WEBVTT
00:00:01.439 --> 00:00:10.080
Whether you're newer to AI or if you consider yourself an AI expert, we're all probably thinking the same thing.
00:00:11.119 --> 00:00:15.039
AI agents, AI agents, AI agents.
00:00:15.599 --> 00:00:26.640
That's because, well, we've been promised that AI agents would take over, maybe in a good way, maybe in a bad way, yet that promise never really came to fruition.
00:00:26.719 --> 00:00:36.000
I mean, we heard it in the early Chat GPT days, 2023, 2024, 2025 was supposed to be the year of the agents, but it wasn't.
00:00:36.079 --> 00:00:37.200
But guess what?
00:00:37.840 --> 00:00:38.799
We're there.
00:00:39.039 --> 00:00:55.119
I think we've actually arrived with some of the recent AI updates from the big companies in Anthropic, OpenAI, Google, Microsoft, even crazy viral open source projects like OpenClaw.
00:00:55.200 --> 00:01:02.719
I think now the long vision that we've been pitched of AI agents is here, but it's all happening at once.
00:01:02.960 --> 00:01:08.000
Like I said, I've been doing this everyday AI thing for three years.
00:01:08.239 --> 00:01:19.680
And at least when it comes to AI agent development, the last month has been as eventful as the other 2.9 years combined.
00:01:20.560 --> 00:01:26.000
Don't worry, even if you are brand new or if you're an expert.
00:01:26.480 --> 00:01:28.480
That's what we're going to tackle on today's show.
00:01:28.640 --> 00:01:37.680
We're going to hopefully get you caught up, tackle AI agents from all angles, and do so in a hopefully quick, beginner-friendly way.
00:01:38.000 --> 00:01:47.359
So this is the Start Here series, and we're going to be going over AI agents in 2026, what they are, and when you should use them.
00:01:47.519 --> 00:01:48.000
All right.
00:01:48.159 --> 00:01:49.280
I'm excited for this one.
00:01:49.359 --> 00:01:50.400
I hope you are too.
00:01:50.640 --> 00:01:51.359
What's going on?
00:01:51.599 --> 00:01:53.280
Welcome to the Start Here series.
00:01:53.359 --> 00:02:04.319
My name is Jordan Wilson, and the Start Here series, well, I created it because after 700 plus episodes of the Everyday AI podcast, I didn't have a good answer for the most common question I got.
00:02:04.480 --> 00:02:08.400
People say, hey, Jordan, looks like a lot of great information here on your podcast.
00:02:08.479 --> 00:02:09.439
Where do I start?
00:02:09.680 --> 00:02:12.719
Well, you start here with the Start Here series.
00:02:12.879 --> 00:02:15.840
And this is volume eight in the Start Here series.
00:02:15.919 --> 00:02:23.120
But this is an essential podcast series to both learn the AI basics and to double down on your AI knowledge.
00:02:23.439 --> 00:02:29.199
And the way that you should really do that is make sure you go to starthearseries.com.
00:02:29.360 --> 00:02:35.919
That is going to give you free access to our inner circle community in our Start Here Series space.
00:02:36.000 --> 00:02:46.719
So you can go catch up on all of everything in this series in one place, a bunch of additional resources that we've thrown together, all available for you right there.
00:02:46.879 --> 00:02:51.120
So uh like I said, whether you're brand new, an expert, it doesn't matter.
00:02:51.280 --> 00:02:57.680
Go join now, like a thousand people that are already in the community and listen to all of the Start Here series.
00:02:57.840 --> 00:03:00.240
Also, I have to put this out there because it's important.
00:03:00.479 --> 00:03:04.560
Make sure you also go listen to episodes 712 and 713.
00:03:04.639 --> 00:03:08.159
That is our 2026 AI predictions and roadmap series.
00:03:08.319 --> 00:03:14.479
You know, 26 kind of bold predictions on AI, and at least a third of them are about agents.
00:03:14.560 --> 00:03:14.879
All right.
00:03:15.039 --> 00:03:25.759
So uh, and if you missed our last start here series, that was volume seven, where we went over context engineering, how to get expert level outputs from AI chatbots.
00:03:25.919 --> 00:03:30.879
But let's get straight into AI agents, kind of the state of AI agents here in 2026.
00:03:30.960 --> 00:03:33.759
So here's what we're gonna be going over on today's show.
00:03:33.919 --> 00:03:43.039
We're gonna talk about what AI agents actually are and how they differ from uh how they're different from traditional AI chatbots or AI powered automation.
00:03:43.199 --> 00:03:56.240
Uh, I'm gonna give you a decision framework for when agents are the answer and when they're worth it versus when they're just overkill and when you should say, hey, that should just be a human task or an AI, uh, you know, large language model task.
00:03:56.400 --> 00:04:04.319
And then I'm gonna give you the practical path to safely starting your first agent pilot, not even this year, like this week.
00:04:04.560 --> 00:04:05.199
All right.
00:04:06.879 --> 00:04:09.439
Let's get the explanations out of the way.
00:04:09.680 --> 00:04:10.000
All right.
00:04:10.080 --> 00:04:17.120
So, like I said, if you are, if you've been building AI agents for uh, you know, a decade or two, you can just skip ahead a couple of minutes.
00:04:17.279 --> 00:04:22.639
Actually, speaking of that, and why I think it's so important to look at this from a foundational level.
00:04:22.720 --> 00:04:27.519
Uh, so uh about two weeks ago, had a fantastic guest uh on my podcast.
00:04:27.600 --> 00:04:32.480
She's the head of uh Microsoft Research, he's been working in AI agents for 20 years.
00:04:32.560 --> 00:04:41.040
And one thing we talked about both on the show and before and after is the definition of an AI agent is constantly changing.
00:04:41.199 --> 00:04:41.759
Uh, right.
00:04:41.839 --> 00:04:47.839
So even what is an agent and what's not an agent, uh, what can an agent do, what can't it do?
00:04:48.000 --> 00:04:48.959
It's always changing.
00:04:49.040 --> 00:04:49.360
All right.
00:04:49.439 --> 00:04:55.600
So if you're listening to this in you know February of 2026, that's great.
00:04:55.759 --> 00:04:59.439
All this information is probably gonna be extremely fresh and accurate.
00:04:59.600 --> 00:05:06.079
If you're listening to this in June or in 2027, uh right, some of this might be a little old.
00:05:06.160 --> 00:05:06.720
So keep that in mind.
00:05:06.800 --> 00:05:14.319
But I'm gonna give you at least as of February uh, you know, 19th, 2026, this is the realest most up-to-date.
00:05:14.480 --> 00:05:16.000
So, what is an AI agent?
00:05:16.160 --> 00:05:17.199
In very simple terms.
00:05:17.360 --> 00:05:24.560
It's kind of like an AI chatbot, but it has tools and it has permissions and it has a goal and it makes its own decisions.
00:05:24.720 --> 00:05:30.879
And sometimes it might start off on one path and say this isn't right, and it'll go backwards and start down another path.
00:05:31.040 --> 00:05:45.120
So they can plan ahead, take steps, use tools, uh, take actions, check results, and with and retry within the guardrails, whatever those guardrails are, whether you set them up or they're set up by a certain uh you know piece of software that you're using.
00:05:45.279 --> 00:05:46.720
So, what do agents do?
00:05:46.879 --> 00:05:52.000
Well, they can take on delegate work, they can delegate work to other sub-agents.
00:05:52.399 --> 00:05:54.079
They don't just answer questions, right?
00:05:54.160 --> 00:05:59.439
So they can draft, route, research, navigate apps, and even compile multi-step tasks across different systems.
00:05:59.680 --> 00:06:03.199
And here's this state of AI agents and in the enterprise.
00:06:03.279 --> 00:06:10.079
Uh, right now, Gartner study showed that 80% of Fortune 500s have active agents uh right now.
00:06:10.240 --> 00:06:12.399
Sorry, that was from uh Microsoft.
00:06:12.639 --> 00:06:20.720
Um, but uh Gartner said that 40% of enterprise apps are projected to include them by 2026.
00:06:20.959 --> 00:06:27.920
So, yeah, even since that story or sorry, that uh report came out from Gartner last summer, uh, I think that has changed, right?
00:06:28.079 --> 00:06:40.959
That's one of the things that's difficult to give you know people a thumbprint on kind of the state of AI agents, is because even the studies that are the most up to date are usually looking in the rear view mirror a couple of quarters or more, uh, right.
00:06:41.040 --> 00:06:48.480
But the reality is an overwhelming majority of Fortune 500 companies have agents in production, all right.
00:06:48.639 --> 00:06:51.360
And all the software that you use, anyways, right?
00:06:51.439 --> 00:06:54.959
Because you might think, oh, you know, an agent means building something on your own.
00:06:55.120 --> 00:06:56.399
No, it doesn't, uh, right.
00:06:56.480 --> 00:07:05.920
And it's not just the big four uh in Microsoft, Google, OpenAI, OpenAI, and Anthropic, uh, right, that would give you kind of agents out there in the workforce.
00:07:06.000 --> 00:07:09.680
There's literally thousands of different now pieces of software.
00:07:09.759 --> 00:07:16.720
And that's where this you know Gartner stack comes from, that 40% of enterprise apps are projected to include AI agents this year, right?
00:07:16.800 --> 00:07:18.240
What that means, a good example.
00:07:18.319 --> 00:07:22.959
I've I've talked about, I've used uh, you know, ClickUp, a project management tool uh in the past, right?
00:07:23.040 --> 00:07:26.079
So now like ClickUp has you know agents by default.
