WEBVTT
00:00:30.879 --> 00:00:35.710
Welcome back to another episode of What's the BUZZ?, where leaders share how they have turned hype into outcome.
00:00:36.070 --> 00:00:41.679
Today we'll talk about what actually changes when your customer is no longer human, but actually an AI agent.
00:00:42.019 --> 00:00:47.409
And who better to talk about it than someone who's been exploring this topic for quite a bit, Michael Carroll.
00:00:47.899 --> 00:00:48.959
Thank you so much for joining.
00:00:49.349 --> 00:00:50.759
Hey, it's fabulous to be here.
00:00:50.759 --> 00:00:52.519
I've really been looking forward to this.
00:00:52.519 --> 00:00:57.009
This, th- this has been top of mind for me for quite some time.
00:00:57.009 --> 00:00:59.989
I think the world's been headed this way.
00:01:00.349 --> 00:01:02.760
We've been experiencing a little bit of it.
00:01:03.010 --> 00:01:11.469
I- it's, it's gonna transform all of our lives in ways we just can't imagine, and I think this conversation's gonna help us at least start to ask better questions.
00:01:11.670 --> 00:01:12.519
Does that make sense?
00:01:12.840 --> 00:01:13.469
Absolutely.
00:01:13.469 --> 00:01:14.560
I'm looking so forward to it.
00:01:14.579 --> 00:01:19.700
we met in person in, Washington DC a, a couple weeks ago- Yeah at an event.
00:01:19.700 --> 00:01:31.719
and, we just didn't have enough time to, to have that conversation, but it was so insightful already that I said, why don't we spend some time on the show and explore it for, the audience benefit as well?" So I'm so thankful, for you making the time to be with us today.
00:01:32.429 --> 00:01:33.760
I'm looking forward to conversation.
00:01:33.780 --> 00:01:34.129
We'll see-
00:01:34.150 --> 00:01:34.299
Yeah
00:01:34.319 --> 00:01:35.329
we'll see where this goes.
00:01:35.379 --> 00:01:35.549
Yeah
00:01:35.549 --> 00:01:38.339
our original conversation was like,"Wow, we gotta do more of this."
00:01:38.339 --> 00:01:39.599
Yes, exactly.
00:01:39.829 --> 00:01:46.060
Now Michael, not everyone in the audience, might know you yet, so I thought it might be a good idea to introduce yourself real quick, who you are and what you do.
00:01:46.479 --> 00:01:46.759
Yeah.
00:01:46.759 --> 00:01:48.519
So I'm Michael Carroll.
00:01:48.619 --> 00:01:51.589
I've been around a long time, that's why I got gray hair, gray beard.
00:01:52.049 --> 00:02:00.890
And, so I'm actually retired, but as a lot of people- in, in my situation can learn that it can be quite liberating.
00:02:00.890 --> 00:02:05.530
So the thing about retirement is that you get to choose the things you want to go do.
00:02:05.920 --> 00:02:13.699
And in this case, I get to go hang out at national events and talk to you and talk about the things that we think are gonna matter and shape the future, right?
00:02:13.900 --> 00:02:19.550
This is one of those things where I think the right perspective and better questions help sha-cha-shape the future.
00:02:19.810 --> 00:02:41.469
So I get to work on advisory boards and do AI research, and in that AI re- AI research, we push the frontiers of kinda how this technology develops and how we can begin to trust this technology in a way that's different than what we would say we hear the hype about.
00:02:41.500 --> 00:02:51.629
So we make it more trustable, more understandable for people, and maybe improve our perspective on what it not only can do, but more importantly, what it should do for us.
00:02:52.030 --> 00:02:52.650
what it should do.
00:02:52.650 --> 00:02:54.129
I think that's the key point here.
00:02:54.519 --> 00:03:06.799
I remember back in November of'25, I was reading this newar- news article that, Perplexity, was being sued by Amazon because Perplexity had just come out with its a- agentic AI browser.
00:03:07.180 --> 00:03:09.669
and Amazon said,"Hey, we...
00:03:10.210 --> 00:03:28.159
we're actually quite opposed to this because if a browser acts on behalf of a user, if an AI system or agent browses our catalog and recommends products, that means we're actually not making as much ad revenue," which was quite interesting to, to see that Amazon feels this business model is being threatened.
00:03:28.479 --> 00:03:30.109
obviously a big revenue source for them.
00:03:30.370 --> 00:03:36.129
And then I think it was in, in March of'26, the court actually said,"Yes, Amazon, you're right." Yeah.
00:03:36.150 --> 00:03:44.405
and companies like Perplexity and, and others are not allowed to use their agentic browsers to browse or scrape your websites exactly for that reason.
00:03:44.674 --> 00:03:52.294
But for me, that triggered the question of what changes, because companies like Amazon obviously, are making a lot of money from their ads.
00:03:53.014 --> 00:03:57.954
I have two books on, on, on Amazon, and certainly there are ads that I place, so more people see them.
00:03:58.504 --> 00:04:04.194
so I'm, I'm as much of a customer, as, as much as I am, some- someone who's being affected by that change.
00:04:04.995 --> 00:04:09.715
Then there's generative engine optimization, answer engine optimization.
00:04:09.775 --> 00:04:12.905
There are ads in ChatGPT and all of that.
00:04:12.905 --> 00:04:23.954
So- Can we actually trust that an agent that acts on our behalf, that shops on our behalf, finds the objectively most valuable, most relevant product for us?
00:04:24.685 --> 00:04:28.675
I think the answer, I think the answer is maybe, right?
00:04:28.764 --> 00:04:33.805
But first some th- but I think we need to say to ourselves what's actually going on, right?
00:04:33.965 --> 00:04:43.204
and i- in the world there's kind of two things that are coming together, and in the middle we got to sort out how we want that trust to develop.
00:04:43.785 --> 00:04:52.644
On one side, you have this ubiquitous connectivity that is actually enabling kind of this degrees of separation of humanity, right?
00:04:52.844 --> 00:05:00.795
and, and so Stanley Milgram done a big study back in the'60s that said, in the current paradigm, people were six degrees of separation.
00:05:00.844 --> 00:05:01.894
And we all remember that.
00:05:02.134 --> 00:05:11.404
And then along come social media and technology, and now if you believe LinkedIn, it's three, and if you believe oth- other studies, it's 3.4, if you wanna be precise about it.
00:05:11.975 --> 00:05:17.094
and so ubiquitous connectivity that's improving all the time is one side of it.
00:05:17.105 --> 00:05:22.159
And then at the other side of it, the edge is starting to become more and more powerful.
00:05:22.159 --> 00:05:31.469
And process compute at the edge to the point that now whatever shape you want it to take, reasoning is going to become a standard at the edge.
00:05:31.469 --> 00:05:38.240
And by what I mean by reasoning is when you talk about the agentic behavior, right?
00:05:38.290 --> 00:05:44.889
and agents are things that act on your behalf to try and shape outcomes for you that use the ubiquitous connectivity.
00:05:45.120 --> 00:05:54.000
What you're looking at is collapsing the distance between you, everybody else, and everything else- from three to one.
00:05:54.509 --> 00:05:54.800
Yeah.
00:05:55.019 --> 00:06:06.490
And the minute that happens, the question you have to ask yourself, if this, if the, if that reasoning is doing it, now it has to operate through an architecture that enables trust and permission.
00:06:07.129 --> 00:06:17.689
And so it's that architecture in the middle that gi- that gives you the enablement, to access the connectivity and then to engage on your behalf.
00:06:18.159 --> 00:06:21.069
So what, what re- what's required, right?
00:06:21.199 --> 00:06:31.240
Well- Yeah now you have to have some kind of trust and some kind of trace of trust, and some- something that says,"My reasoning and my agent is legitimate," right?
00:06:31.240 --> 00:06:31.500
Yeah.
00:06:31.500 --> 00:06:33.540
"And it's defensible," right?
00:06:33.719 --> 00:06:45.269
So now I have to have some way of say- of saying,"You can trust me," without me having to re-litigate my, the, the- Yeah my existence every time I want to go do something.
00:06:45.670 --> 00:06:48.420
So that's, that whole architecture has to be built.
00:06:48.750 --> 00:06:58.629
It also is a w- Possibly it has s- it, it could give us some strife in society, right?
00:06:58.629 --> 00:07:04.589
Because you could segregate society relative to their ability to pay.
