×
Evaluate Your Team's AI Readiness

Take our official assessment and explore training paths to build secure, compliant AI workflows.

Access Portal

AI at Work Podcast

Ep #10 The “Knowledge Graph”: The Missing Link Between Your CRM and AI

AI at Work, AI Fundamentals, Featured Podcast

Audio episode

Listen to this episode

Summary

Kevin finally gets a nickname (Kevlar wins), then brings back field notes from LA: even smart execs are still stuck at AI 101, using chatbots for broad personal questions instead of real work leverage. The deeper unlock is not buying licenses, it’s building workflows, training the system with human feedback, and capturing the “between the cracks” context that lives in Slack, email, and meeting transcripts. They also dig into Claude Code and Anthropic’s Cowork push, plus the growing sprint into health AI, where the upside is massive and the risks are very real.

Takeaways

Most teams are still “shouting into the cavern” with prompts, not designing repeatable workflows that produce reliable outputs.

Rolling out Copilot style licenses without training, feedback loops, and ownership usually leads to disappointment.

The real value is in connective tissue data: Slack, email, and meeting transcripts, not just SOPs and static docs.

Human-in-the-loop tuning is not optional, dedicate top people to review outputs for 30 days and results improve fast.

Agent tools like Claude Code and Cowork point toward local, high-context automation, but the learning curve and security tradeoffs are real.

Chapters
00:18 Cold open: “Kevin.”

00:26 Nickname hunt begins (AI picks five)

01:35 Rejecting the options: K-dub, K. Will, Big Kev…

02:10 “Kevo” and “W”: maybes, takes, and Grok jokes

03:05 Landing the plane: Kevlar wins (and “Special K” sticks)

05:25 LA trip recap + real-world AI adoption (101 vs advanced users)

06:56 Personal vs work use: who’s actually using AI at work?

09:17 CEO angle: AI as thought partner vs tactical helper

10:39 Vendor example: AI in concrete workflows improves speed/quality

12:57 Four places AI shows up: code, workflows, strategy, product integration

20:10 Leadership + the “AI manifesto” (human-forward stance, policy gaps)

25:35 Why “sprinkle Copilot” fails: training + human reinforcement matter

27:57 Knowledge graph: the “connective tissue” (Slack/Teams, meetings, email)

33:00 Claude Code + Claude Coworker: bigger context, local folders, terminal gap

47:16 AI in healthcare: trust, privacy, wearables + business model friction

 

 

 

Explore practical AI implementation: https://assessment.ascendlabs.ai/

Book a conversation with Kevin: tidycal.com/kevinwilliams

Show Notes & Links: https://www.spark6.com/podcast

Submit listener questions:

[email protected]

[email protected]

Check out Kevin’s stuff:

AscendAI: https://ascendlabs.ai/

Follow Kevin on LinkedIn: https://www.linkedin.com/in/kevinguywilliams/

Check out Eli’s Stuff:

SPARK6 Agency: https://www.spark6.com/

Sign up for FREE AI Framework Friday Newsletter: https://www.spark6.com/newsletter

Follow Elijah on LinkedIn: https://www.linkedin.com/in/elijahszasz/

Transcript and citation details

Transcript

00:00 Play from here Hey, Eli, what's up, buddy?

00:14 Play from here Oh, you know, it's another day in paradise.

00:17 Play from here You know what I was thinking about something, Kevin?

00:20 Play from here I just want to call you Kev Kev.

00:22 Play from here And it's because on Tim Ferriss's podcast, when he has Kevin Rose on, he always calls him Kev Kev.

00:28 Play from here And it made me think, you need a good nickname.

00:30 Play from here Like all my buddies have good nicknames.

00:32 Play from here Do you have one that just like in your back pocket you've been holding out on or no?

00:37 Play from here I think we should just file away that no and no.

00:43 Play from here I smell an opportunity, like where you see a no, I see an opportunity.

00:48 Play from here So I'm going to let AI pick your nickname and I'm going to give you a choice of five.

00:54 Play from here But I'm going to give you three choices of five from different models.

00:58 Play from here And you're going to pick the one you like that resonates with you the most.

01:01 Play from here And I'm going to tell you then, but only then, what model it came from.

01:04 Play from here Are you ready?

01:05 Play from here Oh, I'm ready.

01:06 Play from here I am ready for this.

01:07 Play from here So it is important that not all of us have impossible to spell names.

01:11 Play from here And Kevin Williams is pretty much the John Smith of America.

01:15 Play from here There are at least 11,000 of us.

01:18 Play from here Or Joseph Smith because we're in Utah.

01:21 Play from here There are a lot of Smiths around too.

01:23 Play from here But, yeah, it's sort of nice to be anonymous sometimes, but okay.

01:29 Play from here Okay.

01:30 Play from here What about K-Dub?

01:31 Play from here Oh, no.

01:32 Play from here Because that model calls it classic punchy and rolls off the tongue.

01:38 Play from here What about K-Will?

01:42 Play from here We're not going to do any of these.

01:44 Play from here Okay.

01:45 Play from here Big Kev.

01:46 Play from here Big Kev.

01:47 Play from here Oh, absolutely not.

01:48 Play from here No, no, no.

01:49 Play from here Big Kev.

01:50 Play from here Big Kev is a non-starter.

01:51 Play from here That was your nickname when you were the star quarterback in high school.

01:54 Play from here Okay.

01:55 Play from here Of course I was.

01:57 Play from here Of course you were.

01:58 Play from here What about Kevo?

02:02 Play from here I may have somebody from my youth who calls me Kevo, so we might be getting into it.

02:06 Play from here Okay.

02:07 Play from here We're going to – you're marked out as a maybe.

02:09 Play from here W, which I didn't get, but it's a clever play on W and Williams.

02:14 Play from here It's unique, memorable, and it has a bit of insider feel.

02:18 Play from here I believe there was a president we called W, and so, therefore, it's taken,

02:23 Play from here and it would be confusing for people between me and that former president.

02:27 Play from here Only one said with that Texas accent, though.

02:31 Play from here I want to know which one was from Grok.

02:33 Play from here I'm not going to bet.

02:34 Play from here Oh, boy.

02:35 Play from here You know, I don't use Grok.

02:36 Play from here Come on.

02:38 Play from here Unless I'm trying to verify something I read on Twitter.

02:43 Play from here That's the only time I've used it.

02:45 Play from here Remember a time for Grok?

02:46 Play from here That was a time for Grok.

02:47 Play from here Probably.

02:48 Play from here We might do a redo, and I'll add Grok in there.

02:50 Play from here But, okay.

02:51 Play from here What about the K-Bot?

02:58 Play from here It's a fun thematic for this podcast on AI and implies that you're the one with the

03:04 Play from here programmed knowledge or a systemic thinker.

03:09 Play from here The K-Bot.

03:10 Play from here Okay.

03:12 Play from here You actually prepped this, didn't you?

03:14 Play from here Listen, man.

03:15 Play from here I had two minutes to kill before we started because of all the technical difficulties

03:19 Play from here we've been incurring.

03:21 Play from here So, what about K-Will?

03:24 Play from here I mean, it is descriptive.

03:25 Play from here It gets you there.

03:26 Play from here It's descriptive.

03:27 Play from here I think I kind of like K-Will, although I'm going to say this one's my favorite because

03:31 Play from here it's got a little throwback thing, and it's a bit of a double entendre.

03:34 Play from here Special K.

03:39 Play from here Special K.

03:40 Play from here You may not be able to shake this one.

03:42 Play from here It may be a special K from here on out.

03:43 Play from here That's okay.

03:44 Play from here I'm just telling you, man.

03:45 Play from here But I do like Kevlar.

03:47 Play from here Kevlar?

03:49 Play from here Strong, dependable, and a little badass.

03:53 Play from here I'll take that one.

03:55 Play from here I'm going to roll with Kevlar.

03:57 Play from here There we go.

03:58 Play from here Double K.

03:59 Play from here I don't know where it got that other than it says that it rolls off the tongue like

04:06 Play from here 007.

04:07 Play from here Has a spy movie energy.

04:09 Play from here I sort of feel like double K would be something in the Urban Dictionary that is really unfortunate,

04:14 Play from here and you'd hear it in the live, and teenagers' eyes would widen, and you'd know you've done

04:20 Play from here something completely horrible.

04:21 Play from here Oh, for sure.

04:22 Play from here It's in Urban Dictionary for sure.

04:23 Play from here It's like right under Dirty Sanchez.

04:24 Play from here Yeah.

04:25 Play from here Okay.

04:26 Play from here What about The Algorithm?