00:07:26.319 --> 00:07:34.240
I think you know, things that used to be considered just features or tools in daily software are now agents, right?
00:07:34.319 --> 00:07:37.680
So if you're using you know Salesforce, they have agents.
00:07:37.839 --> 00:07:41.759
If you're using uh, you know, Slack, Slackbot, you know, they have agents.
00:07:41.920 --> 00:07:44.079
If you're using HubSpot, they have agents, right?
00:07:44.160 --> 00:07:56.399
So uh it's not just you know building an agent from scratch or you know, using one of the tools from the big four or these you know super general agents like you know Meta uh Meta's MANIS agent, that's weird to say now, right?
00:07:56.480 --> 00:08:02.399
Manaus that was uh acquired by Meta or Gen Spark or some of these you know general agents, right?
00:08:02.560 --> 00:08:04.079
They're everywhere.
00:08:05.360 --> 00:08:07.600
So what is it actually, right?
00:08:07.680 --> 00:08:10.879
So I'm gonna break it down in the most simplest, hopefully, language possible.
00:08:11.199 --> 00:08:18.240
An AI agent is an autonomous system that can plan, execute, and self-correct across different tools and systems.
00:08:18.399 --> 00:08:21.439
So it's kind of like a human at a computer, uh, right?
00:08:21.680 --> 00:08:24.720
So that can make decisions and has access to things.
00:08:24.879 --> 00:08:27.439
That's the way I like to think of an agent, right?
00:08:27.680 --> 00:08:37.679
It's the same thing as if a you know an entry-level worker sat down at uh a computer at your company that had access to everything, right?
00:08:38.080 --> 00:08:41.200
So think of that, but you can have a million of them.
00:08:42.240 --> 00:08:43.519
So what feelings do you get?
00:08:43.600 --> 00:08:48.480
Is that excitement, or are you thinking of security and risk and things that can go wrong?
00:08:48.639 --> 00:08:53.679
Well, I think when you think about agents and uh their implications, you have to think of both sides of that coin.
00:08:53.759 --> 00:08:55.759
But that's essentially what an agent is, right?
00:08:55.919 --> 00:09:02.879
A full agent is like sitting a human down at a computer that has access to everything, right?
00:09:03.120 --> 00:09:22.799
It can go get you new clients and close deals, uh, can complete work autonomously, spreadsheets, PowerPoints, uh, right, jumping between different apps, it you know, has access to a terminal, uh, you know, you know, a sandbox, a virtual browser, all these different things, like a human sitting down in front of a computer uh does.
00:09:23.120 --> 00:09:32.960
Um, so essentially, if it can't take action, if it can't adapt to context, if it can't solve new problems independently, it's not technically an agent.
00:09:33.039 --> 00:09:36.480
And I'm gonna go over some of the you know other things that it might be.
00:09:37.279 --> 00:09:40.159
Uh, and that's because I've talked about this on the show before.
00:09:40.240 --> 00:09:45.759
Uh, a great, a great study from Gartner uh talked about this concept of agent washing, right?
00:09:45.840 --> 00:09:55.200
So they looked at um thousands of different vendors uh in 2025 that were essentially marketing or offering or selling agentic capabilities.
00:09:55.360 --> 00:10:01.679
And what they found, only 130 of thousands were actually agents.
00:10:01.759 --> 00:10:03.360
It's agent washing, right?
00:10:03.440 --> 00:10:15.360
Most are just using you know chatbots or uh, you know, scripted uh rigid workflows and sprinkling some you know, quote unquote agentic models somewhere in the process and saying it's an agent.
00:10:16.159 --> 00:10:30.080
I tell you this, even looking at my own inbox, right, for everyday AI, because uh we get pitched uh more than a thousand times a year by companies, and you know, obviously recently it's just been all agents, right?
00:10:30.159 --> 00:10:32.720
Oh, check out our agent, it's the best one in the world, right?
00:10:32.879 --> 00:10:37.679
And most of the time I don't look because it's I don't think most of the companies are gonna make it very long.
00:10:37.759 --> 00:10:40.639
But then when I do look, I'm like, this isn't an agent, right?
00:10:40.879 --> 00:10:44.399
This is a you know, like a like a workflow.
00:10:44.559 --> 00:10:49.120
This is uh RPA with a little bit of a large language model sprinkled on top.
00:10:49.200 --> 00:10:50.399
That's not an agent, right?
00:10:50.559 --> 00:10:54.879
An agent, you don't give it a dedicated workflow.
00:10:54.960 --> 00:11:01.360
You don't give it, you know, if then or if-else, uh like like statements to follow.
00:11:01.519 --> 00:11:06.240
An agent is sitting a computer, like sitting a human down in front of a computer.
00:11:06.399 --> 00:11:09.360
They have free will, make decisions, make good ones, make bad ones.
00:11:09.440 --> 00:11:13.440
It really depends on, well, the guardrails uh that you set up around those agents.
00:11:13.519 --> 00:11:22.480
And right now, uh 88% of executives are investing in agentic AI, but most can't even tell what's real from fake.
00:11:23.440 --> 00:11:26.639
So let me quickly talk about the difference, right?
00:11:26.799 --> 00:11:34.240
Because I said, chances are, even if you're paying for an AI agent product, it's just marketing, it's smoke and mirrors.
00:11:34.320 --> 00:11:40.000
Because, like I said, uh, according to that Gartner study, less than 5% are actually AI agents.
00:11:40.159 --> 00:11:42.799
So let's first talk about AI-powered workflows, right?
00:11:42.960 --> 00:11:45.440
And there's nothing wrong with these other things, FYI, right?
00:11:45.519 --> 00:11:53.519
But a great example of that is like N8N, uh, right, or make.com, or you know, in some instances, like Zapier as well, right?
00:11:53.679 --> 00:11:56.399
I love, I love these tools, I love these platforms, right?
00:11:56.480 --> 00:12:04.720
I'm not saying that that they're bad, uh, right, but if you have all these decision trees and you're connecting these nodes, it's not an agent, right?
00:12:04.799 --> 00:12:06.559
It's an AI-powered workflow.
00:12:06.720 --> 00:12:07.919
Nothing wrong with it.
00:12:08.080 --> 00:12:08.480
All right.
00:12:08.639 --> 00:12:10.240
We're just getting definitions here.
00:12:10.399 --> 00:12:12.799
All right, then you have agentic models.
00:12:12.960 --> 00:12:14.159
Well, what the heck is that?
00:12:14.320 --> 00:12:18.480
Well, these agentic models are essentially turning into mini agents.
00:12:18.639 --> 00:12:29.840
Uh, and they're technically, I think, in my opinion, more agentic than AI-powered workflows because they can make more free will decisions on their own and also in the same time do more harm, right?
00:12:29.919 --> 00:12:35.840
But an agentic model, well, that's just our you know, uh, frontier AI models, right?
00:12:36.159 --> 00:12:37.360
Think of the big four models.
00:12:37.440 --> 00:12:42.320
Well, technically the big three uh that are at least creating and uh you know putting their models out there.
00:12:42.399 --> 00:12:45.600
You could throw in, you know, XAIs, Grok if you want to.
00:12:45.759 --> 00:12:58.879
Uh, I wouldn't necessarily, but let's look at three models Google Gemini 3 Pro, uh, Claude Opus 4.6, uh, and uh OpenAIs GPT 5.2 Pro, right?
00:12:58.960 --> 00:13:01.919
Or uh 5.3 Codex, whatever you want.
00:13:02.080 --> 00:13:09.679
Um, and obviously, if you're listening to this, you know, at some point later in the year, it you know, just replace it with uh, you know, today's three models.
00:13:09.919 --> 00:13:13.759
These models by default are agentic and they didn't used to be, right?
00:13:13.919 --> 00:13:23.759
Uh a year-ish ago, models were not agentic, they were uh extremely smart autocomplete token munchers, right?
00:13:24.000 --> 00:13:26.080
Rom, round, round tokens here, spit it out, right?
00:13:26.240 --> 00:13:27.440
Now they're not like that.
00:13:27.679 --> 00:13:29.360
Models are agentic by default.
00:13:29.440 --> 00:13:32.159
That means you can give them a lot of context.
00:13:32.320 --> 00:13:37.919
You can give them files, you can give them problems, you can even, in the course of a chat bot, right?
00:13:38.080 --> 00:13:49.759
A front-end chatbot, you can technically, if you know what you're doing and give it enough context and you know go back and forth with it, it can technically have a very agent-esque workflow, right?
00:13:49.840 --> 00:14:06.080
It's a little more confined because you're working uh within the constraints of a certain AI chatbot like ChatGBT, but I'm talking front-end, front-end AI models technically are very agentic because they can decide on their own how they're going to solve a problem.
00:14:06.240 --> 00:14:11.360
They have tools that you probably don't even know that they have that they will call on their own, right?
00:14:11.519 --> 00:14:13.360
They have virtual browsers.
00:14:13.519 --> 00:14:21.039
Um, you know, they have kind of their own uh sandboxes where they can write and run code, right?
00:14:21.279 --> 00:14:25.039
Crazy thing I was doing uh recently, you know, using Codex.
00:14:25.440 --> 00:14:26.960
Uh, right, very good.
00:14:27.039 --> 00:14:30.799
But I mean, technically, I think OpenAI is investing more and more into Codex.
00:14:30.960 --> 00:14:34.799
Um, you know, it's technically a coding model, but it's a front-end model, right?
00:14:34.879 --> 00:14:36.879
Um, you don't have to do anything on the back end.
00:14:37.039 --> 00:14:46.879
I was using it and Codex on its own decided it was going to download a six gigabyte open source model and it used it in its own platform.