00:07:04.839 --> 00:07:10.500
Relative to other standards that we may not fully appreciate in society, if we don't get that architecture right.
00:07:10.939 --> 00:07:16.170
and I remember not too long ago I w- I was seeing some chatter about, airfare for example.
00:07:16.230 --> 00:07:23.199
if you live in a certain zip code or you make certain transactions of certain amounts, we as- assume that you're more affluent than others.
00:07:23.219 --> 00:07:27.889
If you're more affluent, that means you can pay more for the same ticket that somebody else pays less for.
00:07:28.199 --> 00:07:29.370
Now, is this fair?
00:07:29.370 --> 00:07:30.300
is this just?
00:07:31.019 --> 00:07:33.339
That's one of the questions that, that comes to mind.
00:07:33.829 --> 00:07:41.529
the other part for me is, hey, the World Wide Web is, what, 25, 30 years old in its current form.
00:07:41.560 --> 00:07:44.819
Certainly much older than, than that if you think of ARPANET and others.
00:07:44.819 --> 00:07:50.379
But, you started to see companies optimizing, their ranking in search engines like Google.
00:07:50.409 --> 00:07:54.689
You started seeing companies place ads, to, to influence humans.
00:07:55.000 --> 00:07:56.889
Now we have agents.
00:07:56.930 --> 00:08:05.430
We, we- Yeah have on one hand ads in tools like ChatGPT, and I think it's quite interesting to, to see and to understand how that is going to develop.
00:08:05.689 --> 00:08:07.189
Why are you seeing certain ads?
00:08:07.219 --> 00:08:10.800
Is it because you're getting certain responses to your questions in the chat?
00:08:11.230 --> 00:08:14.629
And are these responses already influenced by the ad that you're going to see?
00:08:15.279 --> 00:08:17.180
And let's assume that's the case.
00:08:17.250 --> 00:08:25.269
how can you piggyback onto that as a brand to say, if somebody asks about these questions, show them our ad, recommend that this is the best product"?
00:08:25.769 --> 00:08:27.990
I think it's gonna be the Wild West for a while.
00:08:28.379 --> 00:08:28.519
Yeah.
00:08:28.569 --> 00:08:43.740
be- because figuring that out So if you think about it for a second, we're going from this architecture of the web, as you talk about, that was click-based- architecture and economy to an architecture that's built around what do I want?
00:08:44.379 --> 00:08:44.649
Yes.
00:08:44.769 --> 00:08:53.050
and the reasoning engines were all, are all going to be the collapse of everybody.
00:08:53.050 --> 00:09:00.730
So the way personal agency works is you spend most of your life trying to orchestrate to get what you want, and you do most of the effort for the orchestration.
00:09:01.230 --> 00:09:18.120
What the re- what the reasoning to edge will do is it will do all the orchestration for you based on what you want, and so it will go out and arrange all the orchestration relative to what you want, and you just have to determine what you want, which now means all the effort for you is in determining why you want what you want.
00:09:18.289 --> 00:09:18.559
True.
00:09:18.679 --> 00:09:18.850
Yes.
00:09:18.950 --> 00:09:28.850
and so n- now that's basically all the effort for you and it does all the orchestration, through the architecture using the ubiquitous connectivity, right?
00:09:29.320 --> 00:09:35.509
How do you build a, how do you build a non-click-based economy-
00:09:35.559 --> 00:09:35.710
Yeah
00:09:35.950 --> 00:09:42.809
that allows that to suffice when basically it does all the optimization for you, right?
00:09:42.809 --> 00:09:44.860
So I can I can I'll give you an example, right?
00:09:45.019 --> 00:09:47.980
This is a real recent personal example for me, right?
00:09:48.340 --> 00:09:58.439
Is that my wife decided it was time for a new automobile And be- because you and I are really proficient AI users, right?
00:09:59.100 --> 00:10:07.070
I built tools that went out and secured quotes through internet commerce, right?
00:10:07.080 --> 00:10:15.740
So I built these things that went out and done it, got responses back through the internet commerce engines of the, o- of the, o- of the dealerships, right?
00:10:15.870 --> 00:10:24.399
In the forms that they were needed to reply in, evaluated all o- all of them relative and optimized every one of them, right?
00:10:24.529 --> 00:10:29.250
And pulled it down to the best option and negotiated the deal for me.
00:10:29.590 --> 00:10:30.679
And I didn't have to do anything.
00:10:31.309 --> 00:10:32.950
they have agentic commerce, right
00:10:32.950 --> 00:10:32.960
there.
00:10:32.960 --> 00:10:33.269
Yeah.
00:10:34.159 --> 00:10:34.230
It's a good
00:10:34.230 --> 00:10:34.559
example.
00:10:34.889 --> 00:10:37.570
It is, and it was a g- it was a great ex- great example.
00:10:37.590 --> 00:10:45.919
and, and it found, y- and f- and found all the best deals, all the best discounts, all that stuff for me, right?
00:10:45.929 --> 00:10:49.220
Just with a little bit of guidance on my end.
00:10:49.490 --> 00:10:49.669
Yeah.
00:10:49.919 --> 00:10:50.919
it, it phenomenal.
00:10:51.490 --> 00:10:57.370
a couple of months ago, I spent some time, working with, Cloud Code and some agents there to optimize my website.
00:10:57.549 --> 00:11:15.309
And I found some, some, repositories on GitHub where users put up these agents, and they were so good that they gave me recommendations for how I can improve the website, not just the copy, but the structure, add schema, so agents can process it more easily.
00:11:15.370 --> 00:11:21.389
Add things like, llms.txt as a file of the most important information you want to see.
00:11:21.870 --> 00:11:26.759
optimize it, in, in little chunks and providing, answers to certain questions.
00:11:27.309 --> 00:11:35.960
And Similarly, to, to your scenario, I think that's right now something that many organizations either are already thinking about, and if they're not, they should be thinking about.
00:11:35.960 --> 00:11:47.789
How do I make my business more visible and accessible to agents on the internet so that they can transact with me in that, in, in that car dealership example?
00:11:48.289 --> 00:11:54.909
I, I think that's I think that's what all com- commerce is going to have to learn, right?
00:11:54.929 --> 00:12:21.219
Is that, that ads, in the way that we used to th- used to think about things, aren't what they were, right?
00:12:21.219 --> 00:12:31.759
And that commerce is gonna be more engaging in the way that how do I deal with an agent that's doing the negotiation work for you?
00:12:32.129 --> 00:12:32.279
Yes.
00:12:32.309 --> 00:12:42.589
And that's gonna be a different kind of world because all of it's gonna be optimized through the agent itself doing the mediat- the intermediation work.
00:12:42.769 --> 00:12:46.719
And that's gonna be different for all of us in the op- For- in the optimization.
00:12:47.659 --> 00:13:07.589
Look, so for about the last two years, I've had this slide in my, keynotes and workshops where I said,"Hey, th- this is the, the progression we're going through." We're starting with single agents, then we have two or three agents, then we have agents that are part of a business function, and then at some point we'll have agents that transact on your or your company's behalf with your vendors.
00:13:08.019 --> 00:13:16.339
And, while this is one example from our consumer lives, the same scenario obviously happens in B2B too and is already happening to, to some extent.
00:13:16.339 --> 00:13:28.549
So when you have two agents, vendor and customer transacting with each other, exchanging information, finding information, I think to your point earlier, trust becomes a, a big factor.
00:13:29.129 --> 00:13:33.439
to what level and to what degree are they authorized to transact with each other?
00:13:33.459 --> 00:13:34.500
Is this a repeat order?
00:13:34.500 --> 00:13:35.519
Is this something new?
00:13:35.839 --> 00:13:37.500
What information are they basing it on?
00:13:37.529 --> 00:13:39.829
Is it, a- aligned with our contracts?
00:13:39.849 --> 00:13:43.339
Whose interest, does the agent represent and have in mind?
00:13:43.620 --> 00:13:44.609
Can they be swayed?
00:13:45.159 --> 00:13:45.789
Things like that.
00:13:46.029 --> 00:13:48.120
so I think that's a worthy discussion, right?
00:13:48.120 --> 00:13:51.839
so if you think about organizations for a second, right?
00:13:52.339 --> 00:13:58.019
It, an or- an or- organizations are largely built for two things, right?
00:13:58.259 --> 00:14:03.779
Intermediation, which is basically the translation and transfer of informa- information.
00:14:04.199 --> 00:14:11.769
And then mediation, which is the transformation of things, information or things from one state to the other.