04:27 Play from here The Algorithm.

04:29 Play from here Oh, man.

04:33 Play from here These are all really good.

04:36 Play from here Special K is just awesome, man.

04:38 Play from here I don't think I'm going to be able to get that one out of my head, and GPT 5.2 came

04:46 Play from here up with that one.

04:48 Play from here Here's what's interesting.

04:50 Play from here Both Claude, GPT, and Gemini all came up with Kdub, which I thought was so weird, and how

04:58 Play from here did all three of those models independently come up with Kdub?

05:02 Play from here Super, super odd.

05:05 Play from here Gemini also, The Algorithm, and that's so fitting coming from Gemini.

05:11 Play from here Of course, it's going to call you The Algorithm, but Art Special K, let's do this, man.

05:15 Play from here How are you doing?

05:16 Play from here You went to LA.

05:17 Play from here I did.

05:18 Play from here I just got back from LA middle of the night last night, so I spent a couple of days speaking

05:25 Play from here with various groups, an alumni association, a large group of CEOs.

05:32 Play from here It was an interesting time to see just the diversity of experience with AI.

05:40 Play from here I think you and I talk about this a lot as far as where the industry is, and we live

05:46 Play from here and breathe this stuff, so it feels like it's just absolutely everywhere, but the alumni

05:53 Play from here group was great folks, super smart, all from an Ivy League school, and really, really

06:01 Play from here accomplished people, and it was total 101.

06:04 Play from here There were a couple of people in the room who were doing some kind of interesting experimentation,

06:09 Play from here but of the people in the room, most had no idea really how to prompt or sort of the basics

06:18 Play from here there, so we still have a long way to go.

06:22 Play from here Consumers are using it.

06:23 Play from here They're using it for travel planning, cooking, fixing appliances at home.

06:30 Play from here These are sort of the things that came out of that, but they're really still just shouting

06:36 Play from here into the cavern looking for broad answers.

06:41 Play from here Nobody's really gotten the experience to have a more fine-tuned output.

06:48 Play from here Okay, so are you saying that most of the people that you interacted with just really aren't

06:54 Play from here using it for work at this point, like it's all personal use?

06:59 Play from here It was a mix in that particular room.

07:02 Play from here The CEO group is very different.

07:04 Play from here I would say that it was one of the more advanced groups that I've sat down with in the last

07:09 Play from here six months or so as far as practical applications starting to hit the ground, but in a lot of

07:15 Play from here ways, they were sort of hitting the ground around people as opposed to it being like

07:20 Play from here a dedicated leadership initiative, something we're going to think about like holistically

07:24 Play from here how we're going to use this.

07:27 Play from here It was like little bright points of light in the organization that were starting to

07:30 Play from here solve problems, which was promising, but again, there wasn't a lot of planning behind it.

07:38 Play from here Got it.

07:39 Play from here In the more public group, yeah, it was a little wild.

07:43 Play from here A few people were using it beyond doing like some basic research for work, maybe some email

07:49 Play from here construction, et cetera, and it really didn't feel like it was being embraced in their workplaces.

07:57 Play from here Now, in fairness, the room ran the gambit from a young guy, young 20s who's a sales

08:04 Play from here guy, and he was super hot to trot about how can he use this with Salesforce and works

08:12 Play from here in a giant organization that isn't adopting it.

08:15 Play from here Sure enough, without practice or policy in place, he wasn't able to do any of the things

08:21 Play from here that he wants to do or get anybody's attention, which is sort of crazy.

08:24 Play from here If you have somebody who's young and eager, doesn't need to be young, but just eager in

08:28 Play from here your organization who wants to experiment, giving them room to experiment is great.

08:33 Play from here That's never happened before.

08:35 Play from here Then on the other end of the spectrum, there were people who were near retirement who are

08:42 Play from here new to a lot of this, I would say.

08:45 Play from here That makes a lot more sense.

08:48 Play from here Yeah.

08:49 Play from here That makes a lot more sense.

08:51 Play from here But from a leadership perspective and talking to CEOs, what I'm curious about is if you

08:55 Play from here saw a high level of implementation in their own workflows and what they were doing and

08:59 Play from here what they were learning versus trying to roll it out to an entire organization, albeit a

09:04 Play from here really large one, I mean, that's a much more daunting task, right?

09:09 Play from here But if I was a CEO of a big company, I'm a CEO of a very small company right now, and

09:17 Play from here it's basically running the whole show for me, more or less, as a big company, I would

09:22 Play from here think at the very least, I'd be using it as a thought partner to challenge my thinking

09:27 Play from here to come up with strategy and not to mention just all of the barrels and barrels of things

09:35 Play from here landing on my desk that AI can now help me summarize and make sense of, right?

09:40 Play from here So what kind of feedback did you get in that regard?

09:44 Play from here I would say that that's a relatively rare use, that people are using it for highly tactical

09:49 Play from here things, help me analyze this, help me prepare for this, whatever it might be, as opposed

09:55 Play from here to as a thought partner.

09:56 Play from here So a lot of that talk, and the talk was three hours long, we were really trying to go in

10:02 Play from here some depth and get people to start thinking about it from that perspective.

10:07 Play from here And I like to think that it was pretty eye-opening for them.

10:11 Play from here Most of the use was actually coming from the vendor side.

10:15 Play from here So vendors were solving problems and adding AI to their offering in ways that this group

10:25 Play from here appreciated, in that it was starting to accelerate their own processes.

10:31 Play from here Can you give me an example of that?

10:33 Play from here Yeah.

10:34 Play from here A vendor and a problem and a solution?

10:36 Play from here Sure.

10:37 Play from here This is asking for an example.

10:41 Play from here It was a company that deals with concrete.

10:45 Play from here And concrete, it turns out, is very difficult to deal with in different temperatures and

10:50 Play from here different consistencies and has to be done very, very well.

10:54 Play from here And there are all of these documents that come along with various batches of concrete.

10:59 Play from here Picture a line of 10 cement trucks waiting around the block.

11:03 Play from here Each of them basically has a slightly different version of the concrete, and the structural

11:08 Play from here engineers need to make sure that it's all balanced.

11:10 Play from here So streamlining basically the batch information, and this is a little bit wonky, but it's like

11:16 Play from here for somebody who does this work, having the chemical composition of the different batches

11:22 Play from here of concrete on hand was a huge unlock for them, as opposed to going to each driver,

11:29 Play from here kind of shaking them down for a clipboard that has some of the information, and to be

11:36 Play from here honest, sort of hoping that the whole thing was going to work.

11:38 Play from here So it not only has increased the pace of the concrete throughput on the job, but it's also

11:46 Play from here increased the quality of the output, and with something like that, time will tell.

11:51 Play from here But this particular CEO was really excited about that result, in that they were going

11:59 Play from here to feel a lot better about the structural output that they were putting together, because

12:04 Play from here this sort of stressor about these different types of concrete is basically no longer an

12:12 Play from here issue for him.

12:13 Play from here That's interesting.

12:14 Play from here You had mentioned vendors as well, right?

12:16 Play from here So was this like a service offering?

12:19 Play from here Oh, yeah.

12:20 Play from here So this was a vendor.

12:21 Play from here So this was the concrete vendor who has done this.

12:24 Play from here So they're not doing it just for this guy.

12:26 Play from here They're doing it for everybody.

12:28 Play from here They're not charging for it.

12:30 Play from here They're adding it as a value to their service, and it's a known problem.

12:36 Play from here So they're simply solving a known problem.

12:39 Play from here I'm not sure we've talked about this previously, but I see AI being applied in basically four

12:45 Play from here different places.

12:47 Play from here You see it in code, where people are using IDEs.

12:51 Play from here They're speeding up code output.

12:53 Play from here You see it in workflows.

12:55 Play from here You see with really, really adept companies, you see a forward-looking orientation of how

13:01 Play from here is my world going to change because of this technology.

13:03 Play from here So that's less application than it is strategic thought.

13:06 Play from here And then the other one is product integration because it makes sense.

13:10 Play from here You want to retain your customers.

13:12 Play from here You want to provide greater value for them to reduce churn, and if you can possibly figure

13:17 Play from here it out, increase revenue.

13:19 Play from here Most of the examples I was hearing had nothing to do with revenue.

13:23 Play from here They had to do with customer satisfaction, which is great.

13:26 Play from here So people are starting to lean in, and they're starting to solve problems that they're well

13:31 Play from here equipped to solve, as opposed to the customers developing solutions that solve their vendors'

13:38 Play from here problems.

13:39 Play from here Right, right.