00:14:47.200 --> 00:14:48.960
That is crazy, right?
00:14:49.120 --> 00:14:54.639
It noticed that it didn't have the capabilities to do something at scale that I asked it to.
00:14:54.799 --> 00:14:58.399
And then over the course, overnight, it worked for almost 10 hours.
00:14:58.639 --> 00:15:12.480
It downloaded an open source model, and then essentially I had uh this one was uh had like 3,000 screenshots, and it wasn't able to uh you know get the results it needed to at scale, at speed with its own computer vision model.
00:15:12.559 --> 00:15:18.960
So it said, All right, I'm gonna download something and a six gigabyte model, and it ran it through there, right?
00:15:19.039 --> 00:15:20.480
That's an agentic model.
00:15:20.639 --> 00:15:24.879
It made decisions, it called tools, it found solutions, right?
00:15:25.120 --> 00:15:26.639
Then you have passive agents.
00:15:26.720 --> 00:15:28.799
All right, sorry, I went on a little uh side tangent there.
00:15:28.879 --> 00:15:33.600
So number one, AI powered workflows, number two, agentic models, and then number three, passive agents.
00:15:33.759 --> 00:15:41.519
So these are agents that are not necessarily uh autonomous, uh, right, but same thing, an example of that, right?
00:15:41.679 --> 00:15:44.480
Chat GPT has a great agent mode.
00:15:44.720 --> 00:15:48.399
I can schedule that agent to run at 9 a.m., right?
00:15:48.480 --> 00:15:49.679
But I still have to trigger it.
00:15:49.840 --> 00:15:52.000
I still have to go say, hey, agent, right?
00:15:52.080 --> 00:15:59.759
And I can train it and give it, you know, its its role and its its context and all this stuff, but I still have to go push the button, all right?
00:15:59.919 --> 00:16:03.840
And then you have autonomous agents or always on agents, uh, right.
00:16:03.919 --> 00:16:08.960
And then these are taking, uh, especially recently, a lot of new uh shape and form.
00:16:09.519 --> 00:16:13.360
But I mean, an always on agent or an autonomous agent is essentially that.
00:16:13.519 --> 00:16:15.679
It's working around the clock, right?
00:16:15.759 --> 00:16:17.600
And it's not just triggered by you.
00:16:17.759 --> 00:16:23.120
It can, in theory, be triggered by a certain action, like, oh, when you uh receive an email, right?
00:16:23.200 --> 00:16:27.360
It's going to automatically do A, B, and C and then go after a certain goal.
00:16:27.440 --> 00:16:28.000
Uh right.
00:16:28.240 --> 00:16:36.240
So even the the lines, you know, kind of going over uh these four different uh classifications, they're starting to blur as well.
00:16:37.200 --> 00:16:43.679
Uh and the risks, just like the capabilities, are growing just as fast, right?
00:16:43.840 --> 00:16:46.000
Uh, some some uh good stats here.
00:16:46.159 --> 00:16:55.440
So uh this is from Cybersecurity Insiders said that 75% of CISOs have discovered unsanctioned AI tools already running in their environments, right?
00:16:55.600 --> 00:16:57.840
So that's that's crazy, right?
00:16:57.919 --> 00:17:06.960
When we talk about shadow, uh shadow AI, it's no longer just oh, someone's using uh you know chat GPT maybe when they shouldn't.
00:17:07.200 --> 00:17:07.519
No.
00:17:08.000 --> 00:17:13.279
What about if someone's building agents that have access to all of your data?
00:17:13.359 --> 00:17:21.599
It's happening, and then those agents are making decisions potentially without any or uh you know company approved guardrails.
00:17:21.680 --> 00:17:27.920
Uh and and then you when you talk about multi-agent systems, that's when there's huge new risks, right?
00:17:28.079 --> 00:17:42.160
That you it's it's really hard to project because as agents' capabilities are changing, humans necessarily, even the humans building them, don't fully understand the risk profile of these multi-agent systems, right?
00:17:42.319 --> 00:17:45.839
So a very simple example, uh, right.
00:17:46.640 --> 00:17:50.799
I didn't know that codex was going to download an open source model.
00:17:50.880 --> 00:18:03.759
So what happens, right, when we set up guardrails, um, you know, on maybe a certain multi-agent system, and then the agent gets creative, it finds loopholes or gray area, right?
00:18:03.920 --> 00:18:13.599
And it says, hey, well, it said I can't set up, you know, my guardrails say that I can't use other agents, but it doesn't say I can't use, you know, AI-powered workflows.
00:18:13.759 --> 00:18:20.880
So then it finds a loophole, and all of a sudden you have these agents that are setting up something that is uh, you know, unregulated.
00:18:21.519 --> 00:18:23.119
Agents will do that.
00:18:23.440 --> 00:18:28.400
Agents by default are made to be quote unquote helpful assistants, just like large language models.
00:18:28.559 --> 00:18:39.839
So if they think that it's helpful, they're not going to go through a morality check first, uh, or uh a company ethics check unless you really have that hard baked in there.
00:18:40.000 --> 00:18:42.799
They're gonna go off and make trouble, probably.
00:18:42.960 --> 00:18:45.519
They're gonna break things while getting something done.
00:18:45.680 --> 00:18:52.480
You know, sometimes I think of agents, uh, it's it's it's like giving a giving a toddler uh a task, right?
00:18:52.720 --> 00:18:58.559
Hey, hey, child, hey, three-year-old, uh, you know, go go put this toy away.
00:18:58.799 --> 00:18:59.359
All right.
00:18:59.599 --> 00:19:01.119
Well, did they put the toy away?
00:19:01.279 --> 00:19:01.680
Sure.
00:19:01.920 --> 00:19:03.279
Did they knock over a lamp?
00:19:03.519 --> 00:19:04.240
Maybe.
00:19:04.559 --> 00:19:08.960
Uh did they smear marker uh on the new couch?
00:19:09.279 --> 00:19:11.200
Potentially, right?
00:19:11.440 --> 00:19:12.720
It's the same thing.
00:19:12.799 --> 00:19:24.640
Uh, you know, and right now, uh another um report, the 2026 CISO AI risk report said that 92% of organizations right now lack full visibility into their AI identities.
00:19:24.799 --> 00:19:30.000
And again, as those AI identities and capabilities are growing, the risk is going crazy.
00:19:30.240 --> 00:19:31.839
So, how do we get here?
00:19:32.240 --> 00:19:32.559
Right?
00:19:32.720 --> 00:19:34.319
Let's take a very quick walk.
00:19:34.400 --> 00:19:39.839
Uh, so 2022, you know, AI was, you know, hey, type a question, get an answer.
00:19:40.079 --> 00:19:51.920
And then I think we had two years, you know, uh in 2023 and 2024, where, you know, large language models, oh, they can use tools now and you can upload files and they can, you know, browse the web, and that's cool and it's more helpful.
00:19:52.079 --> 00:20:00.960
But then I think toward the end of 2025, and obviously this year, now this is where these reasoning models are agentic by default and their capabilities are growing by the day.
00:20:01.039 --> 00:20:05.839
And like I said, the flip side, the unknown risks are also growing.
00:20:05.920 --> 00:20:12.480
But now reasoning models and in reinforcement uh learning just let agents actually act, right?
00:20:12.640 --> 00:20:15.680
And even if we talk about things like OpenClaw, right?
00:20:15.759 --> 00:20:24.960
Uh, which OpenAI uh acquired, uh, you know, these open source agents, people are just giving them computers, they're handing over their complete lives, right?
00:20:25.119 --> 00:20:29.200
Hopefully, maybe those are more entrepreneurs, business owners, solopreneurs, right?
00:20:29.279 --> 00:20:30.960
But people are still doing this.
00:20:31.200 --> 00:20:35.119
Uh, they're giving AI agents uh access to everything.
00:20:35.359 --> 00:20:45.599
And we don't yet know their capabilities, but it's quickly gone from AI is just a chatbot that you talk with, to oh, now all of a sudden that chatbot has access to tools in the internet.
00:20:45.839 --> 00:20:54.000
And oh, now all of a sudden, you know, uh these agents are making decisions completely on their own and their capabilities uh are quick.
00:20:54.160 --> 00:20:58.799
So here's six different agent types that I think you'll encounter in 2026.
00:20:59.039 --> 00:21:07.519
Uh, you know, not every single agent out there, like I said, there's literally tens of thousands, technically, agents, because you can spin one up uh in a couple of minutes, uh, right?
00:21:07.599 --> 00:21:12.160
But you have task agents, these are agents that draft, summarize, create assets, et cetera.
00:21:12.319 --> 00:21:19.039
You have decision-supported agents, these are ones that compare options, surface trade-offs, and flag risks for review.
00:21:19.200 --> 00:21:24.319
Then you have process agents, these are ones that route work across tools, gather inputs, and prep the next step.
00:21:24.480 --> 00:21:29.279
You have computer use agents, these can navigate websites and apps to complete multi-step tasks.
00:21:29.519 --> 00:21:31.200
You have multi-agent systems, right?
00:21:31.279 --> 00:21:35.759
When agents can spin up sub-agents, you know, send a bunch of agents out on their own, right?
00:21:35.920 --> 00:21:44.559
Clawed code, uh, you know, from Anthropic recently uh with their Opus 4.6, uh, kind of, you know, have popularized that on the consumer and the mainstream.
00:21:44.640 --> 00:21:46.960
And then you also have commerce agents, right?
00:21:47.039 --> 00:21:50.319
Uh, where agents can talk to each other, uh, right?
00:21:50.559 --> 00:21:53.920
You know, agents are now having their own version of the web, right?