00:14:11.989 --> 00:14:20.219
and when you look at that transformation piece and that permission to transform something from one state to another, right?
00:14:20.569 --> 00:14:30.899
the intermediation part, AI is gonna be really good at doing that because the rules-based state stuff is largely reversible.
00:14:31.459 --> 00:14:37.599
And so I think we're gonna see largely that become autonomous in companies, right?
00:14:37.599 --> 00:14:40.659
So whether it's, if you're familiar with the Produ- Purdue Stack, right?
00:14:40.870 --> 00:14:42.250
You're gonna see a lot of that stuff.
00:14:42.299 --> 00:14:54.000
so the, the categories that exist to support the Purdue Stack will probably become somewhat disrupted, but the functionality is gonna stay- for, for many of those components.
00:14:54.419 --> 00:15:06.579
But when you look at, when you look at the mediation piece, that, that transformation part of things, anything that requires some type of approval or permission.
00:15:07.079 --> 00:15:16.309
So if, i-if you're gonna wear the responsibility of it, you're going to need some type of defensibility because you...
00:15:16.319 --> 00:15:18.019
That responsibility transfers to you.
00:15:18.019 --> 00:15:20.875
A machine is never going to bear the responsibility of it.
00:15:20.875 --> 00:15:21.009
Yes.
00:15:21.009 --> 00:15:23.529
So it has to have a causal chain of reasoning.
00:15:23.769 --> 00:15:25.789
Explainable AI is an abomination.
00:15:26.069 --> 00:15:32.709
it's an in- it's an invitation to re-litigation that we already have too...
00:15:32.719 --> 00:15:38.804
way too many alignment meetings to litigate what it is we think we meant, right- Yes when we agreed to do something.
00:15:39.284 --> 00:15:39.604
Yep.
00:15:39.794 --> 00:15:45.625
And so causal chain of reasoning is why causal exists in the first place, right?
00:15:45.734 --> 00:16:06.774
And I think having that as a part of what trust means, so it's auditable, it's defensible- Yeah it's legitimacy is a part of what it would mean to, for the transformation of one thing, whether it's information to some other type, of information or change of state, is part of what this architecture will have to have.
00:16:07.274 --> 00:16:14.084
But it sounds like we're really just at the very beginning of all of this, where we are figuring out what can we use these agents for.
00:16:14.104 --> 00:16:14.384
we're...
00:16:14.444 --> 00:16:20.104
many of us in, in, in the industry are exploring and at a minimum, on a personal level, what can these agents do?
00:16:20.104 --> 00:16:20.995
How do we build them?
00:16:21.394 --> 00:16:22.684
What are some of the pitfalls?
00:16:23.024 --> 00:16:26.834
Larger organizations, are not just experimenting, but are rolling this out.
00:16:26.834 --> 00:16:31.954
I've seen some good examples in financial services, in, in business functions like customer service and so on.
00:16:32.544 --> 00:16:46.784
But to your point, I think this next part is really a-about building this trust architecture, having, ha-having also a record of what was done by whom, human or agent, at what point in time, based on what information.
00:16:47.284 --> 00:16:54.414
So you can go back and we can actually trace and say,"Okay, this is what happened and why it happened," which hasn't always been that, that easy.
00:16:54.454 --> 00:17:01.834
Thinking back even a couple of years ago with machine learning and looking for things like, SHAP or some of the, explainable, AI methods.
00:17:02.194 --> 00:17:13.365
So Early, but we know where this needs to go if something like this a- agentic commerce is to take off and is to, expand from consumer in- into B2B as well.
00:17:13.865 --> 00:17:16.795
If you can't trust it, it won't happen.
00:17:17.065 --> 00:17:17.444
Yes.
00:17:17.605 --> 00:17:18.634
I'm, I'm, I think that's...
00:17:18.694 --> 00:17:23.355
I think we're probably everybody would be in agreement is if you can't trust it, it won't happen.
00:17:23.355 --> 00:17:26.954
and, and if you go back to our experience at L- with LLMs, right?
00:17:26.964 --> 00:17:26.984
Yeah.
00:17:27.484 --> 00:17:35.275
for those of us that were early into it, we didn't go very far and we went,"Wait a minute," right?
00:17:35.375 --> 00:17:36.714
it's just cool, but it's...
00:17:37.495 --> 00:17:38.995
We better be really good at this.
00:17:39.015 --> 00:17:39.035
Yeah.
00:17:39.214 --> 00:17:41.495
And most of us got really good at it fast.
00:17:41.755 --> 00:17:42.005
Yeah.
00:17:42.825 --> 00:17:47.434
to, to some extent, I think that's still the, the question that needs to be solved next.
00:17:47.474 --> 00:17:53.755
how do we trust these systems to act reliably and repeatedly, so we know that we can give them more autonomy?
00:17:54.495 --> 00:17:56.185
The fact that they can do it, yeah, sure.
00:17:56.275 --> 00:17:58.535
I think that the technical capabilities are there.
00:17:59.154 --> 00:18:00.745
but does it work every time?
00:18:00.785 --> 00:18:02.144
What happens if it doesn't?
00:18:02.365 --> 00:18:04.335
How do you become aware of it?
00:18:04.485 --> 00:18:05.315
Can you jump in?
00:18:05.315 --> 00:18:06.664
Does another agent jump in?
00:18:06.695 --> 00:18:11.225
And, to your point, I think those are the architectural decisions that they need to be made.
00:18:11.965 --> 00:18:12.235
Yeah.
00:18:12.235 --> 00:18:14.215
So that's what, when you hear me talk about causal-
00:18:14.715 --> 00:18:14.845
Yeah
00:18:15.325 --> 00:18:16.875
causal is an architecture, right?
00:18:16.875 --> 00:18:21.505
It's not, it's, it's not, it's not anything else other than a reasoning architecture.
00:18:21.904 --> 00:18:21.924
Yeah.
00:18:21.935 --> 00:18:26.965
and so Pearl basically laid the foundation, a guy named Judea Pearl laid the foundation of it.
00:18:27.365 --> 00:18:31.634
And then so it's the application of how you apply thinking to something.
00:18:31.965 --> 00:18:35.295
And so that, that's, that gets you to the trust part of it.
00:18:35.924 --> 00:18:41.845
But it's not m- a lot of people would say, that means it's deterministic." Causal is not deterministic.
00:18:41.845 --> 00:18:44.785
It's, it, it's actually agentic in na- in nature, right?
00:18:44.785 --> 00:18:50.715
It's the same thing that, that people do to create agency, and that's what it's based on is the human agency model.
00:18:50.985 --> 00:18:54.455
It means that I understand what's true about the world around me.
00:18:54.644 --> 00:19:02.995
I understand what I can do to make things true, and then that's, that's how I build a hypothesis that if I do this, what if, right?
00:19:03.025 --> 00:19:09.279
And then I have the ability to understand the actions I took relative to the evidence of the outcome.
00:19:09.279 --> 00:19:09.585
Yeah.
00:19:09.585 --> 00:19:12.134
And then I learn from that evidence and I iterate forward.
00:19:12.144 --> 00:19:14.205
That's how I shape outcomes for the future.
00:19:14.215 --> 00:19:16.825
That's causal, but it's not deterministic.
00:19:17.495 --> 00:19:19.625
So what do you think, where is this going?
00:19:19.654 --> 00:19:22.545
Again, I'm coming back to the ads that, that we're seeing.
00:19:22.545 --> 00:19:31.144
You search for something on the web, maybe you still use Google to do that, and shortly after, ads start popping up for the same or similar products.
00:19:31.755 --> 00:19:35.894
So you might take an action because you see or you keep seeing these ads.
00:19:36.174 --> 00:19:37.817
I've been listening to a lot of music- Yeah.
00:19:37.817 --> 00:19:40.458
I think- on, on Spotify lately, I keep getting the same ads.
00:19:40.878 --> 00:19:43.837
and again, I know, but I'm not taking action.
00:19:43.858 --> 00:19:57.468
how is that changing with agents and how do you know that an agent acted on verifiable trusted information or because somebody, convinced it that's the best option for them that you should pursue?
00:19:58.057 --> 00:20:00.948
I, I really don't think ads go away, right?
00:20:01.087 --> 00:20:04.688
I think, I, I think you'll still see ads show up on TV.
00:20:04.688 --> 00:20:06.488
I think you'll still see them sh-show up on phone.