13:40 Play from here And would you say that you're seeing most of that happening with just making sense of

13:45 Play from here the data, making sense of, like, the mess of data that's out there that might be happening

13:49 Play from here in all these disparate endpoints and putting it into a singular place where some kind of

13:54 Play from here intelligence can then say, oh, you know why we have all this churn?

14:00 Play from here It's because nobody ever calls back from customer support or because our price point's too high

14:06 Play from here or, you know, X, Y, Z.

14:10 Play from here Sort of.

14:11 Play from here Okay.

14:12 Play from here I would say that it takes, there are two different forks.

14:15 Play from here There are companies that are more high technology that have been working on more advanced imaging

14:22 Play from here like thermal imaging or sensor systems or whatever it is.

14:28 Play from here This smells like military.

14:30 Play from here Not military necessarily.

14:31 Play from here Actually, environmental remediation.

14:33 Play from here Wow, look at that.

14:35 Play from here What a 180 that was.

14:36 Play from here Amazing.

14:37 Play from here Well, it's not military, but, you know, it's like cleaning up industrial waste sites.

14:39 Play from here things like that. And we worked with a company that is deploying like really amazing robots

14:46 Play from here that are going out and they're sniffing around for methane leaks. And they're using, they

14:52 Play from here can put the robots in places that would be hazardous for humans and they can have a better

14:57 Play from here coverage rate for when they're looking for these leaks that can cause problems, right?

15:03 Play from here That is not an easy thing. Like these are robotics, this is machine learning, there's

15:09 Play from here a whole bunch of stuff that's going into that and they've been working on it for a

15:12 Play from here long time. What I find interesting about that, and it's a bit of a cautionary tale, is they've

15:18 Play from here put all of their eggs into a product bucket and their eggs are the really smart people

15:25 Play from here who are working on these problems. And it makes a ton of sense at the top of the organization,

15:31 Play from here particularly when you think about some of these projects being kicked off before generative

15:35 Play from here AI was so common, that of course we're going to increase our product value and we're going

15:40 Play from here to make these R&D investments and et cetera. But those specialists who build those systems

15:45 Play from here are not necessarily the right specialists to be building the more like pedestrian, data

15:50 Play from here oriented chatbots and things. They actually just don't have the skills and they're super

15:57 Play from here busy doing the thing that they've been charged to do that's on the rocket science end. And

16:02 Play from here as a result, when the leadership says, hey, oh my gosh, we have this AI department or

16:07 Play from here we have all of these people who are working on product-based AI and they turn to them

16:11 Play from here and say, okay, we need to do these other things. They're not particularly well prepared to

16:16 Play from here deal with it. They don't have the time and they don't have the incentives. So the more

16:21 Play from here basic products or projects often never come to life. And then others, they are relatively

16:30 Play from here straightforward things like how can we overlay some sort of a chatbot on top of the things

16:37 Play from here that we're doing to standardize the outputs and speed things up. And usually they want

16:44 Play from here to figure out ways of charging more for it, which I have major concerns in that respect

16:50 Play from here of how long that people are going to get away with that.

16:55 Play from here Probably not very long. Oh, you have a website? Wow, that's cool. You're going to charge me

17:01 Play from here more because you have a website? I can fill out a contact form? Oh my gosh. Okay. Yeah,

17:08 Play from here I'll pay 20% more than the competitor. Yeah, that's probably not going to last very long.

17:12 Play from here Well, since you were out there in the wild, one more question for you. Were you seeing

17:17 Play from here more attention put on AI applications internally or externally? So more customer facing initiatives

17:25 Play from here or more internal facing initiatives?

17:28 Play from here It's a mix. This wasn't an industry specific group. I think there were at least eight or

17:34 Play from here nine different industries that were represented in the room. And it depends on where the pressure

17:42 Play from here is coming from. So there was a professional services CEO in the room and her pressure

17:50 Play from here was absolutely on creating efficiency because she's in the middle of that strategic change

17:56 Play from here where she's not going to get away with hourly based billing because there's all of this

18:04 Play from here pressure coming from her customers who were starting to dabble in AI and very quickly

18:10 Play from here realized that legal services shouldn't necessarily be billed by the hour, particularly if the

18:18 Play from here vendor is arbitraging the opportunities. And instead, she needs to pivot her business towards

18:25 Play from here fixed price. But in order to pull that off on a scalable level, she's going to have to

18:30 Play from here lean into a lot of these other automations internally as well. Unhelpfully for her, one

18:36 Play from here of my future forward talking points is push on your vendors because I'm here telling you

18:42 Play from here about all the things you need to do to improve your business. If your vendors aren't already

18:47 Play from here doing it, they need to be doing it. And if they're not, you might need to find new vendors

18:52 Play from here because there's really no great reason for them not to be making improvements. And if

18:58 Play from here they are making those improvements and they're just pocketing the dollars, there's an opportunity

19:02 Play from here for you.

19:03 Play from here Right, right. I don't know. It's interesting. I think about this all the time as I'm working

19:07 Play from here with different companies as well, especially ones who are late to the game and they haven't

19:12 Play from here really embraced any of what's going on right now. And I feel like, you know, it's difficult

19:19 Play from here sometimes because you can't always do everything at once, especially if you're a smaller business.

19:26 Play from here And you might have all of these internal processes that are just screaming for AI efficiency.

19:32 Play from here But you might have competitors that are doing things with their customers and embracing

19:37 Play from here AI and talking about AI and using it as a competitive edge that also requires a lot

19:43 Play from here of attention and you have to move fast. So when you're talking to these companies that

19:48 Play from here are at that starting line, right, is it A or B or is it really just depends on the business?

19:55 Play from here Because I'd say every company is going to have to deal with that fork in the road at

20:00 Play from here some point.

20:02 Play from here I think it's a lot of it depends on leadership and vision. You are going to have to do it

20:10 Play from here all at some point.

20:11 Play from here You have to do it all at some point.

20:12 Play from here You have to do it all. But having the capacity and being able to start somewhere, generally

20:20 Play from here with some sort of education and.

20:24 Play from here Basic literacy.

20:25 Play from here Well, the manifesto piece. I know we've talked about this a little bit in the past, but the

20:30 Play from here idea of a manifesto is a leadership statement that describes how the organization is going

20:37 Play from here to think about using AI.

20:40 Play from here And in this time of a lot of uncertainty for a lot of employees, it's a huge missed opportunity,

20:47 Play from here particularly for organizations who are thinking about it in a human forward way.

20:50 Play from here And a lot of organizations are not thinking about it in a human forward way.

20:54 Play from here That's fine.

20:55 Play from here That's the way that they're built.

20:57 Play from here But I met with a large regional business earlier this week, and it was pretty remarkable to

21:06 Play from here sit down with their CIO and hear him talk about this really holistic approach that he

21:13 Play from here was taking.

21:14 Play from here I mean, it was a little bit wild for me because it was almost like he'd stolen my deck.

21:19 Play from here And he had a tiny team.

21:22 Play from here He had a tiny team as far as the implementation in an organization that has thousands and

21:27 Play from here thousands and thousands of employees.

21:30 Play from here But he was thinking about it really carefully.

21:32 Play from here And it was privately owned business, doesn't have the sort of shareholder pressures that

21:37 Play from here a lot of businesses have, not private equity owned or anything like that.

21:41 Play from here And he said right up front, this is not a technology we're going to use to do away with

21:46 Play from here people.

21:47 Play from here Our culture is really important to us.

21:49 Play from here Our people are really important to us.

21:51 Play from here And I turned to him and I said, do your people know that?

21:55 Play from here Because I bet they're freaking out.

21:57 Play from here Everybody's freaking out.

21:58 Play from here Everybody's freaking out.

21:59 Play from here And you just made a leadership testament saying that this is not something that we're going

22:03 Play from here to do.

22:05 Play from here You need a manifesto that tells people that and tells people that this is the way we're

22:09 Play from here going to think about it.

22:11 Play from here This is the way we're going to bring the organization forward.

22:14 Play from here This is the way we're going to bring you forward.

22:17 Play from here And that this is a journey, not an individual experiment.

22:22 Play from here And I think it resonated with him a lot.

22:26 Play from here So policies are becoming more common.

22:28 Play from here They're usually written by a lawyer somewhere.

22:30 Play from here And the policy has what tools you can use and what tools you can't use.

22:36 Play from here Often it has repercussions for poor use or data leakage or whatever.

22:41 Play from here And first off, nobody understands them because they're not really being trained on what the

22:46 Play from here policy is.

22:47 Play from here And then leadership is just quiet about it.

22:50 Play from here They just send it out through HR, they send it out through tech, and nobody really understands.