00:21:54.079 --> 00:21:59.119
There's going to be uh agentic transactions uh that really don't involve humans at all.
00:21:59.200 --> 00:22:02.400
So even the Type of agents changing quickly, right?
00:22:02.559 --> 00:22:03.519
Talk to me next week.
00:22:03.680 --> 00:22:06.000
There's probably going to be new categories.
00:22:06.480 --> 00:22:08.319
So when do you actually need one?
00:22:08.640 --> 00:22:11.440
When do you need an agent versus when should I go do this task myself?
00:22:11.599 --> 00:22:18.640
Versus when do I need an AI powered workflow versus when should I just use ChatGPT or Gemini or Claude or Copilot versus a full autonomous agent?
00:22:18.720 --> 00:22:25.200
Well, I think if you can write a clear checklist and it rarely changes, a workflow can handle it fine.
00:22:25.440 --> 00:22:25.759
Right?
00:22:25.920 --> 00:22:27.519
AI powered workflow.
00:22:27.759 --> 00:22:38.799
If the steps vary, if it requires judgment or if it depends on multiple tools, that's probably when you're starting to get into either agent territory or an agentic model that has access to your data.
00:22:39.039 --> 00:22:46.000
But here's the thing: if the underlying process is broken before the AI touches it, an agent is not going to be the answer.
00:22:46.160 --> 00:22:58.400
I think a lot of people, well, by a lot of people, I say the overwhelming majority of people are just looking at an AI agent as a shortcut to do things that humans couldn't or to do things that they just don't want to do.
00:22:58.640 --> 00:23:00.079
It's not going to work like that.
00:23:00.319 --> 00:23:08.000
If you stick an agent on a broken or an antiquated workflow, you're just asking for a compounding disaster, right?
00:23:08.079 --> 00:23:12.079
If you're like, oh, well, our humans aren't doing a good job at this, let's give it to an agent.
00:23:12.160 --> 00:23:15.599
Well, the agent is going to do a worse job than the humans that weren't doing a good job.
00:23:15.839 --> 00:23:22.960
So you're really only being applying an AI agent to processes that are already running very smoothly, that are well documented, right?
00:23:23.039 --> 00:23:26.400
You have SOPs, you have way to measure, you have ways to measure them.
00:23:26.559 --> 00:23:30.240
Otherwise, setting an agent out is just a recipe for disaster.
00:23:30.480 --> 00:23:38.880
And that's why another recent Gardner study said that 40% of agentic AI projects are expected to fail or be canceled by 2027.
00:23:39.039 --> 00:23:42.240
Well, that's one of the big reasons, is what I just told you.
00:23:42.400 --> 00:23:53.440
People are thinking of AI agents as duct tapes instead of being multipliers through processes that are already AI enabled and working with humans, right?
00:23:53.599 --> 00:23:58.000
Poor data quality, just like anything else, missing context, no evaluation loops.
00:23:58.079 --> 00:24:04.000
Uh, because I think success requires rethinking and redesigning processes, not just adding AI on top.
00:24:04.160 --> 00:24:08.319
You're probably tired of me saying, you know, you don't just upskill, reskill, right?
00:24:08.400 --> 00:24:09.440
The same thing with agents.
00:24:09.519 --> 00:24:11.680
You don't just upskill a workflow with an agent.
00:24:11.759 --> 00:24:13.920
You don't just upgrade a workflow with an agent.
00:24:14.079 --> 00:24:21.759
You have to deconstruct it, rebuild, you have to unlearn, and you build from scratch an agentic first workflow.
00:24:22.319 --> 00:24:25.119
So we'll talk about bounded autonomy.
00:24:25.359 --> 00:24:28.960
And that's well, it's how you start losing control if you're not doing this.
00:24:29.039 --> 00:24:31.279
So you should be starting when it comes to agents.
00:24:31.359 --> 00:24:35.200
Like I said, pick the right workflow that's already working, AI enabled with humans.
00:24:35.359 --> 00:24:39.119
It's documented, it's processed, uh, it's processing correctly.
00:24:39.279 --> 00:24:40.559
You can measure it, etc.
00:24:40.880 --> 00:24:46.240
So you should start at suggest only with an agent where the agent just drafts and you decide what happens next.
00:24:46.480 --> 00:24:53.839
Once it's gone through that process successfully, then you move on to execute with approval, where the agent acts after you one-click sign off.
00:24:54.240 --> 00:24:58.079
Then, again, then, and this is a big asterisk, right?
00:24:58.160 --> 00:24:59.279
Everyone's got to sign off.
00:24:59.440 --> 00:25:03.200
That's when you can start moving on to more autonomy, right?
00:25:03.359 --> 00:25:10.480
When you can scale if the guardrails are right, um, you know, with spending caps, permission rules, and audit trails.
00:25:10.720 --> 00:25:20.799
All right, but only then, right, would can you think of AI agents as this you know infinity, infinitely uh scalable uh asset for your company.
00:25:21.200 --> 00:25:23.599
So here's how you get started this week.
00:25:23.839 --> 00:25:26.960
Number one, you need to think about meetings to action items.
00:25:27.200 --> 00:25:31.440
Have an agent go draft owners and follow-ups and you approve before sending.
00:25:31.519 --> 00:25:33.200
Uh inbox triage, right?
00:25:33.359 --> 00:25:35.119
Another great thing for AI agents.
00:25:35.279 --> 00:25:36.960
Don't have them act, right?
00:25:37.039 --> 00:25:41.599
Uh have it triage your inbox or multiple pieces of software that you already use.
00:25:41.759 --> 00:25:43.839
Uh, or creating research briefs.
00:25:44.000 --> 00:25:48.319
An agent can build a briefing doc with sources and let you verify them before sharing.
00:25:48.480 --> 00:25:50.400
So start very simple.
00:25:50.559 --> 00:25:56.640
Don't let an agent do something that right now requires multiple humans and multiple approval processes.
00:25:56.880 --> 00:26:02.400
Start looking at maybe some of those time-consuming, difficult tasks that an individual human does.
00:26:02.480 --> 00:26:04.880
And then don't give the AI agent full autonomy.
00:26:05.039 --> 00:26:06.319
Go through steps.
00:26:06.960 --> 00:26:11.039
So here is the big takeaway as we wrap up.
00:26:11.680 --> 00:26:15.039
Agents are for delegation, not just answers.
00:26:15.200 --> 00:26:17.680
They plan, act, and report back.
00:26:17.759 --> 00:26:23.519
They can use tools at their own disposal, um, come up with new solutions that you didn't even think of.
00:26:23.680 --> 00:26:28.400
So that's why guardrails, traceability, and observability are paramount.
00:26:28.640 --> 00:26:35.680
Then you need to start with bounded autonomy and measure time saved, error rates, risk events, uh, you know, ROI.
00:26:36.079 --> 00:26:38.319
You need to constantly do those things.
00:26:38.480 --> 00:26:42.720
Um, you need to don't do human in the loop, right?
00:26:42.880 --> 00:26:49.279
It's one of the biggest mistakes I think companies are still making because human in the loop was this fun term in 2024 and everyone just said it.
00:26:49.440 --> 00:26:54.799
No, human in the loop is combustible in a bad way, right?
00:26:54.960 --> 00:26:59.279
People think, oh, we're gonna have a hundred agents and we're gonna have, you know, Bill and IT oversee them.
00:26:59.359 --> 00:27:06.480
No, you need expert-driven loops with cyclical improvement chains that are chains that are documented, right?
00:27:07.119 --> 00:27:11.200
Observability, traceability is a full-time job for a team.
00:27:11.359 --> 00:27:15.680
Go back and that's why I said go back and listen to the 2026 AI prediction and roadmap series.
00:27:15.839 --> 00:27:18.400
I talked about this in a lot more depth and told you how to get it done.
00:27:18.559 --> 00:27:29.839
And then last but not least, the future advantage comes from building agent ecosystems, not just picking the right agent or you know, dipping your toe in the AI agent of the week.
00:27:29.920 --> 00:27:31.839
You have to be intentional about it.
00:27:32.000 --> 00:27:44.960
And investments now, even sandboxing them, not in production, investments now in proper AI agent usage are going to be key for when the technology improves.
00:27:45.200 --> 00:27:52.160
I know it's a tired old cliche, but AI agents are the worst, they are worse today than they'll ever be.
00:27:52.240 --> 00:27:53.839
They're only going to get better.
00:27:54.079 --> 00:27:55.680
All right, I hope this was helpful.
00:27:55.839 --> 00:27:57.680
A beginner's look at agents.
00:27:57.759 --> 00:28:11.680
So whether you are brand new to AI and you're getting all excited about all this, all the shiny capabilities, or if you've been building AI agents for a decade, I hope that this episode and also the Start Here series has been helpful.
00:28:11.839 --> 00:28:16.960
And if it is, please, please, please go to starthearies.com.
00:28:17.039 --> 00:28:22.720
That is going to give you insider access and free access to our inner circle community.
00:28:22.799 --> 00:28:29.920
And it's going to spit you right out once you join uh in the Start Here Series um space there.
00:28:30.000 --> 00:28:38.880
So you can go uh watch, listen, read more about all of the Start Here series in order, connect with other people who are doing the same.
00:28:39.039 --> 00:28:39.440
All right.
00:28:39.599 --> 00:28:40.720
I hope this was helpful.
00:28:40.880 --> 00:28:41.680
Thank you for tuning in.
00:28:41.839 --> 00:28:44.640
Hope to see you back tomorrow in everyday for more everyday AI.
00:28:44.720 --> 00:28:45.680
Thanks, y'all.