00:20:06.538 --> 00:20:15.758
But if you, if an ad piques your interest, I think what'll happen is that because, people still like to watch YouTube, people still like to watch TV.
00:20:15.758 --> 00:20:15.768
Sure.
00:20:15.768 --> 00:20:19.718
I still think there's room for ads to show up.
00:20:19.718 --> 00:20:32.468
But what I think what'll happen is that people will turn to their agent and say,"I'm interested in this.""Is this the best option for me?" And I think the agent will evaluate that.
00:20:32.817 --> 00:20:45.304
Now- They may say that's their preference, and what the ads will likely do is they will, w- they will actually try and bias you to inform your agent that's your preference, more or less.
00:20:45.364 --> 00:20:53.124
And I think you'll probably see that become more a tone in the ad- to get you to use your agent to inform preference.
00:20:53.244 --> 00:20:55.514
Likely that's what will become true.
00:20:55.964 --> 00:20:57.104
that's an interesting angle.
00:20:57.104 --> 00:21:04.564
so far I've been thinking about it, as, vendors trying to make it into Perplexity's, recommendations, right?
00:21:04.564 --> 00:21:13.334
or ranking or getting into OpenAI's ranking so whenever you ask a question, that's your ad that's being shown, and that's the level of influence.
00:21:13.554 --> 00:21:23.234
Plus, I feel a lot of ads that we listen to or, or that we see, they are, they're geared towards triggering some kind of em- emotional reaction.
00:21:23.284 --> 00:21:38.173
There's a need, there's a paucity or an absence of something, that- that you're feeling or that you didn't know you were feeling, but now after watching it, you are, craving ice cream or fast food, or you see this new car or what have you.
00:21:39.134 --> 00:21:43.964
But there's usually some em- emotional component that, vendors want you to act upon.
00:21:44.384 --> 00:21:47.714
And, and- Now, now with agents, that, that emotional component goes away.
00:21:47.714 --> 00:21:55.354
it's in, in theory, factual and objective, but what are the ways in which your agent will be influenced to take an action?
00:21:55.403 --> 00:21:57.264
so, so let's play that out for a second, right?
00:21:57.264 --> 00:21:57.443
Yeah.
00:21:57.864 --> 00:22:02.584
so who's, so who's gonna build the, who's gonna build the reasoning agents at the edge, right?
00:22:03.044 --> 00:22:13.634
and then who's gonna have the greatest amount of influence over the ability of, so today we have this commerce engine, whether it's Google or Amazon or whatever, right?
00:22:13.943 --> 00:22:23.703
And then the reasoning agents at the edge, who's gonna have the greatest amount of influence over you that basically connects the ad space, right?
00:22:23.714 --> 00:22:36.153
whatever platform you're watching it on, and then that ties to the reasoning agent, and then the l- the last closure loop would be n- probably not the, ubiquitous connectivity.
00:22:36.324 --> 00:22:37.564
It's the architecture of trust.
00:22:37.574 --> 00:22:45.834
and so whoever closes the triad of my ability to influence the reasoning agent, right?
00:22:45.834 --> 00:22:45.974
Yep.
00:22:45.994 --> 00:22:53.163
And the platform of influence, which is the optics that you're watching things through- the media influence, right?
00:22:53.304 --> 00:22:55.634
And that architecture of trust that we talk about.
00:22:55.644 --> 00:23:02.574
If you can close that triad, now you got something that's fairly controllable, right?
00:23:02.574 --> 00:23:02.844
Yes.
00:23:03.744 --> 00:23:04.384
Perfect.
00:23:04.443 --> 00:23:06.114
who ha- who has the power to do that?
00:23:06.114 --> 00:23:07.264
That's a g- that's a good question.
00:23:07.614 --> 00:23:14.274
I think the ones that, that, that have been, called dead, the ones that have been doing it for a long time.
00:23:14.774 --> 00:23:16.724
So the likes of Google, right?
00:23:16.724 --> 00:23:17.314
For example.
00:23:17.324 --> 00:23:17.344
Yeah.
00:23:17.384 --> 00:23:19.144
I think Google is, is not dead,
00:23:19.254 --> 00:23:20.374
Oh, no, no way.
00:23:20.423 --> 00:23:21.564
No way, shape, or form.
00:23:21.574 --> 00:23:21.604
Sure.
00:23:21.614 --> 00:23:24.324
and I think you've, you got the reach to do it too, right?
00:23:24.354 --> 00:23:24.714
It's-
00:23:24.754 --> 00:23:24.953
Yeah
00:23:25.124 --> 00:23:26.734
it, I think it's, that's likely
00:23:27.163 --> 00:23:27.413
Yeah.
00:23:28.364 --> 00:23:36.203
So then it becomes more of an evolution of how to, advertise, search, influence, in- It's orchestration, right?
00:23:36.214 --> 00:23:36.673
that change, right?
00:23:36.903 --> 00:23:37.094
Yeah.
00:23:37.604 --> 00:23:38.464
So fascinating.
00:23:39.274 --> 00:23:46.663
Michael, we've covered a lot of ground, over the last 27 minutes of what changes when your customer all of a sudden is an agent.
00:23:47.403 --> 00:23:55.604
I was wondering if there are any, key takeaways, maybe two, three key takeaways that, that you'd like to take the audience away from our conversation today before we wrap up.
00:23:56.294 --> 00:24:02.294
I think the major key takeaway is this collapse of distance between people and things.
00:24:02.504 --> 00:24:14.084
Is that the reality of our lives is that, that this confluence of technology and is really pointing back to people, right?
00:24:14.134 --> 00:24:21.574
And so all the confluence, so whether, whether it's biomedical, whatever, everything's pointing back to people.
00:24:21.903 --> 00:24:25.534
AI is actually pointing back to people, right?
00:24:25.874 --> 00:24:28.364
And so it's gonna reduce the separation.
00:24:28.364 --> 00:24:36.034
It's not, it's not pulling people apart, it's actually pulling people back closer to everything in your li- in your lives, right?
00:24:36.474 --> 00:24:41.014
And it's separating you by one degree from everything and everybody else, right?
00:24:41.074 --> 00:24:44.804
If you choose for it to be, right?
00:24:44.804 --> 00:24:53.014
And so you gotta ch- you gotta make this choice to, to be more engaged'cause these tools are massively powerful.
00:24:53.834 --> 00:25:09.724
and I think the ability to touch everybody else on the planet that's connected to this, everything else- on the planet that's connected to this, all the biomedical that's connect- that's gonna be connected to this, it is wonderfully exciting.
00:25:10.274 --> 00:25:15.084
But everything's pointed back at the individual, and that's as it should be
00:25:15.584 --> 00:25:18.544
So that means I need to update, my, my phrase, right?
00:25:18.544 --> 00:25:19.834
and my knowledge.
00:25:19.844 --> 00:25:23.183
It's not the six degrees of s- of Kevin Bacon, but the one degree
00:25:23.403 --> 00:25:24.514
It's the one degree.
00:25:24.964 --> 00:25:25.314
Yeah.
00:25:25.494 --> 00:25:25.774
Yeah.
00:25:25.844 --> 00:25:27.044
it's w- liberating.
00:25:27.534 --> 00:25:35.173
l- yes, for sure, and super exciting, especially if you get hands-on and you explore how this works and, think about where this might go eventually.
00:25:35.443 --> 00:25:35.604
Yeah.
00:25:35.984 --> 00:25:37.634
Will there be any downtime for us?
00:25:37.663 --> 00:25:47.514
Because you and I work more than we ever have now, even though, even a- if it said AI, AI would, A- AI would reduce the need for us, it surely hasn't.
00:25:48.054 --> 00:25:50.974
So to me, that's where the human component comes in, to say- Yeah,
00:25:51.153 --> 00:25:51.554
absolutely
00:25:51.834 --> 00:25:57.734
I need a break, and I need to take a break, and I want to take a break." and, the world is still turning when I come back to it.
00:25:57.964 --> 00:25:58.754
Yes, it is.
00:25:59.044 --> 00:25:59.814
So wonderful.
00:25:59.834 --> 00:26:01.614
Michael, it's been a pleasure having you on the show.
00:26:01.614 --> 00:26:09.344
Thank you for sharing your expertise with us, and I really appreciate this, unscripted format and continuing our conversation that started a couple weeks ago in Washington.
00:26:09.413 --> 00:26:10.494
This has been fabulous.
00:26:10.494 --> 00:26:11.913
I thank you so very much.