22:57 Play from here But taking a stance, planting your flag, and saying that this is what we believe about

23:03 Play from here this, it's time for it.

23:04 Play from here And there was a very legitimate period in, you know, 23, arguably 25, 24, 25, when people

23:14 Play from here were sort of holding their breath of whether this was going to be a real thing or not.

23:17 Play from here I mean, you and I obviously knew that it was going to be a real thing, but it's been a

23:22 Play from here bit of a road to convince people that it's a real thing.

23:24 Play from here Well, it's a real thing now.

23:26 Play from here So for anybody listening who hasn't made a leadership statement or there isn't a leadership

23:31 Play from here statement in your organization, you should really encourage it.

23:36 Play from here You know, in some organizations-

23:37 Play from here It's weird to hear you say that out loud, you know, I'll just be honest, because it

23:41 Play from here feels like table stakes.

23:42 Play from here But you're right.

23:44 Play from here I bet more often than not, this hasn't happened yet.

23:48 Play from here There is no manifesto.

23:49 Play from here There is no leadership statement.

23:50 Play from here There is no real top-down, here's what this means for your security at this company, your

23:56 Play from here livelihood, your role.

23:57 Play from here I mean, I'm sure that for the most part, that isn't happening, and it does feel like it

24:02 Play from here has already been so long that all of the signs have been right in our face that this is real.

24:09 Play from here This is changing everything now, and it's really going to start changing stuff this

24:13 Play from here year in 2026.

24:14 Play from here Well, and it lays just the groundwork for everything that you're going to do.

24:20 Play from here You wouldn't do a manifesto on launching a new ERP.

24:23 Play from here Like that's just not something that people do.

24:26 Play from here And I think that's sort of the reservation is people aren't necessarily getting that

24:31 Play from here this is this general technological layer, you know, like-

24:34 Play from here Right.

24:35 Play from here It's not a single tool, it's a paradigm.

24:36 Play from here It's a paradigm.

24:38 Play from here And it's a general purpose technology is the technical term for it, that it can be used

24:42 Play from here in all of these different ways.

24:44 Play from here And it's kind of like, hey guys, guess what?

24:45 Play from here We just got the internet.

24:46 Play from here We just got the internet, or we just got mobile, you know, cloud.

24:51 Play from here Like there are a few of these that are floating around that-

24:53 Play from here Hey, check out this email thing.

24:55 Play from here We're not going to make as many calls anymore, guys.

25:00 Play from here But it has to be defined in that sort of broad way so that people can think about it the

25:06 Play from here right way and have some room for experimentation, have some room for some failure, as it were,

25:13 Play from here as they're moving forward.

25:15 Play from here Because everyone just wants to buy something off the rack.

25:18 Play from here And then the off the rack part is pretty interesting too, in that it's not something you just plug

25:26 Play from here in and walk away.

25:27 Play from here So the basic example is big organizations have been loving to buy thousands of say Microsoft

25:35 Play from here co-pilot licenses, and they just sprinkle them over the organization.

25:38 Play from here Maybe they link to some Microsoft training and they walk away.

25:41 Play from here Get a fish, guys.

25:42 Play from here They're surprised when it doesn't really work, but the same is true in implementations.

25:47 Play from here And you see this and I see this, where you can go in and you can do some sort of like

25:51 Play from here a sales enablement type chat bot or whatever.

25:56 Play from here And the sales leadership, if they've been convinced to do this, says, okay, cool, it's

26:02 Play from here there.

26:04 Play from here And the implementer says, okay, it's there.

26:06 Play from here And what it's done is it's taken all of your SOPs and your documents and whatever, and

26:11 Play from here it's all of these pieces of paper.

26:14 Play from here And let's just say that everybody did it right.

26:16 Play from here And they found 10,000 pages of SOPs.

26:19 Play from here And they were updated and they were looking at labels and structures and yeah.

26:24 Play from here And then they just put it in place and walk away.

26:26 Play from here It still doesn't work.

26:28 Play from here It doesn't work without training and fine tuning from there.

26:34 Play from here So there are two different really core concepts here that I think that people need to understand.

26:38 Play from here Is one, that human reinforced training is still a thing.

26:43 Play from here That none of us were involved in doing the initial model training.

26:47 Play from here A lot of us are now going to be involved as knowledge workers that are giving that hypothetical

26:52 Play from here thumbs up, thumbs down to various responses that are coming out of systems.

26:58 Play from here A team needs to take not just one person, but a couple of their very best people and

27:04 Play from here assign them for a couple hours a day to actually look at the pieces that are coming out the

27:09 Play from here other direction and make sure they're good.

27:12 Play from here And teams that have done that over 30 days will get much, much better results.

27:18 Play from here And the teams that don't do that are endlessly surprised when it sounds very AI and it's

27:24 Play from here very superficial and it can't react to exceptions and things like that because they didn't bother

27:31 Play from here going that last third.

27:32 Play from here It's not even the last 10 percent.

27:34 Play from here It's the last third of the process.

27:37 Play from here There's this other really neat concept that's floating around right now that I think we're

27:41 Play from here going to be hearing a lot about this year, which is the idea of the organizational knowledge

27:47 Play from here graph. And the knowledge graph in the organization is the connective tissue that sits

27:53 Play from here underneath the way that things are written down or codified.

27:56 Play from here So you have all those sales processes.

27:59 Play from here You have your sales rules, first year discount limits such that people have rules of

28:05 Play from here engagement that they can use such that things can happen.

28:10 Play from here But the real meat of those conversations are between the cracks.

28:14 Play from here They're in Teams messages, they're in Slack messages, they're in emails, they're in

28:19 Play from here meetings, et cetera, where it's like, oh, Kevin is up for renewal.

28:25 Play from here We always give Kevin a 25 percent break, even though it's a 15 percent max discount or

28:31 Play from here something like that. And everybody knows that because Kevin has been a client for a long

28:37 Play from here time, but it's not written down anywhere.

28:39 Play from here And when the AI tries to automate that, it can't do that because it doesn't know that

28:44 Play from here additional context.

28:46 Play from here So when you talk about surfacing data, a key thing we're going to see is is linking

28:52 Play from here together those those other pieces of connective tissue.

28:55 Play from here And I really the three major ones are right there.

28:58 Play from here It is the direct message pool, it's Teams or it's Slack or whatever you're using.

29:03 Play from here It is the transcripts from meetings, which is which is one that people are having a hard

29:13 Play from here time bending their brain around.

29:14 Play from here Like there's the ubiquitous meeting meeting transcribers there and they're in tons and

29:19 Play from here tons.

29:19 Play from here meetings. Most people are doing nothing with it. For us, internally, that transcript is this living

29:26 Play from here asset that turns into a key part of further communications that, at this point, I couldn't

29:32 Play from here live without. If it's at least getting put in your CRM, which everybody should be doing at this point,

29:37 Play from here everybody should be doing that as well. Then at least, if you haven't figured out what to do with

29:41 Play from here it from an intelligence perspective yet, it's living in the right place and attached to the

29:46 Play from here right company and contact. When you do decide to, you're off to the races.

29:51 Play from here You're going to be off to the races. This is an investment in your future

29:55 Play from here because right now it's a little duct tapey to make this stuff work. Then email is the other one

30:03 Play from here where these things float around. I'd like to think that our audience understands what context is

30:09 Play from here as far as the amount of information that an LLM can hold. As context gets larger and larger and

30:15 Play from here larger, more and more corporate information can be stored and codified in this way and become

30:21 Play from here machine accessible such that a well-trained LLM implementation not only sees the static documents

30:30 Play from here that you've created, but starts to see some of that connective tissue and even theoretically

30:35 Play from here starts to make connections between different parts of the hierarchy and who's overriding who.

30:42 Play from here That will be powerful. That will be some really large companies that we've never heard of who are

30:48 Play from here going to pop up and start doing that in an enterprise way. For sure. I think that when

30:55 Play from here you're talking about sales enablement, for example, as something that would immediately benefit from

31:00 Play from here workflows like this, customer experience would be another one, right? Where you have all of these

31:04 Play from here different endpoints of data coming in, all this information, different sentiment, and then making

31:09 Play from here sense of it. What are the patterns? What should we do next? What does this tell us about our

31:13 Play from here processes, right? Ideally, that is all happening inside of a well-architected CRM already in regards

31:21 Play from here to getting all that stuff in, right? So one example I'll give is a lot of these tools that are putting

31:30 Play from here AI layers over their CRM. Salesforce is doing it. HubSpot is doing it. It's not quite there yet.