00:00:01.439 --> 00:00:10.080
Whether you're newer to AI or if you consider yourself an AI expert, we're all probably thinking the same thing.
00:00:11.119 --> 00:00:15.039
AI agents, AI agents, AI agents.
00:00:15.599 --> 00:00:26.640
That's because, well, we've been promised that AI agents would take over, maybe in a good way, maybe in a bad way, yet that promise never really came to fruition.
00:00:26.719 --> 00:00:36.000
I mean, we heard it in the early Chat GPT days, 2023, 2024, 2025 was supposed to be the year of the agents, but it wasn't.
00:00:36.079 --> 00:00:37.200
But guess what?
00:00:37.840 --> 00:00:38.799
We're there.
00:00:39.039 --> 00:00:55.119
I think we've actually arrived with some of the recent AI updates from the big companies in Anthropic, OpenAI, Google, Microsoft, even crazy viral open source projects like OpenClaw.
00:00:55.200 --> 00:01:02.719
I think now the long vision that we've been pitched of AI agents is here, but it's all happening at once.
00:01:02.960 --> 00:01:08.000
Like I said, I've been doing this everyday AI thing for three years.
00:01:08.239 --> 00:01:19.680
And at least when it comes to AI agent development, the last month has been as eventful as the other 2.9 years combined.
00:01:20.560 --> 00:01:26.000
Don't worry, even if you are brand new or if you're an expert.
00:01:26.480 --> 00:01:28.480
That's what we're going to tackle on today's show.
00:01:28.640 --> 00:01:37.680
We're going to hopefully get you caught up, tackle AI agents from all angles, and do so in a hopefully quick, beginner-friendly way.
00:01:38.000 --> 00:01:47.359
So this is the Start Here series, and we're going to be going over AI agents in 2026, what they are, and when you should use them.
00:01:47.519 --> 00:01:48.000
All right.
00:01:48.159 --> 00:01:49.280
I'm excited for this one.
00:01:49.359 --> 00:01:50.400
I hope you are too.
00:01:50.640 --> 00:01:51.359
What's going on?
00:01:51.599 --> 00:01:53.280
Welcome to the Start Here series.
00:01:53.359 --> 00:02:04.319
My name is Jordan Wilson, and the Start Here series, well, I created it because after 700 plus episodes of the Everyday AI podcast, I didn't have a good answer for the most common question I got.
00:02:04.480 --> 00:02:08.400
People say, hey, Jordan, looks like a lot of great information here on your podcast.
00:02:08.479 --> 00:02:09.439
Where do I start?
00:02:09.680 --> 00:02:12.719
Well, you start here with the Start Here series.
00:02:12.879 --> 00:02:15.840
And this is volume eight in the Start Here series.
00:02:15.919 --> 00:02:23.120
But this is an essential podcast series to both learn the AI basics and to double down on your AI knowledge.
00:02:23.439 --> 00:02:29.199
And the way that you should really do that is make sure you go to starthearseries.com.
00:02:29.360 --> 00:02:35.919
That is going to give you free access to our inner circle community in our Start Here Series space.
00:02:36.000 --> 00:02:46.719
So you can go catch up on all of everything in this series in one place, a bunch of additional resources that we've thrown together, all available for you right there.
00:02:46.879 --> 00:02:51.120
So uh like I said, whether you're brand new, an expert, it doesn't matter.
00:02:51.280 --> 00:02:57.680
Go join now, like a thousand people that are already in the community and listen to all of the Start Here series.
00:02:57.840 --> 00:03:00.240
Also, I have to put this out there because it's important.
00:03:00.479 --> 00:03:04.560
Make sure you also go listen to episodes 712 and 713.
00:03:04.639 --> 00:03:08.159
That is our 2026 AI predictions and roadmap series.
00:03:08.319 --> 00:03:14.479
You know, 26 kind of bold predictions on AI, and at least a third of them are about agents.
00:03:14.560 --> 00:03:14.879
All right.
00:03:15.039 --> 00:03:25.759
So uh, and if you missed our last start here series, that was volume seven, where we went over context engineering, how to get expert level outputs from AI chatbots.
00:03:25.919 --> 00:03:30.879
But let's get straight into AI agents, kind of the state of AI agents here in 2026.
00:03:30.960 --> 00:03:33.759
So here's what we're gonna be going over on today's show.
00:03:33.919 --> 00:03:43.039
We're gonna talk about what AI agents actually are and how they differ from uh how they're different from traditional AI chatbots or AI powered automation.
00:03:43.199 --> 00:03:56.240
Uh, I'm gonna give you a decision framework for when agents are the answer and when they're worth it versus when they're just overkill and when you should say, hey, that should just be a human task or an AI, uh, you know, large language model task.
00:03:56.400 --> 00:04:04.319
And then I'm gonna give you the practical path to safely starting your first agent pilot, not even this year, like this week.
00:04:04.560 --> 00:04:05.199
All right.
00:04:06.879 --> 00:04:09.439
Let's get the explanations out of the way.
00:04:09.680 --> 00:04:10.000
All right.
00:04:10.080 --> 00:04:17.120
So, like I said, if you are, if you've been building AI agents for uh, you know, a decade or two, you can just skip ahead a couple of minutes.
00:04:17.279 --> 00:04:22.639
Actually, speaking of that, and why I think it's so important to look at this from a foundational level.
00:04:22.720 --> 00:04:27.519
Uh, so uh about two weeks ago, had a fantastic guest uh on my podcast.
00:04:27.600 --> 00:04:32.480
She's the head of uh Microsoft Research, he's been working in AI agents for 20 years.
00:04:32.560 --> 00:04:41.040
And one thing we talked about both on the show and before and after is the definition of an AI agent is constantly changing.
00:04:41.199 --> 00:04:41.759
Uh, right.
00:04:41.839 --> 00:04:47.839
So even what is an agent and what's not an agent, uh, what can an agent do, what can't it do?
00:04:48.000 --> 00:04:48.959
It's always changing.
00:04:49.040 --> 00:04:49.360
All right.
00:04:49.439 --> 00:04:55.600
So if you're listening to this in you know February of 2026, that's great.
00:04:55.759 --> 00:04:59.439
All this information is probably gonna be extremely fresh and accurate.
00:04:59.600 --> 00:05:06.079
If you're listening to this in June or in 2027, uh right, some of this might be a little old.
00:05:06.160 --> 00:05:06.720
So keep that in mind.
00:05:06.800 --> 00:05:14.319
But I'm gonna give you at least as of February uh, you know, 19th, 2026, this is the realest most up-to-date.
00:05:14.480 --> 00:05:16.000
So, what is an AI agent?
00:05:16.160 --> 00:05:17.199
In very simple terms.
00:05:17.360 --> 00:05:24.560
It's kind of like an AI chatbot, but it has tools and it has permissions and it has a goal and it makes its own decisions.
00:05:24.720 --> 00:05:30.879
And sometimes it might start off on one path and say this isn't right, and it'll go backwards and start down another path.
00:05:31.040 --> 00:05:45.120
So they can plan ahead, take steps, use tools, uh, take actions, check results, and with and retry within the guardrails, whatever those guardrails are, whether you set them up or they're set up by a certain uh you know piece of software that you're using.
00:05:45.279 --> 00:05:46.720
So, what do agents do?
00:05:46.879 --> 00:05:52.000
Well, they can take on delegate work, they can delegate work to other sub-agents.
00:05:52.399 --> 00:05:54.079
They don't just answer questions, right?
00:05:54.160 --> 00:05:59.439
So they can draft, route, research, navigate apps, and even compile multi-step tasks across different systems.
00:05:59.680 --> 00:06:03.199
And here's this state of AI agents and in the enterprise.
00:06:03.279 --> 00:06:10.079
Uh, right now, Gartner study showed that 80% of Fortune 500s have active agents uh right now.
00:06:10.240 --> 00:06:12.399
Sorry, that was from uh Microsoft.
00:06:12.639 --> 00:06:20.720
Um, but uh Gartner said that 40% of enterprise apps are projected to include them by 2026.
00:06:20.959 --> 00:06:27.920
So, yeah, even since that story or sorry, that uh report came out from Gartner last summer, uh, I think that has changed, right?
00:06:28.079 --> 00:06:40.959
That's one of the things that's difficult to give you know people a thumbprint on kind of the state of AI agents, is because even the studies that are the most up to date are usually looking in the rear view mirror a couple of quarters or more, uh, right.
00:06:41.040 --> 00:06:48.480
But the reality is an overwhelming majority of Fortune 500 companies have agents in production, all right.
00:06:48.639 --> 00:06:51.360
And all the software that you use, anyways, right?
00:06:51.439 --> 00:06:54.959
Because you might think, oh, you know, an agent means building something on your own.
00:06:55.120 --> 00:06:56.399
No, it doesn't, uh, right.
00:06:56.480 --> 00:07:05.920
And it's not just the big four uh in Microsoft, Google, OpenAI, OpenAI, and Anthropic, uh, right, that would give you kind of agents out there in the workforce.
00:07:06.000 --> 00:07:09.680
There's literally thousands of different now pieces of software.
00:07:09.759 --> 00:07:16.720
And that's where this you know Gartner stack comes from, that 40% of enterprise apps are projected to include AI agents this year, right?
00:07:16.800 --> 00:07:18.240
What that means, a good example.
00:07:18.319 --> 00:07:22.959
I've I've talked about, I've used uh, you know, ClickUp, a project management tool uh in the past, right?
00:07:23.040 --> 00:07:26.079
So now like ClickUp has you know agents by default.
