WEBVTT
00:00:01.600 --> 00:00:04.080
I think there's a new definition of lazy.
00:00:04.240 --> 00:00:07.040
The old version of lazy is I just don't do work.
00:00:07.120 --> 00:00:08.400
Like I don't do the assignment.
00:00:08.560 --> 00:00:10.000
I'm late for the assignment, whatever.
00:00:10.240 --> 00:00:13.759
I think the new lazy is I ask Claude to do it.
00:00:13.919 --> 00:00:15.039
It does the thing.
00:00:15.199 --> 00:00:15.919
It looks cool.
00:00:16.079 --> 00:00:17.039
And then I ship it.
00:00:17.120 --> 00:00:18.480
Or you know, it's my deck.
00:00:18.559 --> 00:00:19.440
It's my paper.
00:00:19.519 --> 00:00:21.039
And that's the new lazy.
00:00:22.719 --> 00:00:24.399
Welcome back to Product Rebels.
00:00:24.480 --> 00:00:25.519
I'm Heather Samron.
00:00:25.600 --> 00:00:26.719
Recording Solo today.
00:00:26.800 --> 00:00:30.079
I'm excited to introduce a new series on the Product Rebels podcast.
00:00:30.239 --> 00:00:40.880
It's our new Ape AI product leader series where we talk with product leaders who are in the thick of building AI products and infusing AI deep into their product development life cycles.
00:00:41.039 --> 00:00:45.280
My guest today is Brian Crofts, Chief Product Officer at Bamboo HR.
00:00:45.520 --> 00:00:48.799
Brian's career is a great reminder there's no single path into product.
00:00:48.960 --> 00:00:56.240
He started out in corporate finance before finding his way to product and intuit, where he spent over a decade leading product for QuickBooks and TurboTax.
00:00:56.479 --> 00:01:04.640
From there, he took on cheap product roles at, namely Pendo, an audit board, building and scaling SaaS products across industries.
00:01:04.799 --> 00:01:08.239
And he was named Best Product Person of the Year in 2018.
00:01:08.400 --> 00:01:15.439
Now at Bamboo HR, he's leading the company's push into AI with the recent launch of Bamboo AI.
00:01:15.680 --> 00:01:20.799
In this conversation, we get into the origin of Bamboo AI and the problem it was built to solve.
00:01:20.959 --> 00:01:31.840
The idea that CPOs no longer have the pattern recognition they once used and must approach product building and leadership from a place of humility in this age of change.
00:01:32.079 --> 00:01:43.519
And his belief that AI makes the fundamentals of product management matter way more, not less, because when it gets easy to build fast, it gets just as easy to build the wrong thing.
00:01:43.680 --> 00:01:48.799
It's a candid look at what building AI is really teaching him about product leadership.
00:01:48.959 --> 00:01:50.239
Let's dive in.
00:01:59.200 --> 00:02:03.200
I am so excited to be part of a podcast that is titled Product Rebels.
00:02:03.280 --> 00:02:03.599
I love it.
00:02:03.840 --> 00:02:04.560
Love the title.
00:02:06.959 --> 00:02:08.319
I was like, that resonates with me.
00:02:08.560 --> 00:02:09.599
Yeah, I love it.
00:02:09.840 --> 00:02:23.199
And this series of episodes for the Product Rebels podcast is really centering on AI and building stories around AI that other product leaders and product managers can learn from.
00:02:23.520 --> 00:02:31.199
And so you have a really interesting experience at Bamboo AI or Bambo AHR, but specifically in your latest release.
00:02:31.520 --> 00:02:32.080
Recently launched, yeah.
00:02:32.400 --> 00:02:33.280
Recently launched.
00:02:33.520 --> 00:02:41.120
I'd love to just hear kind of a little bit about the problem you were solving with Bamboo AI and take us through that journey.
00:02:41.199 --> 00:02:44.800
What did you learn, your approach, and we'll go from there?
00:02:45.120 --> 00:02:45.280
Yeah.
00:02:45.360 --> 00:02:51.039
I mean, I think first and foremost, I mean, I think all of us are looking at AI as uh a change in technology.
00:02:51.120 --> 00:02:56.479
And so the question is like, okay, how can this technology help us build better solutions?
00:02:56.639 --> 00:03:06.080
And not just that, but also really kind of this opportunity to reframe and rethink how we solve the problem in a completely different way.
00:03:06.400 --> 00:03:16.400
And so one of the things I'm off and on sometimes with the chat experience, but really AI is about having conversations with the AI, with the agent.
00:03:16.560 --> 00:03:25.520
And so kind of the first thing that that we did is we thought about well, what's that kind of paradigm shift experience, even if they're not ready for it?
00:03:25.599 --> 00:03:40.879
Because again, this is where I think we spend a lot of time is like, what is the hypothesis around where behavior will change and where it will start to evolve, especially like on the interface layer, like where we know that we have a strong hypothesis that it'll be heavily fragmented.
00:03:40.960 --> 00:03:42.960
And so so what does that mean?
00:03:43.039 --> 00:03:44.960
And and what is the change in behavior?
00:03:45.120 --> 00:03:56.159
And I think largely the adoption of AI is probably slower than the frontier models or the the speed is slower in terms of the adoption because it requires people changing their behavior.
00:03:56.400 --> 00:04:12.879
And so one of the things that you know we've been focused on is this idea of if you think about UI in like three different ways, like one, like navigate, which is kind of the classic SaaS, like, you know, I click through and I do wayfinding and information architecture, and it's all laid out.
00:04:13.039 --> 00:04:19.360
And I kind of have to know the solution and I have to click around and find the solution in the application.
00:04:19.519 --> 00:04:30.399
And I think we're moving to more of this collaborate experience where you you can click around and wayfind and and and kind of have that still present in your app.
00:04:30.639 --> 00:04:34.560
But now there's this like opportunity to talk to it, to work through it.
00:04:34.720 --> 00:04:42.000
And and so our idea lately is is really been around how to are the people that use the product bring their problems.
00:04:42.160 --> 00:04:48.000
So again, not come with the solutions, okay, I need to do this workflow and onboard this person or do this or that.
00:04:48.079 --> 00:04:50.560
But I'm gonna bring my problems to the AI.
00:04:50.639 --> 00:04:53.600
We're gonna chat about it, and it's gonna recommend solutions.
00:04:53.839 --> 00:04:55.759
That I think is is very different.
00:04:55.920 --> 00:05:03.519
It's not what they're asking for, but the idea is like, let's test, is this a preferred way to use our product?
00:05:03.680 --> 00:05:24.959
And then ultimately, so it kind of goes from like navigate to collaborate kind of with the AI, but you still in the existing interface, and then delegate, where I think ultimately where it's headed for us on kind of the administrative side is the admins of HR teams are going to delegate to the agents to do a lot of the administrivia.
00:05:25.040 --> 00:05:29.439
And they kind of then their role changes as they govern the agents and the AI.
00:05:29.680 --> 00:05:34.480
They're not doing the tasks and they're not just collaborating on the tasks with the AI.
00:05:34.639 --> 00:05:37.279
But right now in our evolution, we're not quite there.
00:05:37.439 --> 00:05:50.480
We're more on the let's collaborate in an interface that you're familiar with and see if not only does usage start to shift from clicks to conversations, does sentiment increase and improve?
00:05:50.639 --> 00:05:59.600
Like, do they find that they're actually solving bigger, like they're able to bring their problems and then the AI can recommend, it has access to all the tools.
00:05:59.759 --> 00:06:04.720
And so it may use a tool like an automation or something that you didn't even think of.
00:06:04.879 --> 00:06:07.439
And so I think that's where we are right now.
00:06:07.680 --> 00:06:16.319
But ultimately moving to this idea of delegation, and then it starts to look a lot like the frontier models where it's purely conversation.
00:06:16.480 --> 00:06:25.199
But then I like to think about well, if if I'm an HR uh administrator and I'm building a report or an analytic, like I don't want to just store it under artifacts.
00:06:25.360 --> 00:06:35.519
I want to put it in kind of the drawer or the shelf that's represented in the UI that I can quickly go back to as an artifact and read and bring up or share.
00:06:35.759 --> 00:06:39.680
So that's kind of the call it the user experience that we've been experimenting with.
00:06:40.000 --> 00:06:40.720
Oh, I love it.
00:06:40.879 --> 00:06:54.480
And and Brian, talk a little bit about because this is this is definitely a dramatic change from how traditional product and user experience and product building teams approach SaaS software, right?
00:06:54.639 --> 00:06:59.439
Like you said, it's all about the information architecture and the navigational architecture.
00:06:59.519 --> 00:07:10.079
And and so how has that changed in terms of your approach to experimentation and measuring whether or not to move forward on a feature or not?
00:07:10.240 --> 00:07:20.240
And and talk to me a little bit about just sort of that that change in, I wouldn't even call it development approach, just product building approach on some of the major stages of the PDLC.
00:07:20.399 --> 00:07:21.040
What's different?
00:07:21.360 --> 00:07:22.560
Yeah, it's a great question.
00:07:22.720 --> 00:07:26.319
And we are actually talking about this as a leadership team today in our product team.
00:07:26.399 --> 00:07:38.480
And and one of the probably the most fun things I've seen is with the design team and their ability to, you know, with AI, obviously, building prototypes that like work.
00:07:38.639 --> 00:07:41.279
Like they don't just, you can't just click around, but they work.
00:07:41.360 --> 00:07:42.160
They have data in them.
00:07:42.399 --> 00:07:43.839
Yeah, just not mock-ups like we use.
00:07:44.240 --> 00:07:45.120
They're not just mock-ups.
00:07:45.279 --> 00:07:48.800
And they're and they can take them from an idea.
00:07:49.040 --> 00:07:57.519
And so we have we built this thing called a product sandbox experience, where it's basically a replica of our entire SaaS application.
00:07:57.759 --> 00:08:09.199
And then what you can do is working with an agent that knows our design like components and we call it fabric, but it's all our componentry, all of our patterns, all the libraries.
00:08:09.439 --> 00:08:10.399
It knows them.
00:08:10.560 --> 00:08:19.279
And this is where we route around completely around Figma, where it's like we basically start building, like it knows all of our design styles and all of that, all the components.
00:08:19.439 --> 00:08:30.800
And so we we just start working with AI to build these prototypes to test these different experiences, and they can take it all the way to where it feels like it's in production, but it's it's not, right?
00:08:30.879 --> 00:08:32.559
It's just it's a prototype.
00:08:32.799 --> 00:08:49.360
And so one of the things that we're working on right now is the ability to take not just that all the way to production, because what we're finding is that the gap between getting to that high fidelity prototype and high quality, scalable production level code, there's still quite a gap.
00:08:49.440 --> 00:08:51.679
It's not just the last mile, it's the last five miles.
00:08:51.759 --> 00:08:52.879
Like it's still a lot of work.
00:08:53.039 --> 00:09:07.039
But the idea though is like if we can take high fidelity prototypes that connect with the user's data and then put that in a lab environment where they know that they're experimenting and then we start to get sentiment on like, is it improved?
00:09:07.120 --> 00:09:11.679
Like, is this by doing this, is it actually is it a delight or does it change the way you do things?
00:09:11.759 --> 00:09:13.440
Like, do you love it as an improvement?
00:09:13.679 --> 00:09:21.759
And then so we get that high fidelity user experience, like testing, data, all that stuff, really high fidelity feedback.
00:09:21.919 --> 00:09:25.759
And then we just now have the confidence to then build that thing fast.
00:09:26.000 --> 00:09:28.399
But like it still requires the build.
00:09:28.559 --> 00:09:34.960
And of course, we're you doing agentic engineering, we're doing all that still requires a build to get to that level, but like then we're just building in more confidence.
00:09:35.039 --> 00:09:46.559
And so it's just kind of flipped where we now have we can build a lot, and so you know, the the cost of building kind of starts going down in a way because we can build stuff that's high confidence.
00:09:46.639 --> 00:09:52.720
And and by the way, I would say like we're on that journey, we're not on the other side of it where it's all working.
00:09:52.879 --> 00:09:56.799
Like we are, we are fumbling sometimes, we are slowing down sometimes.
00:09:56.960 --> 00:09:59.919
Some and then sometimes we're like, oh my gosh, this is so much faster.
00:10:00.080 --> 00:10:01.039
This is incredible.
00:10:01.200 --> 00:10:04.720
There's delights along the way, there's there's times that we're like, what do we do?
00:10:04.879 --> 00:10:08.879
You know, so I would just say like it's not it's not perfect for sure.
00:10:09.120 --> 00:10:22.799
But on the design side and the prototyping, that has been one of my favorite call it changes to the SDLC or PDLC, like in terms of how we're building and and working with our customers in a different and faster way has been around.
00:10:23.120 --> 00:10:23.360
Love it.
00:10:23.600 --> 00:10:24.000
Love it.
00:10:24.159 --> 00:10:30.159
And how did you think about organizing the team for this new interaction pattern with customers?
00:10:30.320 --> 00:10:41.360
Like, did you have to bring in, you know, are your UX designers well versed and and from open AI or, you know, from serious AI cultures?
00:10:41.600 --> 00:10:50.879
Or are you feeling like you're able to take the resources you have and the team and the capabilities you've had and sort of just reorient them in a new way?
00:10:51.279 --> 00:10:52.720
Yeah, it's definitely a mix.