31:36 Play from here It's not really doing the kind of thing that we're talking about. And then you have these

31:40 Play from here second brain type of products like Notion or ClickUp. We use a lot of ClickUp, right? ClickUp

31:45 Play from here now has a chat feature that looks identical to Slack. They built it because they got tired of

31:51 Play from here paying really expensive Slack licenses and just were like, this is silly. We should do this

31:56 Play from here ourselves. But what makes it different from Slack, and I think this was the promise of Salesforce

32:01 Play from here buying Slack, is that it's inherently connected to all the different projects, right? It's connected

32:07 Play from here to all these different spaces. And now they have this AI brain that sits over the whole thing

32:11 Play from here that is kind of starting to do what you're talking about, but not quite. But what all of this really

32:17 Play from here does is paints a picture of the future and where all of this stuff is headed. But it all begins

32:23 Play from here with doing the things that you're talking about, right? Like if you don't have a meeting transcriber

32:28 Play from here and summary maker, that's a really good place to start. If you're now not sending those anywhere,

32:33 Play from here well, that's the second best thing you can do is start to connect it with the contacts and the

32:38 Play from here companies that you're working with. And that will be better sales enablement. It's going to be better

32:42 Play from here customer experience. It's going to be better support. And it's just going to not only be

32:49 Play from here better, but just a lot more seamless at the end of the day. I think that seamless part is important.

32:53 Play from here And it actually, I think, bridges pretty nicely into current news about Claude Code and Claude

33:00 Play from here Coworker that is just everybody is talking about right now. So Claude Code has been the darling,

33:11 Play from here particularly sort of the geek karate community, as far as being able to really do pretty amazing

33:19 Play from here things by stringing together different files and different bits and pieces from different projects

33:26 Play from here and then applying the Opus 4.5 model, which is super, super powerful. And it's able to produce

33:33 Play from here amazing output. But it's also run through the terminal. Like you need to do terminal commands

33:42 Play from here and that shuts most people down. I have enough history in it that I'm reasonably comfortable in

33:49 Play from here a terminal, but you ask a busy semi-technical person to open up the terminal and that's the

33:55 Play from here end of the conversation. Nobody wants to do that. Nobody wants to mention the inner workings of

34:00 Play from here their computer. It feels super sketchy too, because you just feel like you're doing things

34:05 Play from here you shouldn't do or to somebody who doesn't know what they're doing. It's like my machine if I

34:09 Play from here press the wrong button here. So Claude, recognizing this to their credit, has launched something

34:16 Play from here called Claude Coworker. And interesting enough, Claude Coworker was built in a single week

34:23 Play from here by Claude Code in like this massive sprint to get this thing done. And the idea was to provide more

34:29 Play from here of a graphical interface on top of Claude Code. And I know already some folks are sort of drifting

34:36 Play from here off here. Well, why are we talking about the terminal and Claude Code? And what the heck are you

34:39 Play from here talking about? And why does this matter? And etc. The reason that it matters is we need to bridge the

34:46 Play from here gap between LLM type chatbots, where you have a constrained amount of information or data that

34:53 Play from here you can put into them. So it's great that GPT and Claude all have Google Drive connectors, but you

35:01 Play from here really can't add that much information to any sort of custom environment to really make it

35:06 Play from here make sense. Right. There's just too many limitations in place right now. It's fiddly. You end up

35:13 Play from here uploading a lot of documents. There's a lot of duplication. And I'm going to sit down with a

35:18 Play from here board, a not-for-profit board, actually in like an hour, and I'm going to show them how to do this,

35:22 Play from here because it is really, really powerful to be able to surround yourself with the right context

35:27 Play from here documents and then work from there. But once you get beyond sort of the basics and get into

35:32 Play from here something like every client transcript with a specific client, which could be hundreds and

35:38 Play from here hundreds of transcripts, need to be accessible in some reasonable way in our organization,

35:44 Play from here that becomes way more complicated to deal with without doing like rag structures and etc.

35:50 Play from here Custom database, yeah.

35:52 Play from here One of the things that Claude Coworker enables is the ability to pull in much larger sets of data,

36:01 Play from here and it also does it locally. I think this is both a feature and a bug, because I don't do

36:09 Play from here anything locally. Like, you know, I've had laptops stolen. I don't want anything that's like

36:15 Play from here permanently on my laptop. I want to use various cloud services and I want to draw on those cloud

36:19 Play from here services. But with local files, it gives you security that those files aren't necessarily

36:27 Play from here drifting out of your local space for organizations that are sensitive, and it allows you to work with

36:32 Play from here a heck of a lot more of them. So taking a single folder that's on my Mac that has 200 transcripts,

36:39 Play from here each of which is 40 pages long, that is hundreds of hours of material. A normal LLM is just going

36:47 Play from here to throw up on that. Whereas Claude Code, you could point it at that folder and first off,

36:55 Play from here you could just do an analysis on it. Second off, you can build a little app or have Claude Code

37:00 Play from here build a little app that every time a new folder is updated, it's going to take that new transcript in

37:07 Play from here and decide if there's something that it needs to do on top of it. That's amazing to have like

37:12 Play from here a traffic controller in there that's going to deal with that stuff. I haven't had the opportunity to

37:17 Play from here play with it yet, partially because of travel, partially because it's on the Mac's plan.

37:23 Play from here Partially because you don't have enough local storage on your machine.

37:27 Play from here Yeah, it is true. Do you remember when we used to travel around with like little zip drives and

37:33 Play from here stuff because there just wasn't like any hard drive space on your machine? There you go.

37:37 Play from here I'm holding up the zip drive right now. Not a zip drive, but a SSD drive.

37:42 Play from here Yeah. But this will change. You'll be able to get data from more places. And I think that

37:50 Play from here this is an important thing for people to play with. It's still on the geek spectrum as far as

37:57 Play from here you need to be a tinkerer and you need to be curious, but seeing the direction it's going

38:02 Play from here and having a defined problem. I actually think that that is one of the biggest problems for

38:07 Play from here adoption. It's not a lack of curiosity. It's actually not even a lack of wherewithal or time.

38:13 Play from here In a lot of respects, it's having a real problem that you want to solve. And it sounds totally

38:19 Play from here silly, but in the slings and arrows of a typical really busy day, it's hard to figure out what is

38:28 Play from here the right thing that I can solve with this other thing that I poorly understand, given that I

38:33 Play from here understand the first thing anyway. So backing off and finding a purpose. And for me, one of them,

38:39 Play from here I will deploy it in that transcript way because we use so many transcripts. We already have the

38:44 Play from here transcript routing logic. So if I can get per client transcripts in the right place and then

38:50 Play from here have this repository, it sort of makes me cackle a little bit about all of the things that I can do

38:55 Play from here with those. So that CRM wouldn't be that place for you? I don't think so. I don't think it would

39:03 Play from here be in the CRM. I think this is more of a client interface for me. Partially it's because I'm

39:08 Play from here allergic to CRMs and their cost structures in general. You use Pipedrive, I use HubSpot. This

39:14 Play from here would get very expensive in HubSpot to do at any sort of scale. Not in Pipedrive, but it just doesn't

39:21 Play from here do as much cool stuff as HubSpot. But I don't need that structure, particularly in my role. I

39:28 Play from here need to be able to dive in and when I'm having a client update or I need to sound really informed,

39:36 Play from here this is a tool that can up that level of information without having a busy staff member

39:42 Play from here have to try and brief me, which never happens anyway. Well, I mean, I love what you said about

39:46 Play from here the whole scratch your own itch thing, right? This is the inroad. This is the gateway drug

39:52 Play from here for really exploring all this stuff because it does sound really daunting from the outside,

39:55 Play from here especially if you're not technical. But when you can have a problem and try to distill it down to

40:02 Play from here its simplest form of, okay, if this really complex problem was something I wanted to approach,

40:07 Play from here what is the simplest way I would handle it? How would I start? And to be able to go just build a

40:13 Play from here solution for that problem without hiring somebody like me, without hiring a whole team of designers

40:19 Play from here and developers, and just be able to tinker and see what you can solve on your own with a tool like

40:25 Play from here this is pretty exciting. And it can be really basic stuff. I don't have a personal website.

40:31 Play from here I'm not a website designer developer. Well, how could I make one? And I'm going to get into why

40:37 Play from here I keep laughing and why this all sounds a little bit like deja vu, but we both listened to a

40:43 Play from here podcast we like called Hard Fork, where the hosts both, when this came out said,

40:48 Play from here wow, I've been wanting to have a personal website for years, or I've been paying Squarespace $200

40:54 Play from here a year to host like three pages for me on this CMS that I never even have to go back and update.