00:00:30.879 --> 00:00:35.710
Welcome back to another episode of What's the BUZZ?, where leaders share how they have turned hype into outcome.
00:00:36.070 --> 00:00:41.679
Today we'll talk about what actually changes when your customer is no longer human, but actually an AI agent.
00:00:42.019 --> 00:00:47.409
And who better to talk about it than someone who's been exploring this topic for quite a bit, Michael Carroll.
00:00:47.899 --> 00:00:48.959
Thank you so much for joining.
00:00:49.349 --> 00:00:50.759
Hey, it's fabulous to be here.
00:00:50.759 --> 00:00:52.519
I've really been looking forward to this.
00:00:52.519 --> 00:00:57.009
This, th- this has been top of mind for me for quite some time.
00:00:57.009 --> 00:00:59.989
I think the world's been headed this way.
00:01:00.349 --> 00:01:02.760
We've been experiencing a little bit of it.
00:01:03.010 --> 00:01:11.469
I- it's, it's gonna transform all of our lives in ways we just can't imagine, and I think this conversation's gonna help us at least start to ask better questions.
00:01:11.670 --> 00:01:12.519
Does that make sense?
00:01:12.840 --> 00:01:13.469
Absolutely.
00:01:13.469 --> 00:01:14.560
I'm looking so forward to it.
00:01:14.579 --> 00:01:19.700
we met in person in, Washington DC a, a couple weeks ago- Yeah at an event.
00:01:19.700 --> 00:01:31.719
and, we just didn't have enough time to, to have that conversation, but it was so insightful already that I said, why don't we spend some time on the show and explore it for, the audience benefit as well?" So I'm so thankful, for you making the time to be with us today.
00:01:32.429 --> 00:01:33.760
I'm looking forward to conversation.
00:01:33.780 --> 00:01:34.129
We'll see-
00:01:34.150 --> 00:01:34.299
Yeah
00:01:34.319 --> 00:01:35.329
we'll see where this goes.
00:01:35.379 --> 00:01:35.549
Yeah
00:01:35.549 --> 00:01:38.339
our original conversation was like,"Wow, we gotta do more of this."
00:01:38.339 --> 00:01:39.599
Yes, exactly.
00:01:39.829 --> 00:01:46.060
Now Michael, not everyone in the audience, might know you yet, so I thought it might be a good idea to introduce yourself real quick, who you are and what you do.
00:01:46.479 --> 00:01:46.759
Yeah.
00:01:46.759 --> 00:01:48.519
So I'm Michael Carroll.
00:01:48.619 --> 00:01:51.589
I've been around a long time, that's why I got gray hair, gray beard.
00:01:52.049 --> 00:02:00.890
And, so I'm actually retired, but as a lot of people- in, in my situation can learn that it can be quite liberating.
00:02:00.890 --> 00:02:05.530
So the thing about retirement is that you get to choose the things you want to go do.
00:02:05.920 --> 00:02:13.699
And in this case, I get to go hang out at national events and talk to you and talk about the things that we think are gonna matter and shape the future, right?
00:02:13.900 --> 00:02:19.550
This is one of those things where I think the right perspective and better questions help sha-cha-shape the future.
00:02:19.810 --> 00:02:41.469
So I get to work on advisory boards and do AI research, and in that AI re- AI research, we push the frontiers of kinda how this technology develops and how we can begin to trust this technology in a way that's different than what we would say we hear the hype about.
00:02:41.500 --> 00:02:51.629
So we make it more trustable, more understandable for people, and maybe improve our perspective on what it not only can do, but more importantly, what it should do for us.
00:02:52.030 --> 00:02:52.650
what it should do.
00:02:52.650 --> 00:02:54.129
I think that's the key point here.
00:02:54.519 --> 00:03:06.799
I remember back in November of'25, I was reading this newar- news article that, Perplexity, was being sued by Amazon because Perplexity had just come out with its a- agentic AI browser.
00:03:07.180 --> 00:03:09.669
and Amazon said,"Hey, we...
00:03:10.210 --> 00:03:28.159
we're actually quite opposed to this because if a browser acts on behalf of a user, if an AI system or agent browses our catalog and recommends products, that means we're actually not making as much ad revenue," which was quite interesting to, to see that Amazon feels this business model is being threatened.
00:03:28.479 --> 00:03:30.109
obviously a big revenue source for them.
00:03:30.370 --> 00:03:36.129
And then I think it was in, in March of'26, the court actually said,"Yes, Amazon, you're right." Yeah.
00:03:36.150 --> 00:03:44.405
and companies like Perplexity and, and others are not allowed to use their agentic browsers to browse or scrape your websites exactly for that reason.
00:03:44.674 --> 00:03:52.294
But for me, that triggered the question of what changes, because companies like Amazon obviously, are making a lot of money from their ads.
00:03:53.014 --> 00:03:57.954
I have two books on, on, on Amazon, and certainly there are ads that I place, so more people see them.
00:03:58.504 --> 00:04:04.194
so I'm, I'm as much of a customer, as, as much as I am, some- someone who's being affected by that change.
00:04:04.995 --> 00:04:09.715
Then there's generative engine optimization, answer engine optimization.
00:04:09.775 --> 00:04:12.905
There are ads in ChatGPT and all of that.
00:04:12.905 --> 00:04:23.954
So- Can we actually trust that an agent that acts on our behalf, that shops on our behalf, finds the objectively most valuable, most relevant product for us?
00:04:24.685 --> 00:04:28.675
I think the answer, I think the answer is maybe, right?
00:04:28.764 --> 00:04:33.805
But first some th- but I think we need to say to ourselves what's actually going on, right?
00:04:33.965 --> 00:04:43.204
and i- in the world there's kind of two things that are coming together, and in the middle we got to sort out how we want that trust to develop.
00:04:43.785 --> 00:04:52.644
On one side, you have this ubiquitous connectivity that is actually enabling kind of this degrees of separation of humanity, right?
00:04:52.844 --> 00:05:00.795
and, and so Stanley Milgram done a big study back in the'60s that said, in the current paradigm, people were six degrees of separation.
00:05:00.844 --> 00:05:01.894
And we all remember that.
00:05:02.134 --> 00:05:11.404
And then along come social media and technology, and now if you believe LinkedIn, it's three, and if you believe oth- other studies, it's 3.4, if you wanna be precise about it.
00:05:11.975 --> 00:05:17.094
and so ubiquitous connectivity that's improving all the time is one side of it.
00:05:17.105 --> 00:05:22.159
And then at the other side of it, the edge is starting to become more and more powerful.
00:05:22.159 --> 00:05:31.469
And process compute at the edge to the point that now whatever shape you want it to take, reasoning is going to become a standard at the edge.
00:05:31.469 --> 00:05:38.240
And by what I mean by reasoning is when you talk about the agentic behavior, right?
00:05:38.290 --> 00:05:44.889
and agents are things that act on your behalf to try and shape outcomes for you that use the ubiquitous connectivity.
00:05:45.120 --> 00:05:54.000
What you're looking at is collapsing the distance between you, everybody else, and everything else- from three to one.
00:05:54.509 --> 00:05:54.800
Yeah.
00:05:55.019 --> 00:06:06.490
And the minute that happens, the question you have to ask yourself, if this, if the, if that reasoning is doing it, now it has to operate through an architecture that enables trust and permission.
00:06:07.129 --> 00:06:17.689
And so it's that architecture in the middle that gi- that gives you the enablement, to access the connectivity and then to engage on your behalf.
00:06:18.159 --> 00:06:21.069
So what, what re- what's required, right?
00:06:21.199 --> 00:06:31.240
Well- Yeah now you have to have some kind of trust and some kind of trace of trust, and some- something that says,"My reasoning and my agent is legitimate," right?
00:06:31.240 --> 00:06:31.500
Yeah.
00:06:31.500 --> 00:06:33.540
"And it's defensible," right?
00:06:33.719 --> 00:06:45.269
So now I have to have some way of say- of saying,"You can trust me," without me having to re-litigate my, the, the- Yeah my existence every time I want to go do something.
00:06:45.670 --> 00:06:48.420
So that's, that whole architecture has to be built.
00:06:48.750 --> 00:06:58.629
It also is a w- Possibly it has s- it, it could give us some strife in society, right?
00:06:58.629 --> 00:07:04.589
Because you could segregate society relative to their ability to pay.
00:07:04.839 --> 00:07:10.500
Relative to other standards that we may not fully appreciate in society, if we don't get that architecture right.