00:10:52.799 --> 00:11:01.200
And one of the things, and and I'll I'll preface this before I answer the question, which is you know, we talk a lot about the different roles and the different structures and all of that.
00:11:01.279 --> 00:11:03.840
And, you know, people wonder like, oh, what's the future of PM?
00:11:03.919 --> 00:11:04.960
What's the future of design?
00:11:05.120 --> 00:11:17.919
Like one thing I I've been trying to remind people a lot is that, and this will ultimately answer the question, but like there is still a lot of scarcity around having great product managers, great designers.
00:11:18.159 --> 00:11:21.120
And the reason why I share that is because like that's not going away.
00:11:21.200 --> 00:11:24.559
Like having great PMs and great designers is not going away.
00:11:24.879 --> 00:11:40.480
And then to answer your question, those designers that are great and scarce, I would say, they're the ones that like grot quickly these new tools and they're constantly exploring, they're curious, they're trying new things.
00:11:40.559 --> 00:11:43.759
Like if they don't know how to do it, they ask the thing, how do I do it?
00:11:44.000 --> 00:11:52.080
And there's just like a tier of designers and PMs that have, you know, I think the the fun term today is like high agency.
00:11:52.320 --> 00:11:55.919
Those high agency designers and PMs are just thriving right now.
00:11:56.000 --> 00:11:59.759
And they're not coming from AI backgrounds or from frontier labs or anything.
00:11:59.919 --> 00:12:01.360
They're they're just great.
00:12:01.519 --> 00:12:04.559
And they're seeing these new tools and new technology.
00:12:04.720 --> 00:12:08.879
And if they're like, if this can help me improve my craft, I'm gonna go do it.
00:12:08.960 --> 00:12:10.080
If not, I don't care.
00:12:10.320 --> 00:12:17.440
And the increase in agency that happens when they take their curiosity, learn skills, and then just go.
00:12:17.600 --> 00:12:19.440
And I'm sure that's true with every team.
00:12:19.519 --> 00:12:26.480
There's just this echelon of designers and PMs that will always be scarce, that have just thrived.
00:12:26.720 --> 00:12:32.720
In fact, tonight we have it's called Founders Award Dinner, and it's the biggest award you can win at Bamboo.
00:12:32.879 --> 00:12:45.519
And one of the winners is a guy from the design team who is high agency building prototypes, rethinking paradigms about like who the end customer should even like just doing so many things.
00:12:45.600 --> 00:12:53.919
And I, you know, I look at his back, it's not necessarily anything special other than like the characteristics of always focused on improving his craft.
00:12:54.080 --> 00:13:00.720
He's always super always curious, not just taking tools for tools' sakes, but trying to get to an outcome.
00:13:00.879 --> 00:13:05.759
And everybody wants works with this guy because whoever he works with, like, oh, it's amazing.
00:13:05.840 --> 00:13:09.440
Our project is is fun and it's exciting and it's innovative.
00:13:09.519 --> 00:13:11.840
And so I'm just like, I want to clone that person.
00:13:12.480 --> 00:13:12.879
Yeah, yeah.
00:13:13.200 --> 00:13:15.840
Man, it is so important to have curiosity.
00:13:15.919 --> 00:13:21.360
It's one of the things we we do a pulse check diagnostic of product building organizations.
00:13:21.440 --> 00:13:37.120
And one of the key dimensions that we sort of probe for is curiosity and a sort of process or an established view on constant experimentation and learning because it's the only way you get there.
00:13:37.360 --> 00:13:44.320
And it's it's so interesting about as a leader, how much trust I have in an individual when they're curious.
00:13:44.960 --> 00:13:49.840
Because I'm not then worried about like how many hours they work or did they hit their commitments.
00:13:50.000 --> 00:13:51.279
Like they're just so curious.
00:13:51.440 --> 00:13:58.159
They want to find the solution to the problem, and they won't stop until they figure it out.
00:13:58.320 --> 00:14:04.080
And it's driven by that foundation of I'm curious, I'm interested, I want to solve this problem.
00:14:04.240 --> 00:14:08.000
And you just infinite trust that goes back to that high agency.
00:14:08.320 --> 00:14:11.279
And not it's it's just not tied to a solution, right?
00:14:11.360 --> 00:14:14.080
They're tied to what is the problem, right?
00:14:14.240 --> 00:14:15.679
That's that's absolutely right.
00:14:16.000 --> 00:14:17.519
And delighting customers, yep.
00:14:17.759 --> 00:14:18.080
Yep.
00:14:18.320 --> 00:14:20.159
Let's talk a little bit more about this, right?
00:14:20.240 --> 00:14:23.279
You you mentioned earlier about, hey, we're not perfect.
00:14:23.360 --> 00:14:27.519
We've got some, you know, we we've slowed ourselves down, we've hit some road bumps.
00:14:27.679 --> 00:14:38.240
I'd love to hear what are the maybe the two biggest challenges that you experienced in this process of launching as of I think July, you said, right?
00:14:38.320 --> 00:14:40.240
You've launched uh Bamboo HR.
00:14:40.399 --> 00:14:48.080
Um, talk to me about kind of what what are the two biggest challenges you and the team had in in delivering that product, that first version.
00:14:48.559 --> 00:14:59.679
Well, I think it was we abandoned some of the process, like in the spirit of like risk on, let's build this vision that we've prototyped.
00:14:59.919 --> 00:15:05.039
And we left some of our processes behind and they and they bit us in the end.
00:15:05.120 --> 00:15:13.279
Like, like being just being able to like clarity on who's building what, how do we hold people accountable, how are we tracking against our commitments?
00:15:13.440 --> 00:15:24.960
Like it was the loosest kind of product development experience that we've we've all had as of late, because we were just doing it so differently, and we did then everything differently.
00:15:25.120 --> 00:15:29.360
And some of it helped us, and then some of it kind of came back to bite us.
00:15:29.440 --> 00:15:37.039
And so when I even think about our V1 launch of Bamboo AI, like I really like it and I love the direction, and we still need to get better.
00:15:37.120 --> 00:15:40.639
And we we can't let go of some of the you know the fundamentals.
00:15:40.879 --> 00:15:45.519
We can't kind of pass up on being able to measure and keep people accountable.
00:15:45.600 --> 00:15:48.159
And it was just it turned into the Wild West a little bit.
00:15:48.320 --> 00:15:51.840
And so that's one where we have to kind of pull that back and say, what went wrong?
00:15:52.000 --> 00:15:53.519
Like, what do we need to kind of bring back?
00:15:53.679 --> 00:15:57.919
Like, how do we make sure that we're both moving in a new direction and a new way?
00:15:58.080 --> 00:16:01.120
But like we were able to do some of the just the core fundamentals.
00:16:01.200 --> 00:16:02.240
So blocking and tackling.
00:16:02.559 --> 00:16:07.919
And so I just think it it was like a Wild West moment, maybe being interpreted from me saying, you know, risk on.
00:16:08.000 --> 00:16:09.840
It's like, okay, just process is in the wind.
00:16:09.919 --> 00:16:13.840
And I think that bit us a little bit, and we've we've caught up and we're recovering from that.
00:16:14.080 --> 00:16:18.000
That would be one thing for sure, and happy to share, share others as well.
00:16:18.240 --> 00:16:19.440
Yeah, this is great.
00:16:19.600 --> 00:16:29.039
And I'd love to, I'm gonna get back to the fundamentals question because I think this is so critical, and this is something that we are running into with a lot of teams, and so I want to have that deeper conversation.
00:16:29.360 --> 00:16:41.759
You also mentioned about in our earlier conversation about bamboo AI, you had mentioned, like, look, we're nowhere near what we want our vision to be, but we don't want to come out ahead of where our customers are.
00:16:41.919 --> 00:16:42.080
Yeah.
00:16:42.240 --> 00:16:52.159
And I thought it was a really interesting discussion to have here, which is how do you think about customer behavior, their readiness for what your vision is?
00:16:52.320 --> 00:16:56.720
And how did you make the decisions you did in your sort of V1 versus what your vision is?
00:16:56.879 --> 00:16:59.360
I'd love to just sort of pick your brain on that one.
00:16:59.679 --> 00:17:03.440
I think sometimes we forget as product people that, like, hey, we're technologists.
00:17:03.519 --> 00:17:09.279
Like I spend two, three hours on YouTube a night listening to Andre Caparthy talk about the future of LLMs.
00:17:09.359 --> 00:17:11.440
Like, like I'm geeking out on this stuff.
00:17:11.519 --> 00:17:19.119
And then when you get to like, okay, what is the thing that we're building and working on, like, oh my gosh, that's not that's not as innovative as like I think it needs to be.
00:17:19.279 --> 00:17:27.680
And so there's like this gap, a little dissonance around like where I think people need to be and in their kind of AI coefficient.
00:17:27.839 --> 00:17:32.880
And then also like the fact that AI has a really bad PR problem, right?
00:17:32.960 --> 00:17:38.079
Like people, there's a negativity around, hey, this AI thing is gonna take your job.
00:17:38.240 --> 00:17:45.599
And now all of a sudden I'm seeing AI feature and functionality start to creep up into the application that I use to do my job.
00:17:45.680 --> 00:17:47.920
And they're kind of like, what are the implications of this?
00:17:48.079 --> 00:17:51.119
And if it's not great, then they're like, get it out of here.
00:17:51.200 --> 00:17:53.519
It's not this is your customer saying this.
00:17:53.839 --> 00:17:55.759
Yeah, it's customers, yeah.
00:17:55.920 --> 00:17:57.519
And they're and so that that's what I mean.
00:17:57.599 --> 00:18:00.400
Like sometimes we forget we're technologists and that we're on the cutting.
00:18:00.720 --> 00:18:02.880
And then we have to remember like, where are our customers?
00:18:02.960 --> 00:18:06.960
Like, they're still just want to run payroll on time and make sure everybody's paid.
00:18:07.119 --> 00:18:14.880
And they're not like, oh my gosh, I need this AI feature that summarizes like, and so we just have to be real with ourselves about now.
00:18:14.960 --> 00:18:19.680
Number one, like we have to just solve big problems and we have to really solve it in a better way.
00:18:19.839 --> 00:18:23.839
And I think then you start to like, but you you definitely have to meet customers where they're at.
00:18:24.000 --> 00:18:28.480
But at the same time, like I don't think they're gonna, I don't think they'll get there without.
00:18:28.640 --> 00:18:33.519
And in fact, we had a customer just the other day said, Hey, our AI strategy is your roadmap.
00:18:33.680 --> 00:18:42.640
Because as you bring in new products and features for us to use that are that are AI native, like that will help us change our process and start to transform more.
00:18:42.720 --> 00:18:44.160
And so you're gonna help us with that.
00:18:44.319 --> 00:18:46.000
So I get really inspired by that.
00:18:46.160 --> 00:18:48.799
But I just think it's And what a what a responsibility, right?
00:18:49.519 --> 00:18:53.599
Yeah, and and it's like we welcome that because we're our mission is to set people free to do great work.
00:18:53.680 --> 00:18:57.759
And we're like, and we think this new technology can really help with that.
00:18:58.240 --> 00:19:02.480
And so we believe in it, but ultimately proof in the putty like does it save you time?
00:19:02.640 --> 00:19:03.599
Does it save you money?
00:19:03.759 --> 00:19:06.640
Like, does it give you more confidence that payroll is run correctly?
00:19:06.799 --> 00:19:11.039
Like these are all things that they care about, not just the bells and whistles of AI.
00:19:11.200 --> 00:19:17.359
And so we just need to make sure that we're bringing the customers along and we're showing them what's possible.
00:19:17.519 --> 00:19:20.240
And so our product managers are very close.
00:19:20.400 --> 00:19:31.200
And then the other thing that we do since we launched is we measure, of course, like sentiment around is like positive, negative sentiment around the feature functionality, you know, and match that with usage.
00:19:31.279 --> 00:19:33.920
And so we it's like high usage, okay, what's the sentiment?
00:19:34.000 --> 00:19:44.319
Because if it's like high usage, bad sentiment, like it's like there's something promising there, but we're not delivering, you know, versus like something nobody really uses, but it's you know, really positive feedback.
00:19:44.400 --> 00:19:46.720
It's like, okay, maybe that's not as impactful.
00:19:46.960 --> 00:19:52.319
But we have another piece of the component, and and the AI picks up on this, which is like this.
00:19:52.559 --> 00:19:55.039
I can't remember what we call, I can look it up.
00:19:55.119 --> 00:19:59.440
We it's I was just looking at this morning, but it's called like AI anxiety.
00:19:59.519 --> 00:20:03.440
So it's like positive, neutral, negative anxiety.
00:20:03.519 --> 00:20:04.559
And it's pulling out.
00:20:04.720 --> 00:20:06.480
That's how you're measuring sent sort of sentiment.
00:20:06.720 --> 00:20:07.519
That's how you're yeah.
00:20:07.839 --> 00:20:16.880
Yeah, because it's like we're we're not necessarily like obviously we're measuring MPS like across the product lines, but within the product lines and the different feature levels, we're trying to capture sentiment.
00:20:16.960 --> 00:20:21.039
And of course, we have usage for these features, but we want to capture sentiment.
00:20:21.279 --> 00:20:24.799
And what's being captured is not just like, is it, you know, is it working?
00:20:24.880 --> 00:20:32.720
Obviously, they can give us feedback and write in comments, but what the summaries are picking up is like there are people that are against you, they don't want they don't like this, it's not meeting the mark.