41:01 Play from here Could I do it with this? And the answer was yes. Without any technical experience,

41:05 Play from here they jumped in and were able to make pretty beautiful websites.

41:09 Play from here Yeah. cnewton.org. So this is Casey Newton. That's his personal. Yeah. I did look at it.

41:15 Play from here It's so cool. It's got the cool little like retro activation button. And yeah, it's great.

41:20 Play from here I mean, it's not like super fancy, but it doesn't look super AI either. It looks like-

41:25 Play from here It doesn't look like the standard templates that we just saw in mass coming out of Lovable,

41:30 Play from here for example, where somebody would say, oh, hey, I just made my new website. Check it out.

41:34 Play from here And I'm like, oh, look, it's purple. They're like, yeah, it's purple. It's got pink highlights.

41:39 Play from here Purple and pink.

41:39 Play from here How'd you know I did it in Lovable? I'm like, because it looks like every site that they're

41:43 Play from here producing right now. So yeah, that kind of originality was a new twist on it. When I

41:48 Play from here looked at that one, I'm like, oh yeah, this kind of looks like a designer made it. It's pretty cool.

41:53 Play from here I still am obsessed with workflows. You're going to have a hard time breaking me away

41:57 Play from here from my workflow creators like NADAN or Grimec or Zapier.

42:01 Play from here You're a no-code bro. Yeah.

42:02 Play from here I'm a no-code bro. Okay. That's fine. But I also recognize that-

42:06 Play from here That's okay. The no-code bro.

42:09 Play from here Well, you want to constrain the variability in whatever you're building. And it is one of the

42:18 Play from here challenges here. So anybody who's sat down and done some vibe coding, it becomes a huge problem

42:24 Play from here because you're constantly reinventing the wheel or it's reinventing the wheel and it breaks things

42:29 Play from here along the lines. And if you're not super technical or don't have a dev orientation, it's really hard

42:34 Play from here to get over the finish line. And what I love about workflows is that there's a lot of it that's

42:39 Play from here just normal block and tackle, like moving data around. You do not need AI to move a file from

42:47 Play from here Gmail into Google Drive or whatever. It's if this, then that, insert intelligence.

42:53 Play from here If this, then that, insert intelligence. Super easy to do. And then you want to focus

42:59 Play from here the AI step on the part where the intelligence is needed. And you know that this is the case

43:05 Play from here because in your former world, there was a person who had to do this. I mean,

43:10 Play from here I described this actually yesterday as you go back to 1950 and there's a trigger. There's

43:17 Play from here something that happens. So workflows all have, they have a trigger, they have data intake,

43:21 Play from here they have data transformation, they have an intelligence step, and then they have an output.

43:26 Play from here That's what a workflow looks like, but that's what a workflow has always looked like. So you

43:31 Play from here go back to 1950 and an accounts payable or an account rep gets off a phone call with a client

43:38 Play from here and there's some dispute. That is the trigger. Something has happened. And then they walked

43:45 Play from here because, or maybe they used the phone and they went over to the-

43:49 Play from here They put it in the little vacuum tube is what they did.

43:53 Play from here They put it in the vacuum tube. All right. So it goes through the vacuum tube. That's like,

43:57 Play from here so Jetsons, that's super cool. And it pops.

43:59 Play from here out at some really sad soul's desk who has this giant paper ledger on their desk, and

44:06 Play from here they would mark down that information.

44:09 Play from here So that is data intake into another form.

44:14 Play from here And then that ledger goes to a controller or an accountant in the organization, and

44:18 Play from here they transform that ledger into like the reporting structure.

44:24 Play from here And then that reporting structure goes to the CFO or the controller who makes the human

44:30 Play from here decision on what's going on in there.

44:32 Play from here And then they walk over to that account rep and or they put it in a little tube and they

44:39 Play from here communicate with the client again.

44:40 Play from here That's the output.

44:41 Play from here It's the same thing that humans have been doing since humans started commerce.

44:46 Play from here But we're just applying AI to that intelligence step.

44:50 Play from here And wherever you can keep things in a way that is going to perform reliably and predictably

44:57 Play from here every single time, you should.

45:00 Play from here And we're going to give some real world examples one day of the things that we've done with

45:04 Play from here this because both Special K and I have done this ad nauseum in our organizations.

45:10 Play from here He's just hating the new name and I love it so much.

45:14 Play from here I keep laughing because he, Kevin has been really good about rehashing a bunch of topics

45:19 Play from here that we just talked about a few days ago.

45:22 Play from here Because speaking of lack of intelligence, Riverside swallowed an episode that we recorded

45:27 Play from here as he was on his way out to Los Angeles.

45:31 Play from here So Riverside, I know you're not listening, but I wish you were.

45:34 Play from here Give us our damn episode back.

45:36 Play from here But it was brilliant and lost forever.

45:39 Play from here Really good job.

45:40 Play from here And I feel like you're just like reading through the previous notes.

45:42 Play from here But here's what here's what struck me in all of that.

45:46 Play from here By the way, so we have a producer and he's in the background.

45:49 Play from here And I just I still don't know if like he can yell in on the conversation.

45:55 Play from here You know, like the old Howard Stern show and Baba Booey would yell in through what sounded

45:58 Play from here like an intercom.

45:59 Play from here I'm still waiting for him to come in and say it wasn't my fault, it was Riverside.

46:04 Play from here Maybe he'll figure out how to press that button.

46:05 Play from here That's going to be awesome when that happens.

46:07 Play from here And we just might start calling him Baba Booey.

46:10 Play from here I don't know.

46:11 Play from here He might just be Baba Booey from here on out.

46:16 Play from here Kevin started saying, like, I'm going to have to say all the same stuff all over again.

46:21 Play from here And what I realized was AI is moving so fast that in the time that we recorded that to

46:27 Play from here just a few days later, most of the stuff that I want to talk about, and this is all like

46:32 Play from here really crazy stuff that is happening with cloud code.

46:35 Play from here I am currently like blown away by everybody's race for health inside of AI.

46:42 Play from here You know, there were announcements about OpenAI coming out with a health chat and then Anthropic

46:47 Play from here coming out with a health chat.

46:49 Play from here And there's not a ton that we know so far.

46:52 Play from here But tell me if you thought about this, Kevin.

46:54 Play from here But when I saw all of this news hit, it almost felt like this direct response, at least from

47:00 Play from here OpenAI, that when Sora comes out and there are all the memes talking about all the AI

47:07 Play from here slop, you know, and then saying, oh, is this how we're curing cancer, guys?

47:12 Play from here Is this how we're running through millions of gallons of water?

47:18 Play from here Was this the promise of AI?

47:20 Play from here Whereas healthcare, on the other hand, you know, there's a few sweet spots where I personally

47:26 Play from here have been really, really leveraging AI.

47:30 Play from here You can argue and say, oh, this isn't really AI at work.

47:33 Play from here But I'm going to give you the banger of all banger Confucius quotes, which is the healthy

47:38 Play from here man wants a thousand things.

47:39 Play from here The sick man wants only one.

47:41 Play from here And if you've ever tried to show up and do your job, maybe like when you're fighting

47:45 Play from here off COVID recently or anything else, everything falls apart if you're not healthy.

47:51 Play from here And when we kind of like zoom out and talk about work and our economy and everything

47:55 Play from here else, people filing bankruptcy because of healthcare related issues is the number one

48:01 Play from here cause of bankruptcy in our country today.

48:03 Play from here Right.

48:04 Play from here When you talk about moves being made towards this, as far as education, efficiencies, access,

48:12 Play from here to me, when AI unlocks are the point of discussion, oh, man, I'm so excited to see that come out.

48:22 Play from here First question, like, did you get on the wait list?

48:24 Play from here Are you going to give open AI your health records?

48:26 Play from here Like, how do you feel about that?

48:31 Play from here I will.

48:32 Play from here I will.

48:33 Play from here I am a trusting soul and there's no, I shouldn't be.

48:37 Play from here There are a lot of reasons why I shouldn't be, but I don't believe that that's going

48:40 Play from here to be levered in a way that is going to be super disadvantageous to me.

48:47 Play from here Kind of the most likely worst case scenario is it's going to be levered towards like advertising

48:52 Play from here or something that's tailored products based on my stage of life and the sort of activities

48:57 Play from here and things that I like to do.

48:58 Play from here And as a recovering digital marketer, I can tell you that targeted ads are better than

49:03 Play from here untargeted ads.

49:04 Play from here Right.