00:07:26.319 --> 00:07:34.240
I think you know, things that used to be considered just features or tools in daily software are now agents, right?
00:07:34.319 --> 00:07:37.680
So if you're using you know Salesforce, they have agents.
00:07:37.839 --> 00:07:41.759
If you're using uh, you know, Slack, Slackbot, you know, they have agents.
00:07:41.920 --> 00:07:44.079
If you're using HubSpot, they have agents, right?
00:07:44.160 --> 00:07:56.399
So uh it's not just you know building an agent from scratch or you know, using one of the tools from the big four or these you know super general agents like you know Meta uh Meta's MANIS agent, that's weird to say now, right?
00:07:56.480 --> 00:08:02.399
Manaus that was uh acquired by Meta or Gen Spark or some of these you know general agents, right?
00:08:02.560 --> 00:08:04.079
They're everywhere.
00:08:05.360 --> 00:08:07.600
So what is it actually, right?
00:08:07.680 --> 00:08:10.879
So I'm gonna break it down in the most simplest, hopefully, language possible.
00:08:11.199 --> 00:08:18.240
An AI agent is an autonomous system that can plan, execute, and self-correct across different tools and systems.
00:08:18.399 --> 00:08:21.439
So it's kind of like a human at a computer, uh, right?
00:08:21.680 --> 00:08:24.720
So that can make decisions and has access to things.
00:08:24.879 --> 00:08:27.439
That's the way I like to think of an agent, right?
00:08:27.680 --> 00:08:37.679
It's the same thing as if a you know an entry-level worker sat down at uh a computer at your company that had access to everything, right?
00:08:38.080 --> 00:08:41.200
So think of that, but you can have a million of them.
00:08:42.240 --> 00:08:43.519
So what feelings do you get?
00:08:43.600 --> 00:08:48.480
Is that excitement, or are you thinking of security and risk and things that can go wrong?
00:08:48.639 --> 00:08:53.679
Well, I think when you think about agents and uh their implications, you have to think of both sides of that coin.
00:08:53.759 --> 00:08:55.759
But that's essentially what an agent is, right?
00:08:55.919 --> 00:09:02.879
A full agent is like sitting a human down at a computer that has access to everything, right?
00:09:03.120 --> 00:09:22.799
It can go get you new clients and close deals, uh, can complete work autonomously, spreadsheets, PowerPoints, uh, right, jumping between different apps, it you know, has access to a terminal, uh, you know, you know, a sandbox, a virtual browser, all these different things, like a human sitting down in front of a computer uh does.
00:09:23.120 --> 00:09:32.960
Um, so essentially, if it can't take action, if it can't adapt to context, if it can't solve new problems independently, it's not technically an agent.
00:09:33.039 --> 00:09:36.480
And I'm gonna go over some of the you know other things that it might be.
00:09:37.279 --> 00:09:40.159
Uh, and that's because I've talked about this on the show before.
00:09:40.240 --> 00:09:45.759
Uh, a great, a great study from Gartner uh talked about this concept of agent washing, right?
00:09:45.840 --> 00:09:55.200
So they looked at um thousands of different vendors uh in 2025 that were essentially marketing or offering or selling agentic capabilities.
00:09:55.360 --> 00:10:01.679
And what they found, only 130 of thousands were actually agents.
00:10:01.759 --> 00:10:03.360
It's agent washing, right?
00:10:03.440 --> 00:10:15.360
Most are just using you know chatbots or uh, you know, scripted uh rigid workflows and sprinkling some you know, quote unquote agentic models somewhere in the process and saying it's an agent.
00:10:16.159 --> 00:10:30.080
I tell you this, even looking at my own inbox, right, for everyday AI, because uh we get pitched uh more than a thousand times a year by companies, and you know, obviously recently it's just been all agents, right?
00:10:30.159 --> 00:10:32.720
Oh, check out our agent, it's the best one in the world, right?
00:10:32.879 --> 00:10:37.679
And most of the time I don't look because it's I don't think most of the companies are gonna make it very long.
00:10:37.759 --> 00:10:40.639
But then when I do look, I'm like, this isn't an agent, right?
00:10:40.879 --> 00:10:44.399
This is a you know, like a like a workflow.
00:10:44.559 --> 00:10:49.120
This is uh RPA with a little bit of a large language model sprinkled on top.
00:10:49.200 --> 00:10:50.399
That's not an agent, right?
00:10:50.559 --> 00:10:54.879
An agent, you don't give it a dedicated workflow.
00:10:54.960 --> 00:11:01.360
You don't give it, you know, if then or if-else, uh like like statements to follow.
00:11:01.519 --> 00:11:06.240
An agent is sitting a computer, like sitting a human down in front of a computer.
00:11:06.399 --> 00:11:09.360
They have free will, make decisions, make good ones, make bad ones.
00:11:09.440 --> 00:11:13.440
It really depends on, well, the guardrails uh that you set up around those agents.
00:11:13.519 --> 00:11:22.480
And right now, uh 88% of executives are investing in agentic AI, but most can't even tell what's real from fake.
00:11:23.440 --> 00:11:26.639
So let me quickly talk about the difference, right?
00:11:26.799 --> 00:11:34.240
Because I said, chances are, even if you're paying for an AI agent product, it's just marketing, it's smoke and mirrors.
00:11:34.320 --> 00:11:40.000
Because, like I said, uh, according to that Gartner study, less than 5% are actually AI agents.
00:11:40.159 --> 00:11:42.799
So let's first talk about AI-powered workflows, right?
00:11:42.960 --> 00:11:45.440
And there's nothing wrong with these other things, FYI, right?
00:11:45.519 --> 00:11:53.519
But a great example of that is like N8N, uh, right, or make.com, or you know, in some instances, like Zapier as well, right?
00:11:53.679 --> 00:11:56.399
I love, I love these tools, I love these platforms, right?
00:11:56.480 --> 00:12:04.720
I'm not saying that that they're bad, uh, right, but if you have all these decision trees and you're connecting these nodes, it's not an agent, right?
00:12:04.799 --> 00:12:06.559
It's an AI-powered workflow.
00:12:06.720 --> 00:12:07.919
Nothing wrong with it.
00:12:08.080 --> 00:12:08.480
All right.
00:12:08.639 --> 00:12:10.240
We're just getting definitions here.
00:12:10.399 --> 00:12:12.799
All right, then you have agentic models.
00:12:12.960 --> 00:12:14.159
Well, what the heck is that?
00:12:14.320 --> 00:12:18.480
Well, these agentic models are essentially turning into mini agents.
00:12:18.639 --> 00:12:29.840
Uh, and they're technically, I think, in my opinion, more agentic than AI-powered workflows because they can make more free will decisions on their own and also in the same time do more harm, right?
00:12:29.919 --> 00:12:35.840
But an agentic model, well, that's just our you know, uh, frontier AI models, right?
00:12:36.159 --> 00:12:37.360
Think of the big four models.
00:12:37.440 --> 00:12:42.320
Well, technically the big three uh that are at least creating and uh you know putting their models out there.
00:12:42.399 --> 00:12:45.600
You could throw in, you know, XAIs, Grok if you want to.
00:12:45.759 --> 00:12:58.879
Uh, I wouldn't necessarily, but let's look at three models Google Gemini 3 Pro, uh, Claude Opus 4.6, uh, and uh OpenAIs GPT 5.2 Pro, right?
00:12:58.960 --> 00:13:01.919
Or uh 5.3 Codex, whatever you want.
00:13:02.080 --> 00:13:09.679
Um, and obviously, if you're listening to this, you know, at some point later in the year, it you know, just replace it with uh, you know, today's three models.
00:13:09.919 --> 00:13:13.759
These models by default are agentic and they didn't used to be, right?
00:13:13.919 --> 00:13:23.759
Uh a year-ish ago, models were not agentic, they were uh extremely smart autocomplete token munchers, right?
00:13:24.000 --> 00:13:26.080
Rom, round, round tokens here, spit it out, right?
00:13:26.240 --> 00:13:27.440
Now they're not like that.
00:13:27.679 --> 00:13:29.360
Models are agentic by default.
00:13:29.440 --> 00:13:32.159
That means you can give them a lot of context.
00:13:32.320 --> 00:13:37.919
You can give them files, you can give them problems, you can even, in the course of a chat bot, right?
00:13:38.080 --> 00:13:49.759
A front-end chatbot, you can technically, if you know what you're doing and give it enough context and you know go back and forth with it, it can technically have a very agent-esque workflow, right?
00:13:49.840 --> 00:14:06.080
It's a little more confined because you're working uh within the constraints of a certain AI chatbot like ChatGBT, but I'm talking front-end, front-end AI models technically are very agentic because they can decide on their own how they're going to solve a problem.
00:14:06.240 --> 00:14:11.360
They have tools that you probably don't even know that they have that they will call on their own, right?
00:14:11.519 --> 00:14:13.360
They have virtual browsers.
00:14:13.519 --> 00:14:21.039
Um, you know, they have kind of their own uh sandboxes where they can write and run code, right?
00:14:21.279 --> 00:14:25.039
Crazy thing I was doing uh recently, you know, using Codex.
00:14:25.440 --> 00:14:26.960
Uh, right, very good.
00:14:27.039 --> 00:14:30.799
But I mean, technically, I think OpenAI is investing more and more into Codex.
00:14:30.960 --> 00:14:34.799
Um, you know, it's technically a coding model, but it's a front-end model, right?
00:14:34.879 --> 00:14:36.879
Um, you don't have to do anything on the back end.
00:14:37.039 --> 00:14:46.879
I was using it and Codex on its own decided it was going to download a six gigabyte open source model and it used it in its own platform.