00:20:32.880 --> 00:20:37.920
There's things people love and like catch the vision, or it's you know, saving them time, money, et cetera.
00:20:38.079 --> 00:20:43.759
And then there's like that other bucket of just AI anxiety that you're you can sense from their comments.
00:20:44.640 --> 00:20:57.279
And so like right now we're seeing that with about 10% of the functionality, like this just gives me AI anxiety, like seeing this in the product, like you can like I don't want anything to do with AI, you know, just stuff like that.
00:20:57.440 --> 00:21:01.039
So it's not even about the feature, the functionality, but like, I don't want anything to do with it.
00:21:01.200 --> 00:21:02.559
And it's like, why is that?
00:21:02.720 --> 00:21:03.680
Let's understand it.
00:21:03.839 --> 00:21:05.200
Let's not just skip that.
00:21:05.359 --> 00:21:11.839
Like, because again, we're trying to change behavior because we think this is a better solution that leads to bigger outcomes.
00:21:12.000 --> 00:21:13.839
You know, otherwise, you know, we wouldn't do it.
00:21:14.000 --> 00:21:31.200
But understanding that sentiment, I think, you know, will kind of give us a little bit of a measure around like where, generally speaking, putting aside solving problems or not, what's their perspective on AI, and how can we help with that and create a better brand around around AI?
00:21:31.519 --> 00:21:32.079
I love it.
00:21:32.319 --> 00:21:47.359
You know, you're bringing up a good question that I want to ask too, which is there's sort of sentiment, AI, comfortability, but there's also this sort of idea of trust, accuracy for highly regulated industries, right?
00:21:47.519 --> 00:21:56.640
There's this what we're finding with a lot of product leaders is that you don't often know whether or not you're losing trust or brand credibility until after you launch the product.
00:21:56.799 --> 00:22:02.160
You gotta watch how the product is being used, which A huge change for most product teams, right?
00:22:02.240 --> 00:22:05.359
It's like ship it and forget it, let's move on to the next feature.
00:22:05.599 --> 00:22:08.400
Your true learning starts after release.
00:22:08.480 --> 00:22:17.039
Like you can't even know whether or not you've maintained, built, or degraded the trust in the brand until after you've seen it used.
00:22:17.279 --> 00:22:21.839
What mechanisms are you putting in place and how has that changed the way your product team works?
00:22:22.079 --> 00:22:26.720
Yeah, this is such a good question because like it's we're dealing with this right now.
00:22:26.880 --> 00:22:32.880
Like there are, let's say there are features that, and again, just just being totally honest, because I think it's helpful.
00:22:33.039 --> 00:22:36.960
I never would want to say we have it all figured out or it's all working great.
00:22:37.039 --> 00:22:38.480
Like there's there's challenges.
00:22:38.640 --> 00:22:45.680
And I think, you know, one of the I don't know if you've experienced this, but definitely in my career, like we hate rolling back products.
00:22:45.839 --> 00:22:47.680
We hate sunsetting products.
00:22:47.759 --> 00:22:52.559
Like it's out there, it's gonna be out there forever, even if I don't improve it or customers don't like it.
00:22:52.640 --> 00:22:57.359
It just, it's just this, I'm sure there's a term for it, but it's just it happens a lot.
00:22:57.599 --> 00:23:08.000
I think now, like if if there is something that we've fully delivered and like the full experience, so sometimes we might roll out an earlier version and it's not the whole vision of what it is.
00:23:08.079 --> 00:23:14.880
And so like you can take negative sentiment with a grain of salt and say, well, they're giving us feedback on this, but it's not really all the way there.
00:23:14.960 --> 00:23:18.640
And so you kind of have to figure out like, do you do that or not in this era?
00:23:18.799 --> 00:23:19.519
Do you know what I mean?
00:23:19.680 --> 00:23:22.319
And yeah, especially when you're dealing with like trust and brand.
00:23:22.559 --> 00:23:40.799
And so I think what we have to do, like you talk about putting in place a mindset to me, and and some tools around, I think it's okay to roll back some of these features that aren't like net positive, net promoter helping grow the brand and the trust as we use AI.
00:23:41.039 --> 00:23:52.720
Because I do think it is a real thing to consider that if you're rolling out AI for AI's sake and you're bolting it on and you're not really solving, you are gonna see degradation in your brand and you're gonna see the trust.
00:23:52.799 --> 00:24:02.480
Like if you just raw out crappy features because I think that's what we're supposed to do, like it is gonna erode trust and it's getting in the way, and people will all of a sudden consider leaving.
00:24:02.559 --> 00:24:04.640
And so we have to take that super serious.
00:24:04.960 --> 00:24:12.640
And I think we just have to get more comfortable with some of the AI tools and functionality that we'll launch won't hit the mark.
00:24:12.880 --> 00:24:24.640
Now, I am really bullish on our vision and what we're building, and that it really will, but it is a journey because it's not just the feature in the old way, it's a it's a paradigm shift.
00:24:24.799 --> 00:24:28.480
And so, how do we how do we show them the better way?
00:24:28.640 --> 00:24:43.440
And then they agree ultimately, it's the better way by watching their usage and their sentiment increase and start to evolve from kind of the navigate kind of user experience to more of that collaborating and delegating type experience.
00:24:43.759 --> 00:24:44.799
Yeah, great.
00:24:45.119 --> 00:24:54.880
Let's shift a little bit now to the how the team is doing this and how you're thinking about building around AI products versus kind of the traditional enterprise software.
00:24:55.119 --> 00:24:56.720
You talk about the fundamentals.
00:24:56.880 --> 00:25:02.400
What fundamentals have changed and what fundamentals have not changed?
00:25:02.799 --> 00:25:03.920
Yeah, I think so.
00:25:04.079 --> 00:25:12.720
Fundamental, you know, I when I think about some of the fundamentals around product and design, it's and I think I call it the like the lost art of great insights.
00:25:12.960 --> 00:25:24.880
You know, it's like we used to, I think, do a better job of spending a lot of time, like you're talking about, like really understanding the problem and the pain point before we jumped into solutions.
00:25:25.119 --> 00:25:31.920
And I think now with AI tools and in their confidence about what you should do or not do, or the deck you built.
00:25:32.079 --> 00:25:35.519
Like I basically I think there's a new definition of lazy.
00:25:36.000 --> 00:25:39.279
And the old version of lazy is I just don't do work.
00:25:39.440 --> 00:25:43.039
Like, you know, I'm I don't do the assignment, I'm late for the assignment, whatever.
00:25:43.200 --> 00:25:48.079
I think the new lazy is I asked Claude to do it, it does the thing.
00:25:48.480 --> 00:25:49.359
It looks cool.
00:25:49.680 --> 00:25:54.079
It looks cool, and then I ship it, uh, or it, you know, it's my deck, it's my paper.
00:25:54.240 --> 00:26:03.519
And that's the new lazy is you actually deliver the thing, the assignment, but you just took at face value whatever the AI gave you.
00:26:03.759 --> 00:26:16.000
And so I'm worried about kind of in the speed of you know, improving the SDLC and moving it forward in this AI era, that we just start, you know, it's the whole slop grenade, right?
00:26:16.079 --> 00:26:19.839
It's the it's the slop discussion that we've been having, you know, as an industry.
00:26:19.920 --> 00:26:21.839
It's like like we've got to make sure.
00:26:21.920 --> 00:26:39.839
And so I think that starts with not being lazy, right, in the new way and really understanding, because then when you are working with AI, you have a discerning eye and a discerning approach of like, this is right, this is wrong, this is helpful, this is not, but it's anchored in a deep, rich customer insight.
00:26:40.000 --> 00:26:43.920
And to get to that insight, it's not just getting a data point.
00:26:44.000 --> 00:26:54.799
It's the combination of data, talking with customers, observing customers, and looking for the thing that you could build a whole roadmap against that durable insight.
00:26:55.039 --> 00:26:58.319
And so I just worry that you know, that we skip that fundamental.
00:26:58.480 --> 00:27:02.400
Like maybe we can do it faster, and maybe we have these great thought partners.
00:27:02.480 --> 00:27:14.960
Like I love to think of my agents as great thought partners, but they need to like, you need to work with them and iterate and challenge to get to the right solution.
00:27:15.200 --> 00:27:17.440
And so I think these tools will help us.
00:27:17.599 --> 00:27:29.599
I just hope that we don't use them as a crutch and say we have insights and say we have deep customer knowledge and empathy and understand the problem when when we when we really don't.
00:27:29.680 --> 00:27:32.400
And that's just gonna lead to sloppy, sloppy products.
00:27:32.559 --> 00:27:34.079
So that's definitely one fundamental.
00:27:34.559 --> 00:27:35.839
Yeah, totally agree.
00:27:36.000 --> 00:27:38.319
I'd love to go even a little deeper on this, right?
00:27:38.400 --> 00:27:40.799
We've got some product leaders, same challenge, right?
00:27:40.880 --> 00:27:51.680
They're really worried that AI puts out some really great looking output, but when they get it in their hands as a leader, they're looking at it going, holy crap, I see this was totally generated by AI.
00:27:51.759 --> 00:27:54.160
It has no sort of fact-based right.
00:27:54.480 --> 00:28:12.880
And so they're doing things like they're developing their own persona as a skill set for AI to sort of dictate to all their product managers you have to use this skill set before you produce anything that's going to be sent to me because I don't want, I want it to be based in, you know, customer reality, the problem we've decided upon, right?
00:28:12.960 --> 00:28:13.759
Like all this stuff.
00:28:14.000 --> 00:28:23.359
Talk to me about some of the tactics that maybe you've thought about in order to minimize slop or to encourage critical thinking or this fundamental.
00:28:23.839 --> 00:28:24.640
Well, it's funny.
00:28:24.880 --> 00:28:31.359
I think it actually first started at like really making fun of some of the slides that the team was producing.
00:28:31.599 --> 00:28:43.920
Like there is a way that Claude writes that you know has tries to have these like the setup, you know, and like these three pithy things, like two products, four solutions, one platform.
00:28:44.000 --> 00:28:46.559
Like this just so clear.
00:28:46.640 --> 00:29:07.200
And like, yeah, yeah, we capture them, we laugh about them as a way to kind of be like, guys, this is like this is not like because I will say, like, there's sometimes there are these like little quips or like these pithy little statements that I love that I did not write, but I'm like, it works, and I love, but like you have to discern and you have to distinguish, like you have to make that, you have to own it.
00:29:07.279 --> 00:29:12.720
It's like owning every word on the slide, owning every product decision, every design decision.
00:29:12.880 --> 00:29:17.200
And so I think you know, it started as just like making fun of people's decks.
00:29:17.279 --> 00:29:20.559
Like, like this is this is crap, this is slop.
00:29:20.640 --> 00:29:20.960
Yeah.
00:29:21.119 --> 00:29:22.720
And like it's gotta be better.
00:29:22.880 --> 00:29:24.720
It's that laziness that I'm talking about.
00:29:24.960 --> 00:29:26.880
And so I think it just kind of started there.
00:29:27.039 --> 00:29:41.200
And then we've come across, you know, I think people are writing their own, like, you know, I've heard like out in the wild, there's like skills that are like for, you know, when presenting to Brian, you know, it's like, what make sure it has this and that, and you know, there's fun stuff like that.
00:29:41.279 --> 00:29:51.359
And then there's like skills like Unslop that takes what you've written and like gets all that stuff out, those crutch phrases and and structures that that seem to be often repeated.
00:29:51.519 --> 00:29:52.720
So it's little things like that.
00:29:52.880 --> 00:30:01.279
But for me, it's just keeps going back to the integrity of like owning, owning the message, the words on the slide, the product decision, the design decision.
00:30:01.440 --> 00:30:08.559
And I hope that AI helped you, but ultimately you own it and you own every word on that slide.
00:30:08.799 --> 00:30:11.839
Because sometimes I'll be like, hey, like what does this mean?
00:30:12.000 --> 00:30:13.759
Like it sounds interesting, but what does it mean?
00:30:13.839 --> 00:30:16.400
And they're like, uh, I don't know, AI wrote that.
00:30:16.559 --> 00:30:18.559
I'm like, why am I reading it then?
00:30:18.720 --> 00:30:19.279
You know?
00:30:19.599 --> 00:30:20.400
Right, right, right.
00:30:20.720 --> 00:30:29.839
And so like, and and by the way, I think tactics are still evolving, but like I think, I think we started just kind of trying to make fun of it in a in a playful way.
00:30:29.920 --> 00:30:38.480
And and again, sharing my definition of lazy and like just mindsets and culture probably have been the biggest tools for us right now.
00:30:38.640 --> 00:30:40.319
You know, it's still still so early.
00:30:40.640 --> 00:30:42.000
Yeah, I mean, it's tough, right?
00:30:42.079 --> 00:30:45.440
And and there is no real best practice yet, right?
00:30:45.599 --> 00:30:48.240
Everyone's sort of this, like you said, wild, wild west.
00:30:48.319 --> 00:30:56.160
We're sort of going out trying to do anything we can to solve some of these pitfalls that we're falling into, and everyone's falling them into them.
00:30:56.480 --> 00:30:58.079
We're just all solving them in different ways.
00:30:58.240 --> 00:31:00.480
So, so super interesting.
00:31:00.720 --> 00:31:07.680
We're coming to an end here, but I want to ask one deep dive just a little bit on one area, and then we'll we'll get it, we'll we'll get our clothes going.