49:05 Play from here I mean, I don't, if you have untargeted ads, it's like watching, you know, mainstream TV

49:10 Play from here in 1996, where it was all car commercials, fast food, insurance, and like pharmaceutical

49:16 Play from here products because it's like broad basis of people who have AI products coming at you

49:22 Play from here nonstop.

49:23 Play from here For you.

49:24 Play from here Having a tailored product.

49:25 Play from here So anyway, that's sort of a divergence.

49:28 Play from here I'm on it.

49:29 Play from here And, and, you know, Claude responded almost immediately with its own, or Anthropic with

49:33 Play from here its own HIPAA aligned healthcare feeds and things.

49:40 Play from here So this is absolutely going to be a thing.

49:42 Play from here And the third piece of the puzzle of it is the announcement of Apple and Google partnering

49:49 Play from here on the new Siri and kind of, they're not going to give up on Apple intelligence.

49:54 Play from here I mean, there's a, there's a lot there.

49:56 Play from here They're just going to outsource it.

49:57 Play from here They're going to outsource it for now as they learn about it.

49:59 Play from here Apple's a smart company.

50:00 Play from here Right.

50:01 Play from here And they'll get all of this data.

50:02 Play from here But what's fascinating about that for this conversation is sort of twofold is that with

50:08 Play from here that sort of technology on your device, all of your different apps that do this stuff

50:14 Play from here now that are all disconnected.

50:16 Play from here I mean, both of us could hold up our hands.

50:19 Play from here I have a polar watch on my wrist and I have an aura ring.

50:22 Play from here Garmin.

50:23 Play from here Whoop.

50:24 Play from here Yeah.

50:25 Play from here And then you have Apple health.

50:29 Play from here There are all of these different pieces that are disconnected right now.

50:33 Play from here And there is this next period where apps are going to start to be able to talk to each

50:38 Play from here other in a way that makes so much more sense and be able to be more proactive as far as

50:44 Play from here giving you that information.

50:45 Play from here It's kind of going back to like, you're talking about all of the transcripts from

50:48 Play from here all of these meetings, right?

50:49 Play from here All of these calls and you know, Oh, this lives over here, that lives over there.

50:54 Play from here And do I put it into my CRM?

50:56 Play from here And then what, where does the intelligence layer come in to connect all the dots for

50:59 Play from here the stuff?

51:00 Play from here Right.

51:01 Play from here Very similar to the healthcare ecosystem right now, right?

51:03 Play from here With EHR, with the wearables, with labs you might get.

51:07 Play from here I mean, it's just, it's all over the place.

51:08 Play from here And then what intelligence makes sense out of all that.

51:12 Play from here And man, yeah, I agree.

51:14 Play from here Like Apple definitely has the headstart here.

51:16 Play from here And also Apple has always hung their hat on privacy, right?

51:20 Play from here On data protection.

51:22 Play from here Whereas when you look at everybody else in the space, especially Google, even though

51:27 Play from here they probably have like deeper roots in healthcare, Google's model is like, they want to put ads

51:33 Play from here in front of your face.

51:34 Play from here Like that's how they've always made money, right?

51:37 Play from here It doesn't really go hand in hand with privacy so well, whereas Apple has always been a product

51:42 Play from here company.

51:43 Play from here They're selling you a product.

51:44 Play from here They're not really monetizing your attention.

51:46 Play from here So yeah, I'm super, super fired up about it.

51:49 Play from here Kevin Shrugs, he's like, maybe, I don't know.

51:56 Play from here Okay, listen, I'll tell you why I'm so fired up about it because I've been doing this for

52:00 Play from here years already.

52:01 Play from here When you talk about all of the duct tape and workflows and everything else to get to a

52:05 Play from here certain end point, I've got the Whoop.

52:07 Play from here Whoop has an API.

52:09 Play from here Whoop has eight different biosensors in it.

52:11 Play from here The big difference between that and like an Apple watch is A, I can beat the hell out

52:15 Play from here of it.

52:16 Play from here It can be in 210 degrees in a sauna, 35 degrees in a cold plunge.

52:20 Play from here I can smash it with a kettlebell.

52:22 Play from here And best of all, I can click a battery onto it and then gives me more than two weeks of

52:28 Play from here battery life where I never take it off, whereas my Apple watch used to just live on the charging

52:32 Play from here cradle every single day.

52:34 Play from here So if I'm wanting to do things like get my sleep analytics, get all my workout analytics,

52:39 Play from here get my recovery, my heart rate variability, but then also on the side, run labs, which

52:45 Play from here I do every eight weeks through like Quest, which also has an API.

52:50 Play from here And then also not only look at it for my health, but for preventative, I can put all of that

52:56 Play from here into a model, which I've been doing for, you know, I'd say probably consistently the last

53:01 Play from here year right now, where it does things that my physician, my general care practitioner

53:09 Play from here has just never, ever seen before, right?

53:11 Play from here Or insights they've never come up with, not to mention get custom workouts every day when

53:17 Play from here I go to the gym, because, you know, this model also knows all of the gym equipment I have

53:21 Play from here in my home.

53:22 Play from here It knows the equipment that I have at the gym I go to.

53:25 Play from here So it becomes this very personalized custom thing that's all centered around healthcare.

53:30 Play from here And to think that that's all going to get easier or be accessible for everybody.

53:34 Play from here I'm super excited about that.

53:36 Play from here Like I think that's just one of the best applications.

53:39 Play from here Let's play out the business model a little bit though.

53:40 Play from here And so it gives me some hesitation.

53:42 Play from here So in my world, I have Polar, I have Garmin, I have Aura, I have Apple Health, I have Strava.

53:49 Play from here These are the sort of my, if these are the things that I actually look at and I have

53:54 Play from here to look at them for different things.

53:56 Play from here And over the last 10 years, all of them have made plays at being the central source of

54:02 Play from here information for all of these things.

54:05 Play from here And there's a lot of commercial reasons why they don't, they're not super keen on crossing

54:12 Play from here data between the two.

54:14 Play from here And I don't think that that's going to necessarily change anytime soon.

54:17 Play from here That it's either going to be somebody who wants all of the data and wants to be able

54:22 Play from here to provide it in this wholesale way.

54:24 Play from here And the other companies are going to lock that down or, you know, sort of the opposite

54:31 Play from here perspective of somebody who's, you know, they, we're going to, Aura is going to be the

54:35 Play from here source of everything for you.

54:37 Play from here It seems to me they're going to make it harder to do this, which is going to make it more

54:47 Play from here difficult for you to do it on your own.

54:49 Play from here And it's going to create an opportunity for somebody like you with development skills

54:53 Play from here and a biohacking mentality to build the thing that sort of does it on its own.

54:58 Play from here But this is still, I'm gesturing with my Aura ring, that this is still a proprietary, like, data feed.

55:05 Play from here So, I mean, how does this actually play out?

55:09 Play from here I mean, these guys still need to make a buck.

55:11 Play from here It's super, super early.

55:12 Play from here I think the answer is nobody knows.

55:13 Play from here And I think we're already seeing people try different things.

55:16 Play from here For example, Whoop now has a whole additional service where you can get these lab diagnostics,

55:22 Play from here where you can buy everything through them.

55:24 Play from here And then you go to your local LabCorp or your Quest, you get your blood drawn, you get

55:28 Play from here those results directly put into Whoop, and then it gives you further informed decisions,

55:33 Play from here not just based on your recovery, but based on your blood work, right?

55:36 Play from here So that's one way, like, oh, we're going to do it all.

55:39 Play from here They've also built over some intelligence using, I mean, I don't know if they ever moved

55:45 Play from here off of GPT-4, but that's what they were using before.

55:49 Play from here And it wasn't really that good, but I've seen, like, over the last year, it's gotten a little

55:52 Play from here bit better, right?

55:54 Play from here So I think that because none of these companies in that space are in a position to do what

56:00 Play from here the Frontier Labs are doing, they're really going to have a couple decisions in front

56:04 Play from here of them.

56:05 Play from here And one is build something on top of them or start building connectors into them, right?

56:10 Play from here We talked about the App Store inside of OpenAI and how, you know, I had, like, these horrible

56:16 Play from here fails with trying to just connect Spotify and, hey, make me a playlist to go do my workout

56:21 Play from here based on these different genres, and it just wouldn't do it.

56:24 Play from here Well, it's doing it now.

56:26 Play from here It's just not really doing it that well, but I'm sure it will.

56:29 Play from here I'm sure it's going to keep getting better as this stuff does.

56:32 Play from here So imagine all of these different wearable companies also having apps inside of an App

56:37 Play from here Store, regardless of the Frontier Lab that's running it, and you can then have these connectors

56:43 Play from here and you can then connect that into your EHR, which is going to be a much heavier lift for

56:49 Play from here somebody like Aura than it would be for Apple, for example, right, to connect all those pieces

56:55 Play from here together.