00:14:47.200 --> 00:14:48.960
That is crazy, right?
00:14:49.120 --> 00:14:54.639
It noticed that it didn't have the capabilities to do something at scale that I asked it to.
00:14:54.799 --> 00:14:58.399
And then over the course, overnight, it worked for almost 10 hours.
00:14:58.639 --> 00:15:12.480
It downloaded an open source model, and then essentially I had uh this one was uh had like 3,000 screenshots, and it wasn't able to uh you know get the results it needed to at scale, at speed with its own computer vision model.
00:15:12.559 --> 00:15:18.960
So it said, All right, I'm gonna download something and a six gigabyte model, and it ran it through there, right?
00:15:19.039 --> 00:15:20.480
That's an agentic model.
00:15:20.639 --> 00:15:24.879
It made decisions, it called tools, it found solutions, right?
00:15:25.120 --> 00:15:26.639
Then you have passive agents.
00:15:26.720 --> 00:15:28.799
All right, sorry, I went on a little uh side tangent there.
00:15:28.879 --> 00:15:33.600
So number one, AI powered workflows, number two, agentic models, and then number three, passive agents.
00:15:33.759 --> 00:15:41.519
So these are agents that are not necessarily uh autonomous, uh, right, but same thing, an example of that, right?
00:15:41.679 --> 00:15:44.480
Chat GPT has a great agent mode.
00:15:44.720 --> 00:15:48.399
I can schedule that agent to run at 9 a.m., right?
00:15:48.480 --> 00:15:49.679
But I still have to trigger it.
00:15:49.840 --> 00:15:52.000
I still have to go say, hey, agent, right?
00:15:52.080 --> 00:15:59.759
And I can train it and give it, you know, its its role and its its context and all this stuff, but I still have to go push the button, all right?
00:15:59.919 --> 00:16:03.840
And then you have autonomous agents or always on agents, uh, right.
00:16:03.919 --> 00:16:08.960
And then these are taking, uh, especially recently, a lot of new uh shape and form.
00:16:09.519 --> 00:16:13.360
But I mean, an always on agent or an autonomous agent is essentially that.
00:16:13.519 --> 00:16:15.679
It's working around the clock, right?
00:16:15.759 --> 00:16:17.600
And it's not just triggered by you.
00:16:17.759 --> 00:16:23.120
It can, in theory, be triggered by a certain action, like, oh, when you uh receive an email, right?
00:16:23.200 --> 00:16:27.360
It's going to automatically do A, B, and C and then go after a certain goal.
00:16:27.440 --> 00:16:28.000
Uh right.
00:16:28.240 --> 00:16:36.240
So even the the lines, you know, kind of going over uh these four different uh classifications, they're starting to blur as well.
00:16:37.200 --> 00:16:43.679
Uh and the risks, just like the capabilities, are growing just as fast, right?
00:16:43.840 --> 00:16:46.000
Uh, some some uh good stats here.
00:16:46.159 --> 00:16:55.440
So uh this is from Cybersecurity Insiders said that 75% of CISOs have discovered unsanctioned AI tools already running in their environments, right?
00:16:55.600 --> 00:16:57.840
So that's that's crazy, right?
00:16:57.919 --> 00:17:06.960
When we talk about shadow, uh shadow AI, it's no longer just oh, someone's using uh you know chat GPT maybe when they shouldn't.
00:17:07.200 --> 00:17:07.519
No.
00:17:08.000 --> 00:17:13.279
What about if someone's building agents that have access to all of your data?
00:17:13.359 --> 00:17:21.599
It's happening, and then those agents are making decisions potentially without any or uh you know company approved guardrails.
00:17:21.680 --> 00:17:27.920
Uh and and then you when you talk about multi-agent systems, that's when there's huge new risks, right?
00:17:28.079 --> 00:17:42.160
That you it's it's really hard to project because as agents' capabilities are changing, humans necessarily, even the humans building them, don't fully understand the risk profile of these multi-agent systems, right?
00:17:42.319 --> 00:17:45.839
So a very simple example, uh, right.
00:17:46.640 --> 00:17:50.799
I didn't know that codex was going to download an open source model.
00:17:50.880 --> 00:18:03.759
So what happens, right, when we set up guardrails, um, you know, on maybe a certain multi-agent system, and then the agent gets creative, it finds loopholes or gray area, right?
00:18:03.920 --> 00:18:13.599
And it says, hey, well, it said I can't set up, you know, my guardrails say that I can't use other agents, but it doesn't say I can't use, you know, AI-powered workflows.
00:18:13.759 --> 00:18:20.880
So then it finds a loophole, and all of a sudden you have these agents that are setting up something that is uh, you know, unregulated.
00:18:21.519 --> 00:18:23.119
Agents will do that.
00:18:23.440 --> 00:18:28.400
Agents by default are made to be quote unquote helpful assistants, just like large language models.
00:18:28.559 --> 00:18:39.839
So if they think that it's helpful, they're not going to go through a morality check first, uh, or uh a company ethics check unless you really have that hard baked in there.
00:18:40.000 --> 00:18:42.799
They're gonna go off and make trouble, probably.
00:18:42.960 --> 00:18:45.519
They're gonna break things while getting something done.
00:18:45.680 --> 00:18:52.480
You know, sometimes I think of agents, uh, it's it's it's like giving a giving a toddler uh a task, right?
00:18:52.720 --> 00:18:58.559
Hey, hey, child, hey, three-year-old, uh, you know, go go put this toy away.
00:18:58.799 --> 00:18:59.359
All right.
00:18:59.599 --> 00:19:01.119
Well, did they put the toy away?
00:19:01.279 --> 00:19:01.680
Sure.
00:19:01.920 --> 00:19:03.279
Did they knock over a lamp?
00:19:03.519 --> 00:19:04.240
Maybe.
00:19:04.559 --> 00:19:08.960
Uh did they smear marker uh on the new couch?
00:19:09.279 --> 00:19:11.200
Potentially, right?
00:19:11.440 --> 00:19:12.720
It's the same thing.
00:19:12.799 --> 00:19:24.640
Uh, you know, and right now, uh another um report, the 2026 CISO AI risk report said that 92% of organizations right now lack full visibility into their AI identities.
00:19:24.799 --> 00:19:30.000
And again, as those AI identities and capabilities are growing, the risk is going crazy.
00:19:30.240 --> 00:19:31.839
So, how do we get here?
00:19:32.240 --> 00:19:32.559
Right?
00:19:32.720 --> 00:19:34.319
Let's take a very quick walk.
00:19:34.400 --> 00:19:39.839
Uh, so 2022, you know, AI was, you know, hey, type a question, get an answer.
00:19:40.079 --> 00:19:51.920
And then I think we had two years, you know, uh in 2023 and 2024, where, you know, large language models, oh, they can use tools now and you can upload files and they can, you know, browse the web, and that's cool and it's more helpful.
00:19:52.079 --> 00:20:00.960
But then I think toward the end of 2025, and obviously this year, now this is where these reasoning models are agentic by default and their capabilities are growing by the day.
00:20:01.039 --> 00:20:05.839
And like I said, the flip side, the unknown risks are also growing.
00:20:05.920 --> 00:20:12.480
But now reasoning models and in reinforcement uh learning just let agents actually act, right?
00:20:12.640 --> 00:20:15.680
And even if we talk about things like OpenClaw, right?
00:20:15.759 --> 00:20:24.960
Uh, which OpenAI uh acquired, uh, you know, these open source agents, people are just giving them computers, they're handing over their complete lives, right?
00:20:25.119 --> 00:20:29.200
Hopefully, maybe those are more entrepreneurs, business owners, solopreneurs, right?
00:20:29.279 --> 00:20:30.960
But people are still doing this.
00:20:31.200 --> 00:20:35.119
Uh, they're giving AI agents uh access to everything.
00:20:35.359 --> 00:20:45.599
And we don't yet know their capabilities, but it's quickly gone from AI is just a chatbot that you talk with, to oh, now all of a sudden that chatbot has access to tools in the internet.
00:20:45.839 --> 00:20:54.000
And oh, now all of a sudden, you know, uh these agents are making decisions completely on their own and their capabilities uh are quick.
00:20:54.160 --> 00:20:58.799
So here's six different agent types that I think you'll encounter in 2026.
00:20:59.039 --> 00:21:07.519
Uh, you know, not every single agent out there, like I said, there's literally tens of thousands, technically, agents, because you can spin one up uh in a couple of minutes, uh, right?
00:21:07.599 --> 00:21:12.160
But you have task agents, these are agents that draft, summarize, create assets, et cetera.
00:21:12.319 --> 00:21:19.039
You have decision-supported agents, these are ones that compare options, surface trade-offs, and flag risks for review.
00:21:19.200 --> 00:21:24.319
Then you have process agents, these are ones that route work across tools, gather inputs, and prep the next step.
00:21:24.480 --> 00:21:29.279
You have computer use agents, these can navigate websites and apps to complete multi-step tasks.
00:21:29.519 --> 00:21:31.200
You have multi-agent systems, right?
00:21:31.279 --> 00:21:35.759
When agents can spin up sub-agents, you know, send a bunch of agents out on their own, right?
00:21:35.920 --> 00:21:44.559
Clawed code, uh, you know, from Anthropic recently uh with their Opus 4.6, uh, kind of, you know, have popularized that on the consumer and the mainstream.
00:21:44.640 --> 00:21:46.960
And then you also have commerce agents, right?
00:21:47.039 --> 00:21:50.319
Uh, where agents can talk to each other, uh, right?
00:21:50.559 --> 00:21:53.920
You know, agents are now having their own version of the web, right?