00:31:07.920 --> 00:31:12.000
Talk to me about how the CPO role has changed.
00:31:12.240 --> 00:31:24.079
How do you view your role sort of before and after in this new age of AI, whether that's AI infused into the SDLC or you're building AI products?
00:31:24.240 --> 00:31:28.400
I'd love to just sort of hear how you define a successful CPO.
00:31:28.720 --> 00:31:30.079
It's such a great question.
00:31:30.160 --> 00:31:33.680
And it's something that I think about, I think about a lot lately.
00:31:33.920 --> 00:31:39.279
And it kind of goes back to like what's happening right now, you know, in the world and in the market.
00:31:39.440 --> 00:31:43.039
And four years ago, I felt like I knew a thing or two.
00:31:43.200 --> 00:31:53.599
Like, like that I could run a product team, I could set a vision, a direction, I know how to manage execution, I know how to develop talent and future leaders.
00:31:53.680 --> 00:31:55.279
Like, check, check, check.
00:31:55.440 --> 00:31:58.400
I feel confident in taking that that job.
00:31:58.559 --> 00:32:04.960
And and for the last 10 years, I feel like, you know, it's it's worked and and we've seen success.
00:32:05.200 --> 00:32:11.519
And I think today that's like all up for the air in terms of like how, well, first of all, where's this thing headed?
00:32:11.759 --> 00:32:14.079
Like, that's the biggest thing is where is this going?
00:32:14.240 --> 00:32:15.119
What is the future?
00:32:15.279 --> 00:32:21.680
Like, I I happen to work for uh an 18-year-old SaaS company, like an old school SaaS company, right?
00:32:21.759 --> 00:32:26.960
Like which a few years ago did not feel like an old school SaaS company, but that's the reality.
00:32:27.119 --> 00:32:29.039
And so, like, where is SaaS headed?
00:32:29.119 --> 00:32:31.759
Like, where is where are the frontier models headed?
00:32:31.839 --> 00:32:33.599
Like, are they gonna verticalize?
00:32:33.759 --> 00:32:36.880
Are they gonna start to sound a lot easier, right?
00:32:37.200 --> 00:32:38.000
Yeah, like so.
00:32:38.400 --> 00:32:59.279
So I would say what's interesting, you know, lately is the competition hasn't isn't really the competitive, like if you look at the marketplace and all the competitors, like they became, I'm like less worried about those competitors, and I'm more worried about, or I I wouldn't say worried, I'm more focused on thinking about the frontier models as the competition.
00:32:59.440 --> 00:33:04.960
And I'm trying to understand what are they doing and how do they see the world and how are they shifting the paradigm?
00:33:05.119 --> 00:33:13.200
In other words, like more than ever, CPOs have to spend time really thinking about what is the future of the company?
00:33:13.440 --> 00:33:16.559
And what are the are there new business models?
00:33:16.799 --> 00:33:18.160
Are there new products?
00:33:18.319 --> 00:33:19.839
Are there adjacent spaces?
00:33:20.000 --> 00:33:22.480
Are there things that we need to shift completely?
00:33:22.640 --> 00:33:30.160
Like we focus really on the admin and the the leadership of the people that running the infrastructure of the company and the people management.
00:33:30.400 --> 00:33:33.440
Like maybe we need to spend more time focused on employees.
00:33:33.599 --> 00:33:39.359
Like, how do we give employees superpowers to be the best, you know, do the best work of their life?
00:33:39.519 --> 00:33:45.920
And then even just things like, okay, well, what's our strategy around how we let others connect to our data?
00:33:46.079 --> 00:33:51.359
Like, so we're a system of record and we have valuable and important data and workflows.
00:33:51.519 --> 00:33:54.160
And that plus AI is really powerful.
00:33:54.319 --> 00:34:12.639
I like to think of it as like in the brain, it's like we're the memory, we're the context, and that powered with intelligence that can, you know, these agents that can reconcile, okay, what day do I have, what tools do I have available, like within the system of record to be able to read and write all these different things that are really important for a company?
00:34:12.880 --> 00:34:15.199
It's like that's a powerful combination.
00:34:15.440 --> 00:34:18.639
But well then, what does that mean for the UI layer?
00:34:18.800 --> 00:34:27.519
Like if it's intelligence plus the system of record, am I spending all my time on the frontier user experience and chatting with that?
00:34:27.679 --> 00:34:29.280
And it's pulling data.
00:34:29.360 --> 00:34:33.679
And I'm am I working more there or am I working still in bamboo?
00:34:33.920 --> 00:34:40.079
And so I think the reality though is that the foundation of either of those scenarios is pretty similar.
00:34:40.239 --> 00:34:44.320
And so that's where where I can't answer the future of UIUX.
00:34:44.719 --> 00:34:48.400
I can understand, like, I need a really awesome data model, right?
00:34:48.559 --> 00:34:59.280
I need to be able to have a trust layer that has incredible governance and access levels, roles, and permissions for all the users and compliance and like things that just don't happen overnight.
00:34:59.440 --> 00:35:18.159
I can have my, you know, I can have my own agent platform where I can build agents that leverage that, but I can also allow others or even my customers, you know, I can even give my customers our SDK to our agent platform to build kind of last mile specialized piece of software that our product doesn't have, but they need to.
00:35:18.400 --> 00:35:25.039
So I spend so much time, like I'm on YouTube a couple hours a night, listening to different podcasts.
00:35:25.199 --> 00:35:26.320
I'm building different things.
00:35:26.400 --> 00:35:30.239
I'm trying to understand the different technology stack that we're operating.
00:35:30.320 --> 00:35:31.840
Like, what is an LLM?
00:35:32.000 --> 00:35:33.679
Like, what powers an LLM?
00:35:33.840 --> 00:35:35.920
What, why, why is there talk about this?
00:35:36.000 --> 00:35:37.360
And why is this at risk?
00:35:37.440 --> 00:35:46.079
And so you I think I'm just spending so much more time on like the direction in the future and staying humble of like, I just don't know.
00:35:46.320 --> 00:35:51.039
And so we have to experiment, we have to watch, we have to know what's happening.
00:35:51.280 --> 00:35:54.239
And and then the employees, like, they're like, well, what do we build?
00:35:54.400 --> 00:36:04.480
Like this is and and so on on the flip side, like I think we have to be a bit more directive of like, it's less right now, just right now, about this bottoms up.
00:36:04.639 --> 00:36:08.320
I know I have this area, and I'm gonna bottoms up, tell you my road.
00:36:08.400 --> 00:36:10.400
I think it's a lot more tops down right now.
00:36:10.800 --> 00:36:13.199
And and that used to be kind of like a bad thing.
00:36:13.280 --> 00:36:16.320
Like, don't tell me what I'll tell you my roadmap and my vision.
00:36:16.559 --> 00:36:21.599
But I think in terms of like having a cohesive, like, hey, we're gonna move in this direction.
00:36:21.760 --> 00:36:26.480
I'm finding that the employees want us to be more directive, like, where is it headed?
00:36:26.639 --> 00:36:33.440
And then show us the evidence that that's the right decision or the the anti-pattern that says, oh no, we should maybe think about it this way.
00:36:33.519 --> 00:36:39.519
And so it's basically like going one day, going, I think I know how to do this, to well, I don't know anything.
00:36:39.679 --> 00:36:40.000
Yeah.
00:36:40.159 --> 00:36:40.400
Right.
00:36:40.960 --> 00:36:43.679
Oh my gosh, what I mean, enlightenment, right?
00:36:43.760 --> 00:36:49.519
Like that to me, one of the things that we keep coming across with product leaders is they're sort of overwhelmed.
00:36:49.599 --> 00:36:51.280
They're like, I don't even know where to begin.
00:36:51.519 --> 00:36:56.480
And the thing here, what I heard was nobody knows where to begin.
00:36:56.639 --> 00:36:56.800
Yeah.
00:36:56.960 --> 00:37:05.920
But it takes the leader to be a thought leader through curiosity, through learning, putting stakes in the ground and learning that you're wrong faster than anyone else.
00:37:06.079 --> 00:37:06.320
Right.
00:37:06.400 --> 00:37:11.760
That's that's sort of the role of a CPO now, as opposed to I have all the answers.
00:37:11.920 --> 00:37:13.519
And I shouldn't say you have all the answers.
00:37:13.679 --> 00:37:16.800
You've just been there, done that, you ask the right questions, you get it done, right?
00:37:17.119 --> 00:37:18.320
But that's not the case here.
00:37:18.719 --> 00:37:19.679
Yes, right, right, right.
00:37:19.760 --> 00:37:21.280
Pattern recognition isn't there yet.
00:37:21.599 --> 00:37:24.239
We have this phrase, learn it all is not know it all.
00:37:24.400 --> 00:37:33.360
It's like, and and that's just to kind of under like underline this idea that both at work, at home, on the week, we need to like spend time learning, whether that's building something.
00:37:33.599 --> 00:37:40.719
Yeah, tinkering, yeah, leveraging a new tool, whether it's just listening and studying and and then working with your agents to like ask questions.
00:37:40.880 --> 00:37:51.119
And like it's so fun to have a really smart, intelligent, smart thing with you that will work any hour of the day that you want to, we'll answer any question.
00:37:51.280 --> 00:37:54.719
Like, it's really fun to riff riff with some of these agents.
00:37:54.800 --> 00:38:03.039
And and so again, like it's just taking the time to like learn and to be curious, like you're saying, curious and experimenting.
00:38:03.119 --> 00:38:17.679
And and I think like a year ago versus where I'm at today, like I actually have a pretty strong point of view on where it's headed, but of course, of course, you know, held loosely, like show me a different way, and I'll but it that seems to happen a lot too.
00:38:17.760 --> 00:38:22.639
But I think a year ago, I think it was okay for me to say, I don't know, but I'm gonna go figure it out.
00:38:22.719 --> 00:38:28.639
And now I'm more like, I have a point of view, but it may be wrong, but I have a point of view and let's move.
00:38:28.800 --> 00:38:32.480
And in moving, we may find the actual solution.
00:38:32.559 --> 00:38:41.760
And so we're in a different place, and that actually feels I feel less anxious, but at the same time, you know, the pace of change is happening so fast.
00:38:41.920 --> 00:38:43.199
So yeah, I love it.
00:38:43.280 --> 00:38:43.760
I love it.
00:38:43.920 --> 00:38:46.159
Okay, last two very short questions.
00:38:46.320 --> 00:38:50.079
First is what are the one or two podcasts that you rely on to learn?
00:38:50.480 --> 00:38:52.719
Yeah, so I was just looking at this.
00:38:52.880 --> 00:38:56.480
So I don't know if you followed Dwarkesh Patel.
00:38:56.639 --> 00:38:57.199
He's great.
00:38:57.599 --> 00:39:00.079
And you know, it's like the guest that he has on.
00:39:00.239 --> 00:39:07.679
Like I mentioned listening to Andre Caparthe talk about kind of the nature of LLMs and intelligence, and like it's so fascinating.
00:39:07.760 --> 00:39:16.719
Sometimes it's a little too technical, I'll admit, but it's so good in terms of trying to understand, like maybe one, like what should I go learn about and and study a little bit more?
00:39:16.800 --> 00:39:23.199
But it's just a re a podcaster that really understands and knows like what's going on, and then his guests are great.
00:39:23.360 --> 00:39:25.519
I think Lenny, Lenny has been great.
00:39:25.760 --> 00:39:28.559
I think he's another guy that gets great guests.
00:39:28.800 --> 00:39:29.039
Yeah.
00:39:29.679 --> 00:39:34.559
So I I I really focus on like who gets great guests, but those are two that come to mind.
00:39:34.800 --> 00:39:38.719
I also like my guilty pleasure is like the all-in podcast.
00:39:38.880 --> 00:39:45.119
Oh, like I think you're like wrong on so many things and they're like and they feel a little captured or like I don't know.
00:39:45.199 --> 00:39:47.360
They're just they're they're interesting to me.
00:39:47.599 --> 00:39:54.880
I like watching them because it's kind of a a high level, like maybe how the market thinks about things in a less, much less detailed way.
00:39:54.960 --> 00:40:02.239
It's kind of a like I don't I don't know if I'm learning a lot on those, but I they have such strong opinion and they're entertaining and kind of goofy.
00:40:02.320 --> 00:40:04.880
And and so like I I enjoy that one as well.
00:40:05.119 --> 00:40:05.519
Awesome.
00:40:05.760 --> 00:40:08.960
And then yeah, there's a few more, but it's like there's just so many out there.
00:40:09.039 --> 00:40:12.400
But those are the those top two I mentioned, I I kind of mix between them.
00:40:12.719 --> 00:40:13.199
Awesome.
00:40:13.440 --> 00:40:15.280
Brian, thank you so much for your time.
00:40:15.440 --> 00:40:18.000
This has been insightful, helpful.
00:40:18.239 --> 00:40:25.760
I just know that the product folks out there listening to this is have picked up some nuggets here from you, and I and I really appreciate your time.
00:40:26.000 --> 00:40:27.039
So thank you.
00:40:27.280 --> 00:40:30.719
Thanks for listening to this episode of the Product Rebels Podcast.
00:40:30.800 --> 00:40:42.639
If you enjoyed this conversation and want to learn more from Product Rebels from companies like Netflix, Amplitude, and beyond, please follow us wherever you listen to podcasts and join us for another impactful interview in about two weeks.