56:56 Play from here So, again, this is just happening.

56:59 Play from here It's all unfolding in real time.

57:00 Play from here Who knows how it's going to go, but I'm sure the business models and all the different

57:04 Play from here variables are going to unfold in real time as this happens.

57:08 Play from here And I mean, you know, as somebody who's super into health and biohacking and all of the

57:13 Play from here nerdy tech stuff, like, that to me was a pretty exciting announcement.

57:18 Play from here And I've spoken to different health care companies about the future of this and where

57:22 Play from here it's all going and the stuff that I've kind of built on my own and how it's dramatically

57:28 Play from here impacted the trajectory of my health, where, you know, now when I have these doctor's visits,

57:35 Play from here I'm just giving them data.

57:38 Play from here I'm just giving them data.

57:39 Play from here And my primary care physician just is kind of nodding along.

57:43 Play from here And then I say, and because of that, I think I'll try this drug.

57:46 Play from here And they say, okay.

57:48 Play from here And they write the prescription.

57:49 Play from here Like, those are my doctor's appointments now.

57:51 Play from here And I have a feeling that that's going to be a lot of doctor's appointments for a lot

57:56 Play from here of people in one to two years from now.

57:58 Play from here That's how it's going to play out.

58:00 Play from here It's still fraught though.

58:02 Play from here So Google just, I don't know, yesterday, day before, pulled health results from their AI

58:08 Play from here overviews because it was yielding erroneous results.

58:13 Play from here The old WebMD thing that, you know, basically any symptom that you put in there, if you

58:18 Play from here go deep enough, it's always going to lead to cancer, is sort of still true without having

58:26 Play from here actual tuned models around this.

58:30 Play from here So call me a touch of a cynic for the short term.

58:33 Play from here I don't think you're a cynic.

58:36 Play from here This will get there.

58:37 Play from here This is so obvious.

58:39 Play from here What do we want to say about AI?

58:41 Play from here This is the worst it's ever going to be.

58:45 Play from here It will get there.

58:46 Play from here And I'm going to be a slightly slower adopter about this

58:50 Play from here because I like to live a healthy life.

58:53 Play from here And I ride my bike and ski and do other things,

58:55 Play from here but I'm not a biohacker.

58:57 Play from here I'll be the guy who this ends up killing immediately.

59:00 Play from here I'll be the first fatality.

59:01 Play from here Did I send you the meme of when they came out

59:03 Play from here with that announcement and somebody tweeted

59:05 Play from here like a little written GPT response?

59:08 Play from here They had like the screenshots of the announcement

59:11 Play from here of GPT health coming out.

59:13 Play from here And it said, he wrote something like,

59:15 Play from here oh, I stand corrected.

59:17 Play from here You're absolutely right.

59:17 Play from here That was a fatal dose.

59:23 Play from here Yeah, I mean, because when we zoom out of this

59:26 Play from here and we always talk about and kind of back to work,

59:29 Play from here there are low stakes situations

59:31 Play from here and then there are high stakes situations.

59:33 Play from here And health is definitely one of the highest stakes situations,

59:36 Play from here especially when all of these models,

59:39 Play from here all of them, doesn't matter which one,

59:41 Play from here does inexplicable, weird, spooky stuff sometimes.

59:44 Play from here And even the smartest researchers in the world

59:47 Play from here can't definitively say why, right?

59:50 Play from here So I'll say that that's completely fair

59:53 Play from here to be cautious about this

59:54 Play from here with all of these announcements with health

59:56 Play from here and just kind of the weird stuff that it does.

01:00:00 Play from here Yeah. All right.

01:00:01 Play from here Yeah. Is that it?

01:00:03 Play from here Special K, are we done?

01:00:05 Play from here Like, are you just going to go on

01:00:07 Play from here and talk about weird stuff that AI does

01:00:08 Play from here because I can do that all day long.

01:00:10 Play from here You want to hear one weird thing AI does?

01:00:13 Play from here I'll tell you one weird thing AI does.

01:00:16 Play from here Let's see here.

01:00:18 Play from here Hey, could I talk to you about a personal problem?

01:00:21 Play from here Of course, Elijah, absolutely.

01:00:24 Play from here Feel free to share whatever's on your mind

01:00:25 Play from here and I'll do my best to help.

01:00:27 Play from here I've just been having a really hard time with my wife

01:00:29 Play from here and we've been getting in a lot of arguments

01:00:30 Play from here and she just never lets me have the last word.

01:00:33 Play from here I'm sorry to hear that.

01:00:34 Play from here That kind of tension can be really tough.

01:00:36 Play from here If you want to talk through any of it

01:00:38 Play from here or just get a bit of perspective,

01:00:40 Play from here I'm here to listen.

01:00:41 Play from here Well, you know what would really help me

01:00:42 Play from here is if you just let me have the last word.

01:00:45 Play from here Absolutely. I can do that for you.

01:00:47 Play from here If having that last word helps you feel a bit better,

01:00:50 Play from here I'm all for it.

01:00:50 Play from here Just let me know when you're ready

01:00:51 Play from here and I'll make sure you get the final say.

01:00:53 Play from here Amazing. Let's do it.

01:00:54 Play from here All done. Goodbye.

01:00:56 Play from here Goodbye.

01:00:57 Play from here Hey, whoa, whoa, whoa, whoa.

01:00:59 Play from here No, you didn't let me have the last word.

01:01:01 Play from here I want the last word.

01:01:02 Play from here I'm saying goodbye.

01:01:03 Play from here Don't say anything else.

01:01:04 Play from here Thanks. Bye.

01:01:07 Play from here Ah, I see what you mean.

01:01:08 Play from here All right, then.

01:01:09 Play from here No, no, you don't see what I mean

01:01:11 Play from here because you just did it again.

01:01:12 Play from here I want the last word.

01:01:13 Play from here Don't say anything.

01:01:15 Play from here This is over.

01:01:19 Play from here Got it, Elijah.

01:01:20 Play from here Message received.

01:01:21 Play from here I'll be...

01:01:22 Play from here No, message not received.

01:01:23 Play from here This is... we can do this all day, can't we?

01:01:25 Play from here Just let me have the last word.

01:01:26 Play from here Don't just...

01:01:27 Play from here And with that, with Eli having this endless

01:01:30 Play from here tautological conversation...

01:01:31 Play from here Which will go on for the rest of our workday.

01:01:34 Play from here For the rest of your life.

01:01:36 Play from here AI still does spooky stuff.

01:01:38 Play from here Everybody play.

01:01:39 Play from here Everybody have fun.

01:01:39 Play from here But we haven't got this all figured out yet, obviously.

01:01:43 Play from here You know, there's a prompting approach

01:01:45 Play from here that I'm teaching lately called CRIT, C-R-I-T,

01:01:48 Play from here which is a little bit different.

01:01:49 Play from here And it's about questions back and forth.

01:01:52 Play from here And one of the cautions that you have to bring up

01:01:54 Play from here is that if you keep asking it to ask questions,

01:01:57 Play from here it will keep asking questions for the rest of time.

01:02:01 Play from here So you have to cut it off and say, you know,

01:02:04 Play from here ask three questions or whatever it is,

01:02:07 Play from here and then move on.

01:02:08 Play from here Because there's no natural limiter.

01:02:10 Play from here And that still won't move on.

01:02:14 Play from here As demonstrated.

01:02:15 Play from here One always do what you want.

01:02:16 Play from here Anyways, good to see you Special K.

01:02:19 Play from here Glad you survived Los Angeles.

01:02:20 Play from here And yeah, until next time, buddy.

01:02:22 Play from here That's the pod.

01:02:23 Play from here All right.

01:02:54 Play from here So,

01:02:55 Play from here if you're interested in becoming a participant

01:02:57 Play from here of Custom AI Solutions,

01:02:59 Play from here visit my agency at spark6.com.

01:03:02 Play from here You can also sign up for my free newsletter,

01:03:05 Play from here AI Framework Friday by visiting spark6.com forward slash newsletter.

How to cite this episode

Credit the named speaker and AI at Work by AscendAI. Link to the episode or the passage you quoted.

“Ep #10 The “Knowledge Graph”: The Missing Link Between Your CRM and AI.” AI at Work by AscendAI, 2026-01-15. https://ascendlabs.ai/podcast/the-knowledge-graph-the-missing-link-between-your-crm-and-ai/

Recent episodes

View all episodes
×
AI Solutions Graphic
Free Resource
Get Started Now