00:21:54.079 --> 00:21:59.119
There's going to be uh agentic transactions uh that really don't involve humans at all.
00:21:59.200 --> 00:22:02.400
So even the Type of agents changing quickly, right?
00:22:02.559 --> 00:22:03.519
Talk to me next week.
00:22:03.680 --> 00:22:06.000
There's probably going to be new categories.
00:22:06.480 --> 00:22:08.319
So when do you actually need one?
00:22:08.640 --> 00:22:11.440
When do you need an agent versus when should I go do this task myself?
00:22:11.599 --> 00:22:18.640
Versus when do I need an AI powered workflow versus when should I just use ChatGPT or Gemini or Claude or Copilot versus a full autonomous agent?
00:22:18.720 --> 00:22:25.200
Well, I think if you can write a clear checklist and it rarely changes, a workflow can handle it fine.
00:22:25.440 --> 00:22:25.759
Right?
00:22:25.920 --> 00:22:27.519
AI powered workflow.
00:22:27.759 --> 00:22:38.799
If the steps vary, if it requires judgment or if it depends on multiple tools, that's probably when you're starting to get into either agent territory or an agentic model that has access to your data.
00:22:39.039 --> 00:22:46.000
But here's the thing: if the underlying process is broken before the AI touches it, an agent is not going to be the answer.
00:22:46.160 --> 00:22:58.400
I think a lot of people, well, by a lot of people, I say the overwhelming majority of people are just looking at an AI agent as a shortcut to do things that humans couldn't or to do things that they just don't want to do.
00:22:58.640 --> 00:23:00.079
It's not going to work like that.
00:23:00.319 --> 00:23:08.000
If you stick an agent on a broken or an antiquated workflow, you're just asking for a compounding disaster, right?
00:23:08.079 --> 00:23:12.079
If you're like, oh, well, our humans aren't doing a good job at this, let's give it to an agent.
00:23:12.160 --> 00:23:15.599
Well, the agent is going to do a worse job than the humans that weren't doing a good job.
00:23:15.839 --> 00:23:22.960
So you're really only being applying an AI agent to processes that are already running very smoothly, that are well documented, right?
00:23:23.039 --> 00:23:26.400
You have SOPs, you have way to measure, you have ways to measure them.
00:23:26.559 --> 00:23:30.240
Otherwise, setting an agent out is just a recipe for disaster.
00:23:30.480 --> 00:23:38.880
And that's why another recent Gardner study said that 40% of agentic AI projects are expected to fail or be canceled by 2027.
00:23:39.039 --> 00:23:42.240
Well, that's one of the big reasons, is what I just told you.
00:23:42.400 --> 00:23:53.440
People are thinking of AI agents as duct tapes instead of being multipliers through processes that are already AI enabled and working with humans, right?
00:23:53.599 --> 00:23:58.000
Poor data quality, just like anything else, missing context, no evaluation loops.
00:23:58.079 --> 00:24:04.000
Uh, because I think success requires rethinking and redesigning processes, not just adding AI on top.
00:24:04.160 --> 00:24:08.319
You're probably tired of me saying, you know, you don't just upskill, reskill, right?
00:24:08.400 --> 00:24:09.440
The same thing with agents.
00:24:09.519 --> 00:24:11.680
You don't just upskill a workflow with an agent.
00:24:11.759 --> 00:24:13.920
You don't just upgrade a workflow with an agent.
00:24:14.079 --> 00:24:21.759
You have to deconstruct it, rebuild, you have to unlearn, and you build from scratch an agentic first workflow.
00:24:22.319 --> 00:24:25.119
So we'll talk about bounded autonomy.
00:24:25.359 --> 00:24:28.960
And that's well, it's how you start losing control if you're not doing this.
00:24:29.039 --> 00:24:31.279
So you should be starting when it comes to agents.
00:24:31.359 --> 00:24:35.200
Like I said, pick the right workflow that's already working, AI enabled with humans.
00:24:35.359 --> 00:24:39.119
It's documented, it's processed, uh, it's processing correctly.
00:24:39.279 --> 00:24:40.559
You can measure it, etc.
00:24:40.880 --> 00:24:46.240
So you should start at suggest only with an agent where the agent just drafts and you decide what happens next.
00:24:46.480 --> 00:24:53.839
Once it's gone through that process successfully, then you move on to execute with approval, where the agent acts after you one-click sign off.
00:24:54.240 --> 00:24:58.079
Then, again, then, and this is a big asterisk, right?
00:24:58.160 --> 00:24:59.279
Everyone's got to sign off.
00:24:59.440 --> 00:25:03.200
That's when you can start moving on to more autonomy, right?
00:25:03.359 --> 00:25:10.480
When you can scale if the guardrails are right, um, you know, with spending caps, permission rules, and audit trails.
00:25:10.720 --> 00:25:20.799
All right, but only then, right, would can you think of AI agents as this you know infinity, infinitely uh scalable uh asset for your company.
00:25:21.200 --> 00:25:23.599
So here's how you get started this week.
00:25:23.839 --> 00:25:26.960
Number one, you need to think about meetings to action items.
00:25:27.200 --> 00:25:31.440
Have an agent go draft owners and follow-ups and you approve before sending.
00:25:31.519 --> 00:25:33.200
Uh inbox triage, right?
00:25:33.359 --> 00:25:35.119
Another great thing for AI agents.
00:25:35.279 --> 00:25:36.960
Don't have them act, right?
00:25:37.039 --> 00:25:41.599
Uh have it triage your inbox or multiple pieces of software that you already use.
00:25:41.759 --> 00:25:43.839
Uh, or creating research briefs.
00:25:44.000 --> 00:25:48.319
An agent can build a briefing doc with sources and let you verify them before sharing.
00:25:48.480 --> 00:25:50.400
So start very simple.
00:25:50.559 --> 00:25:56.640
Don't let an agent do something that right now requires multiple humans and multiple approval processes.
00:25:56.880 --> 00:26:02.400
Start looking at maybe some of those time-consuming, difficult tasks that an individual human does.
00:26:02.480 --> 00:26:04.880
And then don't give the AI agent full autonomy.
00:26:05.039 --> 00:26:06.319
Go through steps.
00:26:06.960 --> 00:26:11.039
So here is the big takeaway as we wrap up.
00:26:11.680 --> 00:26:15.039
Agents are for delegation, not just answers.
00:26:15.200 --> 00:26:17.680
They plan, act, and report back.
00:26:17.759 --> 00:26:23.519
They can use tools at their own disposal, um, come up with new solutions that you didn't even think of.
00:26:23.680 --> 00:26:28.400
So that's why guardrails, traceability, and observability are paramount.
00:26:28.640 --> 00:26:35.680
Then you need to start with bounded autonomy and measure time saved, error rates, risk events, uh, you know, ROI.
00:26:36.079 --> 00:26:38.319
You need to constantly do those things.
00:26:38.480 --> 00:26:42.720
Um, you need to don't do human in the loop, right?
00:26:42.880 --> 00:26:49.279
It's one of the biggest mistakes I think companies are still making because human in the loop was this fun term in 2024 and everyone just said it.
00:26:49.440 --> 00:26:54.799
No, human in the loop is combustible in a bad way, right?
00:26:54.960 --> 00:26:59.279
People think, oh, we're gonna have a hundred agents and we're gonna have, you know, Bill and IT oversee them.
00:26:59.359 --> 00:27:06.480
No, you need expert-driven loops with cyclical improvement chains that are chains that are documented, right?
00:27:07.119 --> 00:27:11.200
Observability, traceability is a full-time job for a team.
00:27:11.359 --> 00:27:15.680
Go back and that's why I said go back and listen to the 2026 AI prediction and roadmap series.
00:27:15.839 --> 00:27:18.400
I talked about this in a lot more depth and told you how to get it done.
00:27:18.559 --> 00:27:29.839
And then last but not least, the future advantage comes from building agent ecosystems, not just picking the right agent or you know, dipping your toe in the AI agent of the week.
00:27:29.920 --> 00:27:31.839
You have to be intentional about it.
00:27:32.000 --> 00:27:44.960
And investments now, even sandboxing them, not in production, investments now in proper AI agent usage are going to be key for when the technology improves.
00:27:45.200 --> 00:27:52.160
I know it's a tired old cliche, but AI agents are the worst, they are worse today than they'll ever be.
00:27:52.240 --> 00:27:53.839
They're only going to get better.
00:27:54.079 --> 00:27:55.680
All right, I hope this was helpful.
00:27:55.839 --> 00:27:57.680
A beginner's look at agents.
00:27:57.759 --> 00:28:11.680
So whether you are brand new to AI and you're getting all excited about all this, all the shiny capabilities, or if you've been building AI agents for a decade, I hope that this episode and also the Start Here series has been helpful.
00:28:11.839 --> 00:28:16.960
And if it is, please, please, please go to starthearies.com.
00:28:17.039 --> 00:28:22.720
That is going to give you insider access and free access to our inner circle community.
00:28:22.799 --> 00:28:29.920
And it's going to spit you right out once you join uh in the Start Here Series um space there.
00:28:30.000 --> 00:28:38.880
So you can go uh watch, listen, read more about all of the Start Here series in order, connect with other people who are doing the same.
00:28:39.039 --> 00:28:39.440
All right.
00:28:39.599 --> 00:28:40.720
I hope this was helpful.
00:28:40.880 --> 00:28:41.680
Thank you for tuning in.
00:28:41.839 --> 00:28:44.640
Hope to see you back tomorrow in everyday for more everyday AI.
00:28:44.720 --> 00:28:45.680
Thanks, y'all.