00:00:01.600 --> 00:00:04.080
I think there's a new definition of lazy.
00:00:04.240 --> 00:00:07.040
The old version of lazy is I just don't do work.
00:00:07.120 --> 00:00:08.400
Like I don't do the assignment.
00:00:08.560 --> 00:00:10.000
I'm late for the assignment, whatever.
00:00:10.240 --> 00:00:13.759
I think the new lazy is I ask Claude to do it.
00:00:13.919 --> 00:00:15.039
It does the thing.
00:00:15.199 --> 00:00:15.919
It looks cool.
00:00:16.079 --> 00:00:17.039
And then I ship it.
00:00:17.120 --> 00:00:18.480
Or you know, it's my deck.
00:00:18.559 --> 00:00:19.440
It's my paper.
00:00:19.519 --> 00:00:21.039
And that's the new lazy.
00:00:22.719 --> 00:00:24.399
Welcome back to Product Rebels.
00:00:24.480 --> 00:00:25.519
I'm Heather Samron.
00:00:25.600 --> 00:00:26.719
Recording Solo today.
00:00:26.800 --> 00:00:30.079
I'm excited to introduce a new series on the Product Rebels podcast.
00:00:30.239 --> 00:00:40.880
It's our new Ape AI product leader series where we talk with product leaders who are in the thick of building AI products and infusing AI deep into their product development life cycles.
00:00:41.039 --> 00:00:45.280
My guest today is Brian Crofts, Chief Product Officer at Bamboo HR.
00:00:45.520 --> 00:00:48.799
Brian's career is a great reminder there's no single path into product.
00:00:48.960 --> 00:00:56.240
He started out in corporate finance before finding his way to product and intuit, where he spent over a decade leading product for QuickBooks and TurboTax.
00:00:56.479 --> 00:01:04.640
From there, he took on cheap product roles at, namely Pendo, an audit board, building and scaling SaaS products across industries.
00:01:04.799 --> 00:01:08.239
And he was named Best Product Person of the Year in 2018.
00:01:08.400 --> 00:01:15.439
Now at Bamboo HR, he's leading the company's push into AI with the recent launch of Bamboo AI.
00:01:15.680 --> 00:01:20.799
In this conversation, we get into the origin of Bamboo AI and the problem it was built to solve.
00:01:20.959 --> 00:01:31.840
The idea that CPOs no longer have the pattern recognition they once used and must approach product building and leadership from a place of humility in this age of change.
00:01:32.079 --> 00:01:43.519
And his belief that AI makes the fundamentals of product management matter way more, not less, because when it gets easy to build fast, it gets just as easy to build the wrong thing.
00:01:43.680 --> 00:01:48.799
It's a candid look at what building AI is really teaching him about product leadership.
00:01:48.959 --> 00:01:50.239
Let's dive in.
00:01:59.200 --> 00:02:03.200
I am so excited to be part of a podcast that is titled Product Rebels.
00:02:03.280 --> 00:02:03.599
I love it.
00:02:03.840 --> 00:02:04.560
Love the title.
00:02:06.959 --> 00:02:08.319
I was like, that resonates with me.
00:02:08.560 --> 00:02:09.599
Yeah, I love it.
00:02:09.840 --> 00:02:23.199
And this series of episodes for the Product Rebels podcast is really centering on AI and building stories around AI that other product leaders and product managers can learn from.
00:02:23.520 --> 00:02:31.199
And so you have a really interesting experience at Bamboo AI or Bambo AHR, but specifically in your latest release.
00:02:31.520 --> 00:02:32.080
Recently launched, yeah.
00:02:32.400 --> 00:02:33.280
Recently launched.
00:02:33.520 --> 00:02:41.120
I'd love to just hear kind of a little bit about the problem you were solving with Bamboo AI and take us through that journey.
00:02:41.199 --> 00:02:44.800
What did you learn, your approach, and we'll go from there?
00:02:45.120 --> 00:02:45.280
Yeah.
00:02:45.360 --> 00:02:51.039
I mean, I think first and foremost, I mean, I think all of us are looking at AI as uh a change in technology.
00:02:51.120 --> 00:02:56.479
And so the question is like, okay, how can this technology help us build better solutions?
00:02:56.639 --> 00:03:06.080
And not just that, but also really kind of this opportunity to reframe and rethink how we solve the problem in a completely different way.
00:03:06.400 --> 00:03:16.400
And so one of the things I'm off and on sometimes with the chat experience, but really AI is about having conversations with the AI, with the agent.
00:03:16.560 --> 00:03:25.520
And so kind of the first thing that that we did is we thought about well, what's that kind of paradigm shift experience, even if they're not ready for it?
00:03:25.599 --> 00:03:40.879
Because again, this is where I think we spend a lot of time is like, what is the hypothesis around where behavior will change and where it will start to evolve, especially like on the interface layer, like where we know that we have a strong hypothesis that it'll be heavily fragmented.
00:03:40.960 --> 00:03:42.960
And so so what does that mean?
00:03:43.039 --> 00:03:44.960
And and what is the change in behavior?
00:03:45.120 --> 00:03:56.159
And I think largely the adoption of AI is probably slower than the frontier models or the the speed is slower in terms of the adoption because it requires people changing their behavior.
00:03:56.400 --> 00:04:12.879
And so one of the things that you know we've been focused on is this idea of if you think about UI in like three different ways, like one, like navigate, which is kind of the classic SaaS, like, you know, I click through and I do wayfinding and information architecture, and it's all laid out.
00:04:13.039 --> 00:04:19.360
And I kind of have to know the solution and I have to click around and find the solution in the application.
00:04:19.519 --> 00:04:30.399
And I think we're moving to more of this collaborate experience where you you can click around and wayfind and and and kind of have that still present in your app.
00:04:30.639 --> 00:04:34.560
But now there's this like opportunity to talk to it, to work through it.
00:04:34.720 --> 00:04:42.000
And and so our idea lately is is really been around how to are the people that use the product bring their problems.
00:04:42.160 --> 00:04:48.000
So again, not come with the solutions, okay, I need to do this workflow and onboard this person or do this or that.
00:04:48.079 --> 00:04:50.560
But I'm gonna bring my problems to the AI.
00:04:50.639 --> 00:04:53.600
We're gonna chat about it, and it's gonna recommend solutions.
00:04:53.839 --> 00:04:55.759
That I think is is very different.
00:04:55.920 --> 00:05:03.519
It's not what they're asking for, but the idea is like, let's test, is this a preferred way to use our product?
00:05:03.680 --> 00:05:24.959
And then ultimately, so it kind of goes from like navigate to collaborate kind of with the AI, but you still in the existing interface, and then delegate, where I think ultimately where it's headed for us on kind of the administrative side is the admins of HR teams are going to delegate to the agents to do a lot of the administrivia.
00:05:25.040 --> 00:05:29.439
And they kind of then their role changes as they govern the agents and the AI.
00:05:29.680 --> 00:05:34.480
They're not doing the tasks and they're not just collaborating on the tasks with the AI.
00:05:34.639 --> 00:05:37.279
But right now in our evolution, we're not quite there.
00:05:37.439 --> 00:05:50.480
We're more on the let's collaborate in an interface that you're familiar with and see if not only does usage start to shift from clicks to conversations, does sentiment increase and improve?
00:05:50.639 --> 00:05:59.600
Like, do they find that they're actually solving bigger, like they're able to bring their problems and then the AI can recommend, it has access to all the tools.
00:05:59.759 --> 00:06:04.720
And so it may use a tool like an automation or something that you didn't even think of.
00:06:04.879 --> 00:06:07.439
And so I think that's where we are right now.
00:06:07.680 --> 00:06:16.319
But ultimately moving to this idea of delegation, and then it starts to look a lot like the frontier models where it's purely conversation.
00:06:16.480 --> 00:06:25.199
But then I like to think about well, if if I'm an HR uh administrator and I'm building a report or an analytic, like I don't want to just store it under artifacts.
00:06:25.360 --> 00:06:35.519
I want to put it in kind of the drawer or the shelf that's represented in the UI that I can quickly go back to as an artifact and read and bring up or share.
00:06:35.759 --> 00:06:39.680
So that's kind of the call it the user experience that we've been experimenting with.
00:06:40.000 --> 00:06:40.720
Oh, I love it.
00:06:40.879 --> 00:06:54.480
And and Brian, talk a little bit about because this is this is definitely a dramatic change from how traditional product and user experience and product building teams approach SaaS software, right?
00:06:54.639 --> 00:06:59.439
Like you said, it's all about the information architecture and the navigational architecture.
00:06:59.519 --> 00:07:10.079
And and so how has that changed in terms of your approach to experimentation and measuring whether or not to move forward on a feature or not?
00:07:10.240 --> 00:07:20.240
And and talk to me a little bit about just sort of that that change in, I wouldn't even call it development approach, just product building approach on some of the major stages of the PDLC.
00:07:20.399 --> 00:07:21.040
What's different?
00:07:21.360 --> 00:07:22.560
Yeah, it's a great question.
00:07:22.720 --> 00:07:26.319
And we are actually talking about this as a leadership team today in our product team.
00:07:26.399 --> 00:07:38.480
And and one of the probably the most fun things I've seen is with the design team and their ability to, you know, with AI, obviously, building prototypes that like work.
00:07:38.639 --> 00:07:41.279
Like they don't just, you can't just click around, but they work.
00:07:41.360 --> 00:07:42.160
They have data in them.
00:07:42.399 --> 00:07:43.839
Yeah, just not mock-ups like we use.
00:07:44.240 --> 00:07:45.120
They're not just mock-ups.
00:07:45.279 --> 00:07:48.800
And they're and they can take them from an idea.
00:07:49.040 --> 00:07:57.519
And so we have we built this thing called a product sandbox experience, where it's basically a replica of our entire SaaS application.
00:07:57.759 --> 00:08:09.199
And then what you can do is working with an agent that knows our design like components and we call it fabric, but it's all our componentry, all of our patterns, all the libraries.
00:08:09.439 --> 00:08:10.399
It knows them.
00:08:10.560 --> 00:08:19.279
And this is where we route around completely around Figma, where it's like we basically start building, like it knows all of our design styles and all of that, all the components.
00:08:19.439 --> 00:08:30.800
And so we we just start working with AI to build these prototypes to test these different experiences, and they can take it all the way to where it feels like it's in production, but it's it's not, right?
00:08:30.879 --> 00:08:32.559
It's just it's a prototype.
00:08:32.799 --> 00:08:49.360
And so one of the things that we're working on right now is the ability to take not just that all the way to production, because what we're finding is that the gap between getting to that high fidelity prototype and high quality, scalable production level code, there's still quite a gap.
00:08:49.440 --> 00:08:51.679
It's not just the last mile, it's the last five miles.
00:08:51.759 --> 00:08:52.879
Like it's still a lot of work.
00:08:53.039 --> 00:09:07.039
But the idea though is like if we can take high fidelity prototypes that connect with the user's data and then put that in a lab environment where they know that they're experimenting and then we start to get sentiment on like, is it improved?
00:09:07.120 --> 00:09:11.679
Like, is this by doing this, is it actually is it a delight or does it change the way you do things?
00:09:11.759 --> 00:09:13.440
Like, do you love it as an improvement?
00:09:13.679 --> 00:09:21.759
And then so we get that high fidelity user experience, like testing, data, all that stuff, really high fidelity feedback.
00:09:21.919 --> 00:09:25.759
And then we just now have the confidence to then build that thing fast.
00:09:26.000 --> 00:09:28.399
But like it still requires the build.
00:09:28.559 --> 00:09:34.960
And of course, we're you doing agentic engineering, we're doing all that still requires a build to get to that level, but like then we're just building in more confidence.
00:09:35.039 --> 00:09:46.559
And so it's just kind of flipped where we now have we can build a lot, and so you know, the the cost of building kind of starts going down in a way because we can build stuff that's high confidence.
00:09:46.639 --> 00:09:52.720
And and by the way, I would say like we're on that journey, we're not on the other side of it where it's all working.
00:09:52.879 --> 00:09:56.799
Like we are, we are fumbling sometimes, we are slowing down sometimes.
00:09:56.960 --> 00:09:59.919
Some and then sometimes we're like, oh my gosh, this is so much faster.
00:10:00.080 --> 00:10:01.039
This is incredible.
00:10:01.200 --> 00:10:04.720
There's delights along the way, there's there's times that we're like, what do we do?
00:10:04.879 --> 00:10:08.879
You know, so I would just say like it's not it's not perfect for sure.
00:10:09.120 --> 00:10:22.799
But on the design side and the prototyping, that has been one of my favorite call it changes to the SDLC or PDLC, like in terms of how we're building and and working with our customers in a different and faster way has been around.
00:10:23.120 --> 00:10:23.360
Love it.
00:10:23.600 --> 00:10:24.000
Love it.
00:10:24.159 --> 00:10:30.159
And how did you think about organizing the team for this new interaction pattern with customers?
00:10:30.320 --> 00:10:41.360
Like, did you have to bring in, you know, are your UX designers well versed and and from open AI or, you know, from serious AI cultures?
00:10:41.600 --> 00:10:50.879
Or are you feeling like you're able to take the resources you have and the team and the capabilities you've had and sort of just reorient them in a new way?
00:10:51.279 --> 00:10:52.720
Yeah, it's definitely a mix.