00:07:10.939 --> 00:07:16.170
and I remember not too long ago I w- I was seeing some chatter about, airfare for example.
00:07:16.230 --> 00:07:23.199
if you live in a certain zip code or you make certain transactions of certain amounts, we as- assume that you're more affluent than others.
00:07:23.219 --> 00:07:27.889
If you're more affluent, that means you can pay more for the same ticket that somebody else pays less for.
00:07:28.199 --> 00:07:29.370
Now, is this fair?
00:07:29.370 --> 00:07:30.300
is this just?
00:07:31.019 --> 00:07:33.339
That's one of the questions that, that comes to mind.
00:07:33.829 --> 00:07:41.529
the other part for me is, hey, the World Wide Web is, what, 25, 30 years old in its current form.
00:07:41.560 --> 00:07:44.819
Certainly much older than, than that if you think of ARPANET and others.
00:07:44.819 --> 00:07:50.379
But, you started to see companies optimizing, their ranking in search engines like Google.
00:07:50.409 --> 00:07:54.689
You started seeing companies place ads, to, to influence humans.
00:07:55.000 --> 00:07:56.889
Now we have agents.
00:07:56.930 --> 00:08:05.430
We, we- Yeah have on one hand ads in tools like ChatGPT, and I think it's quite interesting to, to see and to understand how that is going to develop.
00:08:05.689 --> 00:08:07.189
Why are you seeing certain ads?
00:08:07.219 --> 00:08:10.800
Is it because you're getting certain responses to your questions in the chat?
00:08:11.230 --> 00:08:14.629
And are these responses already influenced by the ad that you're going to see?
00:08:15.279 --> 00:08:17.180
And let's assume that's the case.
00:08:17.250 --> 00:08:25.269
how can you piggyback onto that as a brand to say, if somebody asks about these questions, show them our ad, recommend that this is the best product"?
00:08:25.769 --> 00:08:27.990
I think it's gonna be the Wild West for a while.
00:08:28.379 --> 00:08:28.519
Yeah.
00:08:28.569 --> 00:08:43.740
be- because figuring that out So if you think about it for a second, we're going from this architecture of the web, as you talk about, that was click-based- architecture and economy to an architecture that's built around what do I want?
00:08:44.379 --> 00:08:44.649
Yes.
00:08:44.769 --> 00:08:53.050
and the reasoning engines were all, are all going to be the collapse of everybody.
00:08:53.050 --> 00:09:00.730
So the way personal agency works is you spend most of your life trying to orchestrate to get what you want, and you do most of the effort for the orchestration.
00:09:01.230 --> 00:09:18.120
What the re- what the reasoning to edge will do is it will do all the orchestration for you based on what you want, and so it will go out and arrange all the orchestration relative to what you want, and you just have to determine what you want, which now means all the effort for you is in determining why you want what you want.
00:09:18.289 --> 00:09:18.559
True.
00:09:18.679 --> 00:09:18.850
Yes.
00:09:18.950 --> 00:09:28.850
and so n- now that's basically all the effort for you and it does all the orchestration, through the architecture using the ubiquitous connectivity, right?
00:09:29.320 --> 00:09:35.509
How do you build a, how do you build a non-click-based economy-
00:09:35.559 --> 00:09:35.710
Yeah
00:09:35.950 --> 00:09:42.809
that allows that to suffice when basically it does all the optimization for you, right?
00:09:42.809 --> 00:09:44.860
So I can I can I'll give you an example, right?
00:09:45.019 --> 00:09:47.980
This is a real recent personal example for me, right?
00:09:48.340 --> 00:09:58.439
Is that my wife decided it was time for a new automobile And be- because you and I are really proficient AI users, right?
00:09:59.100 --> 00:10:07.070
I built tools that went out and secured quotes through internet commerce, right?
00:10:07.080 --> 00:10:15.740
So I built these things that went out and done it, got responses back through the internet commerce engines of the, o- of the, o- of the dealerships, right?
00:10:15.870 --> 00:10:24.399
In the forms that they were needed to reply in, evaluated all o- all of them relative and optimized every one of them, right?
00:10:24.529 --> 00:10:29.250
And pulled it down to the best option and negotiated the deal for me.
00:10:29.590 --> 00:10:30.679
And I didn't have to do anything.
00:10:31.309 --> 00:10:32.950
they have agentic commerce, right
00:10:32.950 --> 00:10:32.960
there.
00:10:32.960 --> 00:10:33.269
Yeah.
00:10:34.159 --> 00:10:34.230
It's a good
00:10:34.230 --> 00:10:34.559
example.
00:10:34.889 --> 00:10:37.570
It is, and it was a g- it was a great ex- great example.
00:10:37.590 --> 00:10:45.919
and, and it found, y- and f- and found all the best deals, all the best discounts, all that stuff for me, right?
00:10:45.929 --> 00:10:49.220
Just with a little bit of guidance on my end.
00:10:49.490 --> 00:10:49.669
Yeah.
00:10:49.919 --> 00:10:50.919
it, it phenomenal.
00:10:51.490 --> 00:10:57.370
a couple of months ago, I spent some time, working with, Cloud Code and some agents there to optimize my website.
00:10:57.549 --> 00:11:15.309
And I found some, some, repositories on GitHub where users put up these agents, and they were so good that they gave me recommendations for how I can improve the website, not just the copy, but the structure, add schema, so agents can process it more easily.
00:11:15.370 --> 00:11:21.389
Add things like, llms.txt as a file of the most important information you want to see.
00:11:21.870 --> 00:11:26.759
optimize it, in, in little chunks and providing, answers to certain questions.
00:11:27.309 --> 00:11:35.960
And Similarly, to, to your scenario, I think that's right now something that many organizations either are already thinking about, and if they're not, they should be thinking about.
00:11:35.960 --> 00:11:47.789
How do I make my business more visible and accessible to agents on the internet so that they can transact with me in that, in, in that car dealership example?
00:11:48.289 --> 00:11:54.909
I, I think that's I think that's what all com- commerce is going to have to learn, right?
00:11:54.929 --> 00:12:21.219
Is that, that ads, in the way that we used to th- used to think about things, aren't what they were, right?
00:12:21.219 --> 00:12:31.759
And that commerce is gonna be more engaging in the way that how do I deal with an agent that's doing the negotiation work for you?
00:12:32.129 --> 00:12:32.279
Yes.
00:12:32.309 --> 00:12:42.589
And that's gonna be a different kind of world because all of it's gonna be optimized through the agent itself doing the mediat- the intermediation work.
00:12:42.769 --> 00:12:46.719
And that's gonna be different for all of us in the op- For- in the optimization.
00:12:47.659 --> 00:13:07.589
Look, so for about the last two years, I've had this slide in my, keynotes and workshops where I said,"Hey, th- this is the, the progression we're going through." We're starting with single agents, then we have two or three agents, then we have agents that are part of a business function, and then at some point we'll have agents that transact on your or your company's behalf with your vendors.
00:13:08.019 --> 00:13:16.339
And, while this is one example from our consumer lives, the same scenario obviously happens in B2B too and is already happening to, to some extent.
00:13:16.339 --> 00:13:28.549
So when you have two agents, vendor and customer transacting with each other, exchanging information, finding information, I think to your point earlier, trust becomes a, a big factor.
00:13:29.129 --> 00:13:33.439
to what level and to what degree are they authorized to transact with each other?
00:13:33.459 --> 00:13:34.500
Is this a repeat order?
00:13:34.500 --> 00:13:35.519
Is this something new?
00:13:35.839 --> 00:13:37.500
What information are they basing it on?
00:13:37.529 --> 00:13:39.829
Is it, a- aligned with our contracts?
00:13:39.849 --> 00:13:43.339
Whose interest, does the agent represent and have in mind?
00:13:43.620 --> 00:13:44.609
Can they be swayed?
00:13:45.159 --> 00:13:45.789
Things like that.
00:13:46.029 --> 00:13:48.120
so I think that's a worthy discussion, right?
00:13:48.120 --> 00:13:51.839
so if you think about organizations for a second, right?
00:13:52.339 --> 00:13:58.019
It, an or- an or- organizations are largely built for two things, right?
00:13:58.259 --> 00:14:03.779
Intermediation, which is basically the translation and transfer of informa- information.
00:14:04.199 --> 00:14:11.769
And then mediation, which is the transformation of things, information or things from one state to the other.