00:10:52.799 --> 00:11:01.200
And one of the things, and and I'll I'll preface this before I answer the question, which is you know, we talk a lot about the different roles and the different structures and all of that.
00:11:01.279 --> 00:11:03.840
And, you know, people wonder like, oh, what's the future of PM?
00:11:03.919 --> 00:11:04.960
What's the future of design?
00:11:05.120 --> 00:11:17.919
Like one thing I I've been trying to remind people a lot is that, and this will ultimately answer the question, but like there is still a lot of scarcity around having great product managers, great designers.
00:11:18.159 --> 00:11:21.120
And the reason why I share that is because like that's not going away.
00:11:21.200 --> 00:11:24.559
Like having great PMs and great designers is not going away.
00:11:24.879 --> 00:11:40.480
And then to answer your question, those designers that are great and scarce, I would say, they're the ones that like grot quickly these new tools and they're constantly exploring, they're curious, they're trying new things.
00:11:40.559 --> 00:11:43.759
Like if they don't know how to do it, they ask the thing, how do I do it?
00:11:44.000 --> 00:11:52.080
And there's just like a tier of designers and PMs that have, you know, I think the the fun term today is like high agency.
00:11:52.320 --> 00:11:55.919
Those high agency designers and PMs are just thriving right now.
00:11:56.000 --> 00:11:59.759
And they're not coming from AI backgrounds or from frontier labs or anything.
00:11:59.919 --> 00:12:01.360
They're they're just great.
00:12:01.519 --> 00:12:04.559
And they're seeing these new tools and new technology.
00:12:04.720 --> 00:12:08.879
And if they're like, if this can help me improve my craft, I'm gonna go do it.
00:12:08.960 --> 00:12:10.080
If not, I don't care.
00:12:10.320 --> 00:12:17.440
And the increase in agency that happens when they take their curiosity, learn skills, and then just go.
00:12:17.600 --> 00:12:19.440
And I'm sure that's true with every team.
00:12:19.519 --> 00:12:26.480
There's just this echelon of designers and PMs that will always be scarce, that have just thrived.
00:12:26.720 --> 00:12:32.720
In fact, tonight we have it's called Founders Award Dinner, and it's the biggest award you can win at Bamboo.
00:12:32.879 --> 00:12:45.519
And one of the winners is a guy from the design team who is high agency building prototypes, rethinking paradigms about like who the end customer should even like just doing so many things.
00:12:45.600 --> 00:12:53.919
And I, you know, I look at his back, it's not necessarily anything special other than like the characteristics of always focused on improving his craft.
00:12:54.080 --> 00:13:00.720
He's always super always curious, not just taking tools for tools' sakes, but trying to get to an outcome.
00:13:00.879 --> 00:13:05.759
And everybody wants works with this guy because whoever he works with, like, oh, it's amazing.
00:13:05.840 --> 00:13:09.440
Our project is is fun and it's exciting and it's innovative.
00:13:09.519 --> 00:13:11.840
And so I'm just like, I want to clone that person.
00:13:12.480 --> 00:13:12.879
Yeah, yeah.
00:13:13.200 --> 00:13:15.840
Man, it is so important to have curiosity.
00:13:15.919 --> 00:13:21.360
It's one of the things we we do a pulse check diagnostic of product building organizations.
00:13:21.440 --> 00:13:37.120
And one of the key dimensions that we sort of probe for is curiosity and a sort of process or an established view on constant experimentation and learning because it's the only way you get there.
00:13:37.360 --> 00:13:44.320
And it's it's so interesting about as a leader, how much trust I have in an individual when they're curious.
00:13:44.960 --> 00:13:49.840
Because I'm not then worried about like how many hours they work or did they hit their commitments.
00:13:50.000 --> 00:13:51.279
Like they're just so curious.
00:13:51.440 --> 00:13:58.159
They want to find the solution to the problem, and they won't stop until they figure it out.
00:13:58.320 --> 00:14:04.080
And it's driven by that foundation of I'm curious, I'm interested, I want to solve this problem.
00:14:04.240 --> 00:14:08.000
And you just infinite trust that goes back to that high agency.
00:14:08.320 --> 00:14:11.279
And not it's it's just not tied to a solution, right?
00:14:11.360 --> 00:14:14.080
They're tied to what is the problem, right?
00:14:14.240 --> 00:14:15.679
That's that's absolutely right.
00:14:16.000 --> 00:14:17.519
And delighting customers, yep.
00:14:17.759 --> 00:14:18.080
Yep.
00:14:18.320 --> 00:14:20.159
Let's talk a little bit more about this, right?
00:14:20.240 --> 00:14:23.279
You you mentioned earlier about, hey, we're not perfect.
00:14:23.360 --> 00:14:27.519
We've got some, you know, we we've slowed ourselves down, we've hit some road bumps.
00:14:27.679 --> 00:14:38.240
I'd love to hear what are the maybe the two biggest challenges that you experienced in this process of launching as of I think July, you said, right?
00:14:38.320 --> 00:14:40.240
You've launched uh Bamboo HR.
00:14:40.399 --> 00:14:48.080
Um, talk to me about kind of what what are the two biggest challenges you and the team had in in delivering that product, that first version.
00:14:48.559 --> 00:14:59.679
Well, I think it was we abandoned some of the process, like in the spirit of like risk on, let's build this vision that we've prototyped.
00:14:59.919 --> 00:15:05.039
And we left some of our processes behind and they and they bit us in the end.
00:15:05.120 --> 00:15:13.279
Like, like being just being able to like clarity on who's building what, how do we hold people accountable, how are we tracking against our commitments?
00:15:13.440 --> 00:15:24.960
Like it was the loosest kind of product development experience that we've we've all had as of late, because we were just doing it so differently, and we did then everything differently.
00:15:25.120 --> 00:15:29.360
And some of it helped us, and then some of it kind of came back to bite us.
00:15:29.440 --> 00:15:37.039
And so when I even think about our V1 launch of Bamboo AI, like I really like it and I love the direction, and we still need to get better.
00:15:37.120 --> 00:15:40.639
And we we can't let go of some of the you know the fundamentals.
00:15:40.879 --> 00:15:45.519
We can't kind of pass up on being able to measure and keep people accountable.
00:15:45.600 --> 00:15:48.159
And it was just it turned into the Wild West a little bit.
00:15:48.320 --> 00:15:51.840
And so that's one where we have to kind of pull that back and say, what went wrong?
00:15:52.000 --> 00:15:53.519
Like, what do we need to kind of bring back?
00:15:53.679 --> 00:15:57.919
Like, how do we make sure that we're both moving in a new direction and a new way?
00:15:58.080 --> 00:16:01.120
But like we were able to do some of the just the core fundamentals.
00:16:01.200 --> 00:16:02.240
So blocking and tackling.
00:16:02.559 --> 00:16:07.919
And so I just think it it was like a Wild West moment, maybe being interpreted from me saying, you know, risk on.
00:16:08.000 --> 00:16:09.840
It's like, okay, just process is in the wind.
00:16:09.919 --> 00:16:13.840
And I think that bit us a little bit, and we've we've caught up and we're recovering from that.
00:16:14.080 --> 00:16:18.000
That would be one thing for sure, and happy to share, share others as well.
00:16:18.240 --> 00:16:19.440
Yeah, this is great.
00:16:19.600 --> 00:16:29.039
And I'd love to, I'm gonna get back to the fundamentals question because I think this is so critical, and this is something that we are running into with a lot of teams, and so I want to have that deeper conversation.
00:16:29.360 --> 00:16:41.759
You also mentioned about in our earlier conversation about bamboo AI, you had mentioned, like, look, we're nowhere near what we want our vision to be, but we don't want to come out ahead of where our customers are.
00:16:41.919 --> 00:16:42.080
Yeah.
00:16:42.240 --> 00:16:52.159
And I thought it was a really interesting discussion to have here, which is how do you think about customer behavior, their readiness for what your vision is?
00:16:52.320 --> 00:16:56.720
And how did you make the decisions you did in your sort of V1 versus what your vision is?
00:16:56.879 --> 00:16:59.360
I'd love to just sort of pick your brain on that one.
00:16:59.679 --> 00:17:03.440
I think sometimes we forget as product people that, like, hey, we're technologists.
00:17:03.519 --> 00:17:09.279
Like I spend two, three hours on YouTube a night listening to Andre Caparthy talk about the future of LLMs.
00:17:09.359 --> 00:17:11.440
Like, like I'm geeking out on this stuff.
00:17:11.519 --> 00:17:19.119
And then when you get to like, okay, what is the thing that we're building and working on, like, oh my gosh, that's not that's not as innovative as like I think it needs to be.
00:17:19.279 --> 00:17:27.680
And so there's like this gap, a little dissonance around like where I think people need to be and in their kind of AI coefficient.
00:17:27.839 --> 00:17:32.880
And then also like the fact that AI has a really bad PR problem, right?
00:17:32.960 --> 00:17:38.079
Like people, there's a negativity around, hey, this AI thing is gonna take your job.
00:17:38.240 --> 00:17:45.599
And now all of a sudden I'm seeing AI feature and functionality start to creep up into the application that I use to do my job.
00:17:45.680 --> 00:17:47.920
And they're kind of like, what are the implications of this?
00:17:48.079 --> 00:17:51.119
And if it's not great, then they're like, get it out of here.
00:17:51.200 --> 00:17:53.519
It's not this is your customer saying this.
00:17:53.839 --> 00:17:55.759
Yeah, it's customers, yeah.
00:17:55.920 --> 00:17:57.519
And they're and so that that's what I mean.
00:17:57.599 --> 00:18:00.400
Like sometimes we forget we're technologists and that we're on the cutting.
00:18:00.720 --> 00:18:02.880
And then we have to remember like, where are our customers?
00:18:02.960 --> 00:18:06.960
Like, they're still just want to run payroll on time and make sure everybody's paid.
00:18:07.119 --> 00:18:14.880
And they're not like, oh my gosh, I need this AI feature that summarizes like, and so we just have to be real with ourselves about now.
00:18:14.960 --> 00:18:19.680
Number one, like we have to just solve big problems and we have to really solve it in a better way.
00:18:19.839 --> 00:18:23.839
And I think then you start to like, but you you definitely have to meet customers where they're at.
00:18:24.000 --> 00:18:28.480
But at the same time, like I don't think they're gonna, I don't think they'll get there without.
00:18:28.640 --> 00:18:33.519
And in fact, we had a customer just the other day said, Hey, our AI strategy is your roadmap.
00:18:33.680 --> 00:18:42.640
Because as you bring in new products and features for us to use that are that are AI native, like that will help us change our process and start to transform more.
00:18:42.720 --> 00:18:44.160
And so you're gonna help us with that.
00:18:44.319 --> 00:18:46.000
So I get really inspired by that.
00:18:46.160 --> 00:18:48.799
But I just think it's And what a what a responsibility, right?
00:18:49.519 --> 00:18:53.599
Yeah, and and it's like we welcome that because we're our mission is to set people free to do great work.
00:18:53.680 --> 00:18:57.759
And we're like, and we think this new technology can really help with that.
00:18:58.240 --> 00:19:02.480
And so we believe in it, but ultimately proof in the putty like does it save you time?
00:19:02.640 --> 00:19:03.599
Does it save you money?
00:19:03.759 --> 00:19:06.640
Like, does it give you more confidence that payroll is run correctly?
00:19:06.799 --> 00:19:11.039
Like these are all things that they care about, not just the bells and whistles of AI.
00:19:11.200 --> 00:19:17.359
And so we just need to make sure that we're bringing the customers along and we're showing them what's possible.
00:19:17.519 --> 00:19:20.240
And so our product managers are very close.
00:19:20.400 --> 00:19:31.200
And then the other thing that we do since we launched is we measure, of course, like sentiment around is like positive, negative sentiment around the feature functionality, you know, and match that with usage.
00:19:31.279 --> 00:19:33.920
And so we it's like high usage, okay, what's the sentiment?
00:19:34.000 --> 00:19:44.319
Because if it's like high usage, bad sentiment, like it's like there's something promising there, but we're not delivering, you know, versus like something nobody really uses, but it's you know, really positive feedback.
00:19:44.400 --> 00:19:46.720
It's like, okay, maybe that's not as impactful.
00:19:46.960 --> 00:19:52.319
But we have another piece of the component, and and the AI picks up on this, which is like this.
00:19:52.559 --> 00:19:55.039
I can't remember what we call, I can look it up.
00:19:55.119 --> 00:19:59.440
We it's I was just looking at this morning, but it's called like AI anxiety.
00:19:59.519 --> 00:20:03.440
So it's like positive, neutral, negative anxiety.
00:20:03.519 --> 00:20:04.559
And it's pulling out.
00:20:04.720 --> 00:20:06.480
That's how you're measuring sent sort of sentiment.
00:20:06.720 --> 00:20:07.519
That's how you're yeah.
00:20:07.839 --> 00:20:16.880
Yeah, because it's like we're we're not necessarily like obviously we're measuring MPS like across the product lines, but within the product lines and the different feature levels, we're trying to capture sentiment.
00:20:16.960 --> 00:20:21.039
And of course, we have usage for these features, but we want to capture sentiment.
00:20:21.279 --> 00:20:24.799
And what's being captured is not just like, is it, you know, is it working?
00:20:24.880 --> 00:20:32.720
Obviously, they can give us feedback and write in comments, but what the summaries are picking up is like there are people that are against you, they don't want they don't like this, it's not meeting the mark.
00:20:32.880 --> 00:20:37.920
There's things people love and like catch the vision, or it's you know, saving them time, money, et cetera.
00:20:38.079 --> 00:20:43.759
And then there's like that other bucket of just AI anxiety that you're you can sense from their comments.
00:20:44.640 --> 00:20:57.279
And so like right now we're seeing that with about 10% of the functionality, like this just gives me AI anxiety, like seeing this in the product, like you can like I don't want anything to do with AI, you know, just stuff like that.
00:20:57.440 --> 00:21:01.039
So it's not even about the feature, the functionality, but like, I don't want anything to do with it.
00:21:01.200 --> 00:21:02.559
And it's like, why is that?
00:21:02.720 --> 00:21:03.680
Let's understand it.
00:21:03.839 --> 00:21:05.200
Let's not just skip that.
00:21:05.359 --> 00:21:11.839
Like, because again, we're trying to change behavior because we think this is a better solution that leads to bigger outcomes.
00:21:12.000 --> 00:21:13.839
You know, otherwise, you know, we wouldn't do it.
00:21:14.000 --> 00:21:31.200
But understanding that sentiment, I think, you know, will kind of give us a little bit of a measure around like where, generally speaking, putting aside solving problems or not, what's their perspective on AI, and how can we help with that and create a better brand around around AI?
00:21:31.519 --> 00:21:32.079
I love it.
00:21:32.319 --> 00:21:47.359
You know, you're bringing up a good question that I want to ask too, which is there's sort of sentiment, AI, comfortability, but there's also this sort of idea of trust, accuracy for highly regulated industries, right?
00:21:47.519 --> 00:21:56.640
There's this what we're finding with a lot of product leaders is that you don't often know whether or not you're losing trust or brand credibility until after you launch the product.
00:21:56.799 --> 00:22:02.160
You gotta watch how the product is being used, which A huge change for most product teams, right?
00:22:02.240 --> 00:22:05.359
It's like ship it and forget it, let's move on to the next feature.
00:22:05.599 --> 00:22:08.400
Your true learning starts after release.
00:22:08.480 --> 00:22:17.039
Like you can't even know whether or not you've maintained, built, or degraded the trust in the brand until after you've seen it used.
00:22:17.279 --> 00:22:21.839
What mechanisms are you putting in place and how has that changed the way your product team works?
00:22:22.079 --> 00:22:26.720
Yeah, this is such a good question because like it's we're dealing with this right now.
00:22:26.880 --> 00:22:32.880
Like there are, let's say there are features that, and again, just just being totally honest, because I think it's helpful.
00:22:33.039 --> 00:22:36.960
I never would want to say we have it all figured out or it's all working great.
00:22:37.039 --> 00:22:38.480
Like there's there's challenges.
00:22:38.640 --> 00:22:45.680
And I think, you know, one of the I don't know if you've experienced this, but definitely in my career, like we hate rolling back products.
00:22:45.839 --> 00:22:47.680
We hate sunsetting products.
00:22:47.759 --> 00:22:52.559
Like it's out there, it's gonna be out there forever, even if I don't improve it or customers don't like it.
00:22:52.640 --> 00:22:57.359
It just, it's just this, I'm sure there's a term for it, but it's just it happens a lot.
00:22:57.599 --> 00:23:08.000
I think now, like if if there is something that we've fully delivered and like the full experience, so sometimes we might roll out an earlier version and it's not the whole vision of what it is.
00:23:08.079 --> 00:23:14.880
And so like you can take negative sentiment with a grain of salt and say, well, they're giving us feedback on this, but it's not really all the way there.
00:23:14.960 --> 00:23:18.640
And so you kind of have to figure out like, do you do that or not in this era?
00:23:18.799 --> 00:23:19.519
Do you know what I mean?
00:23:19.680 --> 00:23:22.319
And yeah, especially when you're dealing with like trust and brand.
00:23:22.559 --> 00:23:40.799
And so I think what we have to do, like you talk about putting in place a mindset to me, and and some tools around, I think it's okay to roll back some of these features that aren't like net positive, net promoter helping grow the brand and the trust as we use AI.
00:23:41.039 --> 00:23:52.720
Because I do think it is a real thing to consider that if you're rolling out AI for AI's sake and you're bolting it on and you're not really solving, you are gonna see degradation in your brand and you're gonna see the trust.
00:23:52.799 --> 00:24:02.480
Like if you just raw out crappy features because I think that's what we're supposed to do, like it is gonna erode trust and it's getting in the way, and people will all of a sudden consider leaving.
00:24:02.559 --> 00:24:04.640
And so we have to take that super serious.
00:24:04.960 --> 00:24:12.640
And I think we just have to get more comfortable with some of the AI tools and functionality that we'll launch won't hit the mark.
00:24:12.880 --> 00:24:24.640
Now, I am really bullish on our vision and what we're building, and that it really will, but it is a journey because it's not just the feature in the old way, it's a it's a paradigm shift.
00:24:24.799 --> 00:24:28.480
And so, how do we how do we show them the better way?
00:24:28.640 --> 00:24:43.440
And then they agree ultimately, it's the better way by watching their usage and their sentiment increase and start to evolve from kind of the navigate kind of user experience to more of that collaborating and delegating type experience.
00:24:43.759 --> 00:24:44.799
Yeah, great.
00:24:45.119 --> 00:24:54.880
Let's shift a little bit now to the how the team is doing this and how you're thinking about building around AI products versus kind of the traditional enterprise software.
00:24:55.119 --> 00:24:56.720
You talk about the fundamentals.
00:24:56.880 --> 00:25:02.400
What fundamentals have changed and what fundamentals have not changed?
00:25:02.799 --> 00:25:03.920
Yeah, I think so.
00:25:04.079 --> 00:25:12.720
Fundamental, you know, I when I think about some of the fundamentals around product and design, it's and I think I call it the like the lost art of great insights.
00:25:12.960 --> 00:25:24.880
You know, it's like we used to, I think, do a better job of spending a lot of time, like you're talking about, like really understanding the problem and the pain point before we jumped into solutions.
00:25:25.119 --> 00:25:31.920
And I think now with AI tools and in their confidence about what you should do or not do, or the deck you built.
00:25:32.079 --> 00:25:35.519
Like I basically I think there's a new definition of lazy.
00:25:36.000 --> 00:25:39.279
And the old version of lazy is I just don't do work.
00:25:39.440 --> 00:25:43.039
Like, you know, I'm I don't do the assignment, I'm late for the assignment, whatever.
00:25:43.200 --> 00:25:48.079
I think the new lazy is I asked Claude to do it, it does the thing.
00:25:48.480 --> 00:25:49.359
It looks cool.
00:25:49.680 --> 00:25:54.079
It looks cool, and then I ship it, uh, or it, you know, it's my deck, it's my paper.
00:25:54.240 --> 00:26:03.519
And that's the new lazy is you actually deliver the thing, the assignment, but you just took at face value whatever the AI gave you.
00:26:03.759 --> 00:26:16.000
And so I'm worried about kind of in the speed of you know, improving the SDLC and moving it forward in this AI era, that we just start, you know, it's the whole slop grenade, right?
00:26:16.079 --> 00:26:19.839
It's the it's the slop discussion that we've been having, you know, as an industry.
00:26:19.920 --> 00:26:21.839
It's like like we've got to make sure.
00:26:21.920 --> 00:26:39.839
And so I think that starts with not being lazy, right, in the new way and really understanding, because then when you are working with AI, you have a discerning eye and a discerning approach of like, this is right, this is wrong, this is helpful, this is not, but it's anchored in a deep, rich customer insight.
00:26:40.000 --> 00:26:43.920
And to get to that insight, it's not just getting a data point.
00:26:44.000 --> 00:26:54.799
It's the combination of data, talking with customers, observing customers, and looking for the thing that you could build a whole roadmap against that durable insight.
00:26:55.039 --> 00:26:58.319
And so I just worry that you know, that we skip that fundamental.
00:26:58.480 --> 00:27:02.400
Like maybe we can do it faster, and maybe we have these great thought partners.
00:27:02.480 --> 00:27:14.960
Like I love to think of my agents as great thought partners, but they need to like, you need to work with them and iterate and challenge to get to the right solution.
00:27:15.200 --> 00:27:17.440
And so I think these tools will help us.
00:27:17.599 --> 00:27:29.599
I just hope that we don't use them as a crutch and say we have insights and say we have deep customer knowledge and empathy and understand the problem when when we when we really don't.
00:27:29.680 --> 00:27:32.400
And that's just gonna lead to sloppy, sloppy products.
00:27:32.559 --> 00:27:34.079
So that's definitely one fundamental.
00:27:34.559 --> 00:27:35.839
Yeah, totally agree.
00:27:36.000 --> 00:27:38.319
I'd love to go even a little deeper on this, right?
00:27:38.400 --> 00:27:40.799
We've got some product leaders, same challenge, right?
00:27:40.880 --> 00:27:51.680
They're really worried that AI puts out some really great looking output, but when they get it in their hands as a leader, they're looking at it going, holy crap, I see this was totally generated by AI.
00:27:51.759 --> 00:27:54.160
It has no sort of fact-based right.
00:27:54.480 --> 00:28:12.880
And so they're doing things like they're developing their own persona as a skill set for AI to sort of dictate to all their product managers you have to use this skill set before you produce anything that's going to be sent to me because I don't want, I want it to be based in, you know, customer reality, the problem we've decided upon, right?
00:28:12.960 --> 00:28:13.759
Like all this stuff.
00:28:14.000 --> 00:28:23.359
Talk to me about some of the tactics that maybe you've thought about in order to minimize slop or to encourage critical thinking or this fundamental.
00:28:23.839 --> 00:28:24.640
Well, it's funny.
00:28:24.880 --> 00:28:31.359
I think it actually first started at like really making fun of some of the slides that the team was producing.
00:28:31.599 --> 00:28:43.920
Like there is a way that Claude writes that you know has tries to have these like the setup, you know, and like these three pithy things, like two products, four solutions, one platform.
00:28:44.000 --> 00:28:46.559
Like this just so clear.
00:28:46.640 --> 00:29:07.200
And like, yeah, yeah, we capture them, we laugh about them as a way to kind of be like, guys, this is like this is not like because I will say, like, there's sometimes there are these like little quips or like these pithy little statements that I love that I did not write, but I'm like, it works, and I love, but like you have to discern and you have to distinguish, like you have to make that, you have to own it.
00:29:07.279 --> 00:29:12.720
It's like owning every word on the slide, owning every product decision, every design decision.
00:29:12.880 --> 00:29:17.200
And so I think you know, it started as just like making fun of people's decks.
00:29:17.279 --> 00:29:20.559
Like, like this is this is crap, this is slop.
00:29:20.640 --> 00:29:20.960
Yeah.
00:29:21.119 --> 00:29:22.720
And like it's gotta be better.
00:29:22.880 --> 00:29:24.720
It's that laziness that I'm talking about.
00:29:24.960 --> 00:29:26.880
And so I think it just kind of started there.
00:29:27.039 --> 00:29:41.200
And then we've come across, you know, I think people are writing their own, like, you know, I've heard like out in the wild, there's like skills that are like for, you know, when presenting to Brian, you know, it's like, what make sure it has this and that, and you know, there's fun stuff like that.
00:29:41.279 --> 00:29:51.359
And then there's like skills like Unslop that takes what you've written and like gets all that stuff out, those crutch phrases and and structures that that seem to be often repeated.
00:29:51.519 --> 00:29:52.720
So it's little things like that.
00:29:52.880 --> 00:30:01.279
But for me, it's just keeps going back to the integrity of like owning, owning the message, the words on the slide, the product decision, the design decision.
00:30:01.440 --> 00:30:08.559
And I hope that AI helped you, but ultimately you own it and you own every word on that slide.
00:30:08.799 --> 00:30:11.839
Because sometimes I'll be like, hey, like what does this mean?
00:30:12.000 --> 00:30:13.759
Like it sounds interesting, but what does it mean?
00:30:13.839 --> 00:30:16.400
And they're like, uh, I don't know, AI wrote that.
00:30:16.559 --> 00:30:18.559
I'm like, why am I reading it then?
00:30:18.720 --> 00:30:19.279
You know?
00:30:19.599 --> 00:30:20.400
Right, right, right.
00:30:20.720 --> 00:30:29.839
And so like, and and by the way, I think tactics are still evolving, but like I think, I think we started just kind of trying to make fun of it in a in a playful way.
00:30:29.920 --> 00:30:38.480
And and again, sharing my definition of lazy and like just mindsets and culture probably have been the biggest tools for us right now.
00:30:38.640 --> 00:30:40.319
You know, it's still still so early.
00:30:40.640 --> 00:30:42.000
Yeah, I mean, it's tough, right?
00:30:42.079 --> 00:30:45.440
And and there is no real best practice yet, right?
00:30:45.599 --> 00:30:48.240
Everyone's sort of this, like you said, wild, wild west.
00:30:48.319 --> 00:30:56.160
We're sort of going out trying to do anything we can to solve some of these pitfalls that we're falling into, and everyone's falling them into them.
00:30:56.480 --> 00:30:58.079
We're just all solving them in different ways.
00:30:58.240 --> 00:31:00.480
So, so super interesting.
00:31:00.720 --> 00:31:07.680
We're coming to an end here, but I want to ask one deep dive just a little bit on one area, and then we'll we'll get it, we'll we'll get our clothes going.