00:14:11.989 --> 00:14:20.219
and when you look at that transformation piece and that permission to transform something from one state to another, right?
00:14:20.569 --> 00:14:30.899
the intermediation part, AI is gonna be really good at doing that because the rules-based state stuff is largely reversible.
00:14:31.459 --> 00:14:37.599
And so I think we're gonna see largely that become autonomous in companies, right?
00:14:37.599 --> 00:14:40.659
So whether it's, if you're familiar with the Produ- Purdue Stack, right?
00:14:40.870 --> 00:14:42.250
You're gonna see a lot of that stuff.
00:14:42.299 --> 00:14:54.000
so the, the categories that exist to support the Purdue Stack will probably become somewhat disrupted, but the functionality is gonna stay- for, for many of those components.
00:14:54.419 --> 00:15:06.579
But when you look at, when you look at the mediation piece, that, that transformation part of things, anything that requires some type of approval or permission.
00:15:07.079 --> 00:15:16.309
So if, i-if you're gonna wear the responsibility of it, you're going to need some type of defensibility because you...
00:15:16.319 --> 00:15:18.019
That responsibility transfers to you.
00:15:18.019 --> 00:15:20.875
A machine is never going to bear the responsibility of it.
00:15:20.875 --> 00:15:21.009
Yes.
00:15:21.009 --> 00:15:23.529
So it has to have a causal chain of reasoning.
00:15:23.769 --> 00:15:25.789
Explainable AI is an abomination.
00:15:26.069 --> 00:15:32.709
it's an in- it's an invitation to re-litigation that we already have too...
00:15:32.719 --> 00:15:38.804
way too many alignment meetings to litigate what it is we think we meant, right- Yes when we agreed to do something.
00:15:39.284 --> 00:15:39.604
Yep.
00:15:39.794 --> 00:15:45.625
And so causal chain of reasoning is why causal exists in the first place, right?
00:15:45.734 --> 00:16:06.774
And I think having that as a part of what trust means, so it's auditable, it's defensible- Yeah it's legitimacy is a part of what it would mean to, for the transformation of one thing, whether it's information to some other type, of information or change of state, is part of what this architecture will have to have.
00:16:07.274 --> 00:16:14.084
But it sounds like we're really just at the very beginning of all of this, where we are figuring out what can we use these agents for.
00:16:14.104 --> 00:16:14.384
we're...
00:16:14.444 --> 00:16:20.104
many of us in, in, in the industry are exploring and at a minimum, on a personal level, what can these agents do?
00:16:20.104 --> 00:16:20.995
How do we build them?
00:16:21.394 --> 00:16:22.684
What are some of the pitfalls?
00:16:23.024 --> 00:16:26.834
Larger organizations, are not just experimenting, but are rolling this out.
00:16:26.834 --> 00:16:31.954
I've seen some good examples in financial services, in, in business functions like customer service and so on.
00:16:32.544 --> 00:16:46.784
But to your point, I think this next part is really a-about building this trust architecture, having, ha-having also a record of what was done by whom, human or agent, at what point in time, based on what information.
00:16:47.284 --> 00:16:54.414
So you can go back and we can actually trace and say,"Okay, this is what happened and why it happened," which hasn't always been that, that easy.
00:16:54.454 --> 00:17:01.834
Thinking back even a couple of years ago with machine learning and looking for things like, SHAP or some of the, explainable, AI methods.
00:17:02.194 --> 00:17:13.365
So Early, but we know where this needs to go if something like this a- agentic commerce is to take off and is to, expand from consumer in- into B2B as well.
00:17:13.865 --> 00:17:16.795
If you can't trust it, it won't happen.
00:17:17.065 --> 00:17:17.444
Yes.
00:17:17.605 --> 00:17:18.634
I'm, I'm, I think that's...
00:17:18.694 --> 00:17:23.355
I think we're probably everybody would be in agreement is if you can't trust it, it won't happen.
00:17:23.355 --> 00:17:26.954
and, and if you go back to our experience at L- with LLMs, right?
00:17:26.964 --> 00:17:26.984
Yeah.
00:17:27.484 --> 00:17:35.275
for those of us that were early into it, we didn't go very far and we went,"Wait a minute," right?
00:17:35.375 --> 00:17:36.714
it's just cool, but it's...
00:17:37.495 --> 00:17:38.995
We better be really good at this.
00:17:39.015 --> 00:17:39.035
Yeah.
00:17:39.214 --> 00:17:41.495
And most of us got really good at it fast.
00:17:41.755 --> 00:17:42.005
Yeah.
00:17:42.825 --> 00:17:47.434
to, to some extent, I think that's still the, the question that needs to be solved next.
00:17:47.474 --> 00:17:53.755
how do we trust these systems to act reliably and repeatedly, so we know that we can give them more autonomy?
00:17:54.495 --> 00:17:56.185
The fact that they can do it, yeah, sure.
00:17:56.275 --> 00:17:58.535
I think that the technical capabilities are there.
00:17:59.154 --> 00:18:00.745
but does it work every time?
00:18:00.785 --> 00:18:02.144
What happens if it doesn't?
00:18:02.365 --> 00:18:04.335
How do you become aware of it?
00:18:04.485 --> 00:18:05.315
Can you jump in?
00:18:05.315 --> 00:18:06.664
Does another agent jump in?
00:18:06.695 --> 00:18:11.225
And, to your point, I think those are the architectural decisions that they need to be made.
00:18:11.965 --> 00:18:12.235
Yeah.
00:18:12.235 --> 00:18:14.215
So that's what, when you hear me talk about causal-
00:18:14.715 --> 00:18:14.845
Yeah
00:18:15.325 --> 00:18:16.875
causal is an architecture, right?
00:18:16.875 --> 00:18:21.505
It's not, it's, it's not, it's not anything else other than a reasoning architecture.
00:18:21.904 --> 00:18:21.924
Yeah.
00:18:21.935 --> 00:18:26.965
and so Pearl basically laid the foundation, a guy named Judea Pearl laid the foundation of it.
00:18:27.365 --> 00:18:31.634
And then so it's the application of how you apply thinking to something.
00:18:31.965 --> 00:18:35.295
And so that, that's, that gets you to the trust part of it.
00:18:35.924 --> 00:18:41.845
But it's not m- a lot of people would say, that means it's deterministic." Causal is not deterministic.
00:18:41.845 --> 00:18:44.785
It's, it, it's actually agentic in na- in nature, right?
00:18:44.785 --> 00:18:50.715
It's the same thing that, that people do to create agency, and that's what it's based on is the human agency model.
00:18:50.985 --> 00:18:54.455
It means that I understand what's true about the world around me.
00:18:54.644 --> 00:19:02.995
I understand what I can do to make things true, and then that's, that's how I build a hypothesis that if I do this, what if, right?
00:19:03.025 --> 00:19:09.279
And then I have the ability to understand the actions I took relative to the evidence of the outcome.
00:19:09.279 --> 00:19:09.585
Yeah.
00:19:09.585 --> 00:19:12.134
And then I learn from that evidence and I iterate forward.
00:19:12.144 --> 00:19:14.205
That's how I shape outcomes for the future.
00:19:14.215 --> 00:19:16.825
That's causal, but it's not deterministic.
00:19:17.495 --> 00:19:19.625
So what do you think, where is this going?
00:19:19.654 --> 00:19:22.545
Again, I'm coming back to the ads that, that we're seeing.
00:19:22.545 --> 00:19:31.144
You search for something on the web, maybe you still use Google to do that, and shortly after, ads start popping up for the same or similar products.
00:19:31.755 --> 00:19:35.894
So you might take an action because you see or you keep seeing these ads.
00:19:36.174 --> 00:19:37.817
I've been listening to a lot of music- Yeah.
00:19:37.817 --> 00:19:40.458
I think- on, on Spotify lately, I keep getting the same ads.
00:19:40.878 --> 00:19:43.837
and again, I know, but I'm not taking action.
00:19:43.858 --> 00:19:57.468
how is that changing with agents and how do you know that an agent acted on verifiable trusted information or because somebody, convinced it that's the best option for them that you should pursue?
00:19:58.057 --> 00:20:00.948
I, I really don't think ads go away, right?
00:20:01.087 --> 00:20:04.688
I think, I, I think you'll still see ads show up on TV.
00:20:04.688 --> 00:20:06.488
I think you'll still see them sh-show up on phone.
00:20:06.538 --> 00:20:15.758
But if you, if an ad piques your interest, I think what'll happen is that because, people still like to watch YouTube, people still like to watch TV.
00:20:15.758 --> 00:20:15.768
Sure.
00:20:15.768 --> 00:20:19.718
I still think there's room for ads to show up.
00:20:19.718 --> 00:20:32.468
But what I think what'll happen is that people will turn to their agent and say,"I'm interested in this.""Is this the best option for me?" And I think the agent will evaluate that.
00:20:32.817 --> 00:20:45.304
Now- They may say that's their preference, and what the ads will likely do is they will, w- they will actually try and bias you to inform your agent that's your preference, more or less.
00:20:45.364 --> 00:20:53.124
And I think you'll probably see that become more a tone in the ad- to get you to use your agent to inform preference.
00:20:53.244 --> 00:20:55.514
Likely that's what will become true.
00:20:55.964 --> 00:20:57.104
that's an interesting angle.
00:20:57.104 --> 00:21:04.564
so far I've been thinking about it, as, vendors trying to make it into Perplexity's, recommendations, right?
00:21:04.564 --> 00:21:13.334
or ranking or getting into OpenAI's ranking so whenever you ask a question, that's your ad that's being shown, and that's the level of influence.
00:21:13.554 --> 00:21:23.234
Plus, I feel a lot of ads that we listen to or, or that we see, they are, they're geared towards triggering some kind of em- emotional reaction.
00:21:23.284 --> 00:21:38.173
There's a need, there's a paucity or an absence of something, that- that you're feeling or that you didn't know you were feeling, but now after watching it, you are, craving ice cream or fast food, or you see this new car or what have you.
00:21:39.134 --> 00:21:43.964
But there's usually some em- emotional component that, vendors want you to act upon.
00:21:44.384 --> 00:21:47.714
And, and- Now, now with agents, that, that emotional component goes away.
00:21:47.714 --> 00:21:55.354
it's in, in theory, factual and objective, but what are the ways in which your agent will be influenced to take an action?
00:21:55.403 --> 00:21:57.264
so, so let's play that out for a second, right?
00:21:57.264 --> 00:21:57.443
Yeah.
00:21:57.864 --> 00:22:02.584
so who's, so who's gonna build the, who's gonna build the reasoning agents at the edge, right?
00:22:03.044 --> 00:22:13.634
and then who's gonna have the greatest amount of influence over the ability of, so today we have this commerce engine, whether it's Google or Amazon or whatever, right?
00:22:13.943 --> 00:22:23.703
And then the reasoning agents at the edge, who's gonna have the greatest amount of influence over you that basically connects the ad space, right?
00:22:23.714 --> 00:22:36.153
whatever platform you're watching it on, and then that ties to the reasoning agent, and then the l- the last closure loop would be n- probably not the, ubiquitous connectivity.
00:22:36.324 --> 00:22:37.564
It's the architecture of trust.
00:22:37.574 --> 00:22:45.834
and so whoever closes the triad of my ability to influence the reasoning agent, right?
00:22:45.834 --> 00:22:45.974
Yep.
00:22:45.994 --> 00:22:53.163
And the platform of influence, which is the optics that you're watching things through- the media influence, right?
00:22:53.304 --> 00:22:55.634
And that architecture of trust that we talk about.
00:22:55.644 --> 00:23:02.574
If you can close that triad, now you got something that's fairly controllable, right?
00:23:02.574 --> 00:23:02.844
Yes.
00:23:03.744 --> 00:23:04.384
Perfect.
00:23:04.443 --> 00:23:06.114
who ha- who has the power to do that?
00:23:06.114 --> 00:23:07.264
That's a g- that's a good question.
00:23:07.614 --> 00:23:14.274
I think the ones that, that, that have been, called dead, the ones that have been doing it for a long time.
00:23:14.774 --> 00:23:16.724
So the likes of Google, right?
00:23:16.724 --> 00:23:17.314
For example.
00:23:17.324 --> 00:23:17.344
Yeah.
00:23:17.384 --> 00:23:19.144
I think Google is, is not dead,
00:23:19.254 --> 00:23:20.374
Oh, no, no way.
00:23:20.423 --> 00:23:21.564
No way, shape, or form.
00:23:21.574 --> 00:23:21.604
Sure.
00:23:21.614 --> 00:23:24.324
and I think you've, you got the reach to do it too, right?
00:23:24.354 --> 00:23:24.714
It's-
00:23:24.754 --> 00:23:24.953
Yeah
00:23:25.124 --> 00:23:26.734
it, I think it's, that's likely
00:23:27.163 --> 00:23:27.413
Yeah.
00:23:28.364 --> 00:23:36.203
So then it becomes more of an evolution of how to, advertise, search, influence, in- It's orchestration, right?
00:23:36.214 --> 00:23:36.673
that change, right?
00:23:36.903 --> 00:23:37.094
Yeah.
00:23:37.604 --> 00:23:38.464
So fascinating.
00:23:39.274 --> 00:23:46.663
Michael, we've covered a lot of ground, over the last 27 minutes of what changes when your customer all of a sudden is an agent.
00:23:47.403 --> 00:23:55.604
I was wondering if there are any, key takeaways, maybe two, three key takeaways that, that you'd like to take the audience away from our conversation today before we wrap up.
00:23:56.294 --> 00:24:02.294
I think the major key takeaway is this collapse of distance between people and things.
00:24:02.504 --> 00:24:14.084
Is that the reality of our lives is that, that this confluence of technology and is really pointing back to people, right?
00:24:14.134 --> 00:24:21.574
And so all the confluence, so whether, whether it's biomedical, whatever, everything's pointing back to people.
00:24:21.903 --> 00:24:25.534
AI is actually pointing back to people, right?
00:24:25.874 --> 00:24:28.364
And so it's gonna reduce the separation.
00:24:28.364 --> 00:24:36.034
It's not, it's not pulling people apart, it's actually pulling people back closer to everything in your li- in your lives, right?
00:24:36.474 --> 00:24:41.014
And it's separating you by one degree from everything and everybody else, right?
00:24:41.074 --> 00:24:44.804
If you choose for it to be, right?
00:24:44.804 --> 00:24:53.014
And so you gotta ch- you gotta make this choice to, to be more engaged'cause these tools are massively powerful.
00:24:53.834 --> 00:25:09.724
and I think the ability to touch everybody else on the planet that's connected to this, everything else- on the planet that's connected to this, all the biomedical that's connect- that's gonna be connected to this, it is wonderfully exciting.
00:25:10.274 --> 00:25:15.084
But everything's pointed back at the individual, and that's as it should be
00:25:15.584 --> 00:25:18.544
So that means I need to update, my, my phrase, right?
00:25:18.544 --> 00:25:19.834
and my knowledge.
00:25:19.844 --> 00:25:23.183
It's not the six degrees of s- of Kevin Bacon, but the one degree
00:25:23.403 --> 00:25:24.514
It's the one degree.
00:25:24.964 --> 00:25:25.314
Yeah.
00:25:25.494 --> 00:25:25.774
Yeah.
00:25:25.844 --> 00:25:27.044
it's w- liberating.
00:25:27.534 --> 00:25:35.173
l- yes, for sure, and super exciting, especially if you get hands-on and you explore how this works and, think about where this might go eventually.
00:25:35.443 --> 00:25:35.604
Yeah.
00:25:35.984 --> 00:25:37.634
Will there be any downtime for us?
00:25:37.663 --> 00:25:47.514
Because you and I work more than we ever have now, even though, even a- if it said AI, AI would, A- AI would reduce the need for us, it surely hasn't.
00:25:48.054 --> 00:25:50.974
So to me, that's where the human component comes in, to say- Yeah,
00:25:51.153 --> 00:25:51.554
absolutely
00:25:51.834 --> 00:25:57.734
I need a break, and I need to take a break, and I want to take a break." and, the world is still turning when I come back to it.
00:25:57.964 --> 00:25:58.754
Yes, it is.
00:25:59.044 --> 00:25:59.814
So wonderful.
00:25:59.834 --> 00:26:01.614
Michael, it's been a pleasure having you on the show.
00:26:01.614 --> 00:26:09.344
Thank you for sharing your expertise with us, and I really appreciate this, unscripted format and continuing our conversation that started a couple weeks ago in Washington.
00:26:09.413 --> 00:26:10.494
This has been fabulous.
00:26:10.494 --> 00:26:11.913
I thank you so very much.