00:31:07.920 --> 00:31:12.000
Talk to me about how the CPO role has changed.
00:31:12.240 --> 00:31:24.079
How do you view your role sort of before and after in this new age of AI, whether that's AI infused into the SDLC or you're building AI products?
00:31:24.240 --> 00:31:28.400
I'd love to just sort of hear how you define a successful CPO.
00:31:28.720 --> 00:31:30.079
It's such a great question.
00:31:30.160 --> 00:31:33.680
And it's something that I think about, I think about a lot lately.
00:31:33.920 --> 00:31:39.279
And it kind of goes back to like what's happening right now, you know, in the world and in the market.
00:31:39.440 --> 00:31:43.039
And four years ago, I felt like I knew a thing or two.
00:31:43.200 --> 00:31:53.599
Like, like that I could run a product team, I could set a vision, a direction, I know how to manage execution, I know how to develop talent and future leaders.
00:31:53.680 --> 00:31:55.279
Like, check, check, check.
00:31:55.440 --> 00:31:58.400
I feel confident in taking that that job.
00:31:58.559 --> 00:32:04.960
And and for the last 10 years, I feel like, you know, it's it's worked and and we've seen success.
00:32:05.200 --> 00:32:11.519
And I think today that's like all up for the air in terms of like how, well, first of all, where's this thing headed?
00:32:11.759 --> 00:32:14.079
Like, that's the biggest thing is where is this going?
00:32:14.240 --> 00:32:15.119
What is the future?
00:32:15.279 --> 00:32:21.680
Like, I I happen to work for uh an 18-year-old SaaS company, like an old school SaaS company, right?
00:32:21.759 --> 00:32:26.960
Like which a few years ago did not feel like an old school SaaS company, but that's the reality.
00:32:27.119 --> 00:32:29.039
And so, like, where is SaaS headed?
00:32:29.119 --> 00:32:31.759
Like, where is where are the frontier models headed?
00:32:31.839 --> 00:32:33.599
Like, are they gonna verticalize?
00:32:33.759 --> 00:32:36.880
Are they gonna start to sound a lot easier, right?
00:32:37.200 --> 00:32:38.000
Yeah, like so.
00:32:38.400 --> 00:32:59.279
So I would say what's interesting, you know, lately is the competition hasn't isn't really the competitive, like if you look at the marketplace and all the competitors, like they became, I'm like less worried about those competitors, and I'm more worried about, or I I wouldn't say worried, I'm more focused on thinking about the frontier models as the competition.
00:32:59.440 --> 00:33:04.960
And I'm trying to understand what are they doing and how do they see the world and how are they shifting the paradigm?
00:33:05.119 --> 00:33:13.200
In other words, like more than ever, CPOs have to spend time really thinking about what is the future of the company?
00:33:13.440 --> 00:33:16.559
And what are the are there new business models?
00:33:16.799 --> 00:33:18.160
Are there new products?
00:33:18.319 --> 00:33:19.839
Are there adjacent spaces?
00:33:20.000 --> 00:33:22.480
Are there things that we need to shift completely?
00:33:22.640 --> 00:33:30.160
Like we focus really on the admin and the the leadership of the people that running the infrastructure of the company and the people management.
00:33:30.400 --> 00:33:33.440
Like maybe we need to spend more time focused on employees.
00:33:33.599 --> 00:33:39.359
Like, how do we give employees superpowers to be the best, you know, do the best work of their life?
00:33:39.519 --> 00:33:45.920
And then even just things like, okay, well, what's our strategy around how we let others connect to our data?
00:33:46.079 --> 00:33:51.359
Like, so we're a system of record and we have valuable and important data and workflows.
00:33:51.519 --> 00:33:54.160
And that plus AI is really powerful.
00:33:54.319 --> 00:34:12.639
I like to think of it as like in the brain, it's like we're the memory, we're the context, and that powered with intelligence that can, you know, these agents that can reconcile, okay, what day do I have, what tools do I have available, like within the system of record to be able to read and write all these different things that are really important for a company?
00:34:12.880 --> 00:34:15.199
It's like that's a powerful combination.
00:34:15.440 --> 00:34:18.639
But well then, what does that mean for the UI layer?
00:34:18.800 --> 00:34:27.519
Like if it's intelligence plus the system of record, am I spending all my time on the frontier user experience and chatting with that?
00:34:27.679 --> 00:34:29.280
And it's pulling data.
00:34:29.360 --> 00:34:33.679
And I'm am I working more there or am I working still in bamboo?
00:34:33.920 --> 00:34:40.079
And so I think the reality though is that the foundation of either of those scenarios is pretty similar.
00:34:40.239 --> 00:34:44.320
And so that's where where I can't answer the future of UIUX.
00:34:44.719 --> 00:34:48.400
I can understand, like, I need a really awesome data model, right?
00:34:48.559 --> 00:34:59.280
I need to be able to have a trust layer that has incredible governance and access levels, roles, and permissions for all the users and compliance and like things that just don't happen overnight.
00:34:59.440 --> 00:35:18.159
I can have my, you know, I can have my own agent platform where I can build agents that leverage that, but I can also allow others or even my customers, you know, I can even give my customers our SDK to our agent platform to build kind of last mile specialized piece of software that our product doesn't have, but they need to.
00:35:18.400 --> 00:35:25.039
So I spend so much time, like I'm on YouTube a couple hours a night, listening to different podcasts.
00:35:25.199 --> 00:35:26.320
I'm building different things.
00:35:26.400 --> 00:35:30.239
I'm trying to understand the different technology stack that we're operating.
00:35:30.320 --> 00:35:31.840
Like, what is an LLM?
00:35:32.000 --> 00:35:33.679
Like, what powers an LLM?
00:35:33.840 --> 00:35:35.920
What, why, why is there talk about this?
00:35:36.000 --> 00:35:37.360
And why is this at risk?
00:35:37.440 --> 00:35:46.079
And so you I think I'm just spending so much more time on like the direction in the future and staying humble of like, I just don't know.
00:35:46.320 --> 00:35:51.039
And so we have to experiment, we have to watch, we have to know what's happening.
00:35:51.280 --> 00:35:54.239
And and then the employees, like, they're like, well, what do we build?
00:35:54.400 --> 00:36:04.480
Like this is and and so on on the flip side, like I think we have to be a bit more directive of like, it's less right now, just right now, about this bottoms up.
00:36:04.639 --> 00:36:08.320
I know I have this area, and I'm gonna bottoms up, tell you my road.
00:36:08.400 --> 00:36:10.400
I think it's a lot more tops down right now.
00:36:10.800 --> 00:36:13.199
And and that used to be kind of like a bad thing.
00:36:13.280 --> 00:36:16.320
Like, don't tell me what I'll tell you my roadmap and my vision.
00:36:16.559 --> 00:36:21.599
But I think in terms of like having a cohesive, like, hey, we're gonna move in this direction.
00:36:21.760 --> 00:36:26.480
I'm finding that the employees want us to be more directive, like, where is it headed?
00:36:26.639 --> 00:36:33.440
And then show us the evidence that that's the right decision or the the anti-pattern that says, oh no, we should maybe think about it this way.
00:36:33.519 --> 00:36:39.519
And so it's basically like going one day, going, I think I know how to do this, to well, I don't know anything.
00:36:39.679 --> 00:36:40.000
Yeah.
00:36:40.159 --> 00:36:40.400
Right.
00:36:40.960 --> 00:36:43.679
Oh my gosh, what I mean, enlightenment, right?
00:36:43.760 --> 00:36:49.519
Like that to me, one of the things that we keep coming across with product leaders is they're sort of overwhelmed.
00:36:49.599 --> 00:36:51.280
They're like, I don't even know where to begin.
00:36:51.519 --> 00:36:56.480
And the thing here, what I heard was nobody knows where to begin.
00:36:56.639 --> 00:36:56.800
Yeah.
00:36:56.960 --> 00:37:05.920
But it takes the leader to be a thought leader through curiosity, through learning, putting stakes in the ground and learning that you're wrong faster than anyone else.
00:37:06.079 --> 00:37:06.320
Right.
00:37:06.400 --> 00:37:11.760
That's that's sort of the role of a CPO now, as opposed to I have all the answers.
00:37:11.920 --> 00:37:13.519
And I shouldn't say you have all the answers.
00:37:13.679 --> 00:37:16.800
You've just been there, done that, you ask the right questions, you get it done, right?
00:37:17.119 --> 00:37:18.320
But that's not the case here.
00:37:18.719 --> 00:37:19.679
Yes, right, right, right.
00:37:19.760 --> 00:37:21.280
Pattern recognition isn't there yet.
00:37:21.599 --> 00:37:24.239
We have this phrase, learn it all is not know it all.
00:37:24.400 --> 00:37:33.360
It's like, and and that's just to kind of under like underline this idea that both at work, at home, on the week, we need to like spend time learning, whether that's building something.
00:37:33.599 --> 00:37:40.719
Yeah, tinkering, yeah, leveraging a new tool, whether it's just listening and studying and and then working with your agents to like ask questions.
00:37:40.880 --> 00:37:51.119
And like it's so fun to have a really smart, intelligent, smart thing with you that will work any hour of the day that you want to, we'll answer any question.
00:37:51.280 --> 00:37:54.719
Like, it's really fun to riff riff with some of these agents.
00:37:54.800 --> 00:38:03.039
And and so again, like it's just taking the time to like learn and to be curious, like you're saying, curious and experimenting.
00:38:03.119 --> 00:38:17.679
And and I think like a year ago versus where I'm at today, like I actually have a pretty strong point of view on where it's headed, but of course, of course, you know, held loosely, like show me a different way, and I'll but it that seems to happen a lot too.
00:38:17.760 --> 00:38:22.639
But I think a year ago, I think it was okay for me to say, I don't know, but I'm gonna go figure it out.
00:38:22.719 --> 00:38:28.639
And now I'm more like, I have a point of view, but it may be wrong, but I have a point of view and let's move.
00:38:28.800 --> 00:38:32.480
And in moving, we may find the actual solution.
00:38:32.559 --> 00:38:41.760
And so we're in a different place, and that actually feels I feel less anxious, but at the same time, you know, the pace of change is happening so fast.
00:38:41.920 --> 00:38:43.199
So yeah, I love it.
00:38:43.280 --> 00:38:43.760
I love it.
00:38:43.920 --> 00:38:46.159
Okay, last two very short questions.
00:38:46.320 --> 00:38:50.079
First is what are the one or two podcasts that you rely on to learn?
00:38:50.480 --> 00:38:52.719
Yeah, so I was just looking at this.
00:38:52.880 --> 00:38:56.480
So I don't know if you followed Dwarkesh Patel.
00:38:56.639 --> 00:38:57.199
He's great.
00:38:57.599 --> 00:39:00.079
And you know, it's like the guest that he has on.
00:39:00.239 --> 00:39:07.679
Like I mentioned listening to Andre Caparthe talk about kind of the nature of LLMs and intelligence, and like it's so fascinating.
00:39:07.760 --> 00:39:16.719
Sometimes it's a little too technical, I'll admit, but it's so good in terms of trying to understand, like maybe one, like what should I go learn about and and study a little bit more?
00:39:16.800 --> 00:39:23.199
But it's just a re a podcaster that really understands and knows like what's going on, and then his guests are great.
00:39:23.360 --> 00:39:25.519
I think Lenny, Lenny has been great.
00:39:25.760 --> 00:39:28.559
I think he's another guy that gets great guests.
00:39:28.800 --> 00:39:29.039
Yeah.
00:39:29.679 --> 00:39:34.559
So I I I really focus on like who gets great guests, but those are two that come to mind.
00:39:34.800 --> 00:39:38.719
I also like my guilty pleasure is like the all-in podcast.
00:39:38.880 --> 00:39:45.119
Oh, like I think you're like wrong on so many things and they're like and they feel a little captured or like I don't know.
00:39:45.199 --> 00:39:47.360
They're just they're they're interesting to me.
00:39:47.599 --> 00:39:54.880
I like watching them because it's kind of a a high level, like maybe how the market thinks about things in a less, much less detailed way.
00:39:54.960 --> 00:40:02.239
It's kind of a like I don't I don't know if I'm learning a lot on those, but I they have such strong opinion and they're entertaining and kind of goofy.
00:40:02.320 --> 00:40:04.880
And and so like I I enjoy that one as well.
00:40:05.119 --> 00:40:05.519
Awesome.
00:40:05.760 --> 00:40:08.960
And then yeah, there's a few more, but it's like there's just so many out there.
00:40:09.039 --> 00:40:12.400
But those are the those top two I mentioned, I I kind of mix between them.
00:40:12.719 --> 00:40:13.199
Awesome.
00:40:13.440 --> 00:40:15.280
Brian, thank you so much for your time.
00:40:15.440 --> 00:40:18.000
This has been insightful, helpful.
00:40:18.239 --> 00:40:25.760
I just know that the product folks out there listening to this is have picked up some nuggets here from you, and I and I really appreciate your time.
00:40:26.000 --> 00:40:27.039
So thank you.
00:40:27.280 --> 00:40:30.719
Thanks for listening to this episode of the Product Rebels Podcast.
00:40:30.800 --> 00:40:42.639
If you enjoyed this conversation and want to learn more from Product Rebels from companies like Netflix, Amplitude, and beyond, please follow us wherever you listen to podcasts and join us for another impactful interview in about two weeks.
