YOU are the AI Cavalry

Speaker 1 (00:04): Welcome to Raw Data with Rob Collie, real talk about AI and data for business impact. And now CEO and founder of P3 Adaptive, your host, Rob Collie.

Rob Collie (00:20): Hello, friends. Did you hear that slightly new intro? It's different in a few ways. First, the name of our podcast, Raw Data by P3 Adaptive never really rolled off the tongue, did it, but Raw Data with Rob Collie has that nice smooth alliterative flow. And then there's the velvety luxuriant voice of our new announcer telling you what we're about, real talk about AI and data for business impact. Another difference is that depending upon which podcast client you use and how often it downloads updates, you will soon, if not already, also see a change to our logo. I won't spoil it here if you haven't seen it, but yes, the beloved mullet people are gone replaced by a logo that is designed to better grab your attention, but all of those changes are simply designed to tell new listeners what we're about. If you've already been a regular listener, you're going to find that everything important remains the same.

(01:16): Yes, it is now called Raw Data with Rob Collie, but Justin Manhart remains firmly planted in the co-host chair with the exception of these occasional Rob solo episodes like this one, I simply could not imagine doing this show without him. And yes, even though we've been a little lazy about using it lately, we're very much sticking with the twangy guitar music we've had since episode one nearly five years ago. So the changes are about attracting new listeners to the show at an increased rate. We think we're filling a valuable gap in the information space these days and we want to get more credit for that and we want to reach more people. There are plenty of shows about AI and data that are essentially high-level hype, and then there are plenty which are down-in-the-weeds technical. In between though there's a need for communicators like us who are technical enough to understand the tech, but then able to translate it back into practical business context.

(02:08): We see that as our mission and we like to think we're pretty good at it. If you agree, we'd really appreciate it if you took the time to give us a rating on your podcast app because even a single new positive rating does get noticed by the algorithm. This show is a labor of love. We don't pay wallet and we don't sell ads, so if you want to say thank you by way of a rating or a review, we'd really appreciate it. Going back to the new intro for a moment, you might've noticed that we say the show is about AI and data in that order, not data and AI, AI and data. Now, this show has been mostly about AI for quite a while now already, so again, it's really just the same, but that we're going so far as to formalize it in that order.

(02:51): AI and data tells you something important. If your career has been about data, it's going to have to be about AI. You might as well start trying on the idea, even if you're currently kind of on the sidelines of AI, keeping tabs on it, but kind of only using it for your own personal, even professional, but personal usage, you should start thinking, "I'm going to need to be an AI professional." And that at first sounds a lot like the anxiety inducing clickbait you see on LinkedIn, right? Scary, you must be doing X. If you aren't doing X, your career is at risk. But that kind of fearmongering is just the easiest, laziest way to draw attention, and it misses the better message, which happens to be the truth. If your career has in any way been about data or adjacent to data, it absolutely can be about AI.

(03:45): So it's not just it has to be about AI, it's also completely achievable for you, and it's not going to be as scary as it might seem. That's been my own journey over the past year or so. I've been pretty deep down the AI rabbit hole and I've gone from nervous to excited over the course of that journey, and I'm very happy to report back that AI is just, you can hear the air quotes there, "AI is just the next phase of our data careers." For example, if you're a business leader and you're listening to this podcast, well, you're someone who thinks about data, otherwise you wouldn't be listening. And whether you're early in your company's data journey or have reached a point of relatively modernized maturity with data, if your company looks to you for direction about data-driven business improvement, they're going to be looking to you for AI leadership as well.

(04:33): And you're so much better positioned to meet that challenge than you might think because of your background as a data thinker. Or if you're a hands-on data practitioner like a power BI developer or a data engineer, the way your brain is wired to think about data is going to remain a massive, massive asset for you. Now, your job is, as a practitioner, your job is going to have to change more than the data-driven business leader's jobs will need to change, but in both cases, whether you're a business leader or a data practitioner, your foundation in the data world is your bridge to AI. For leaders and practitioners alike, the knowledge you've amassed, the wisdom you've gained, and frankly, just the fact that you're the rare human being who likes thinking about these sorts of things, all of those are home field advantage for the game of AI, and here's kind of the central reason I've come to that conclusion.

(05:24): It's kind of a funny thing to say, so I'm going to say it twice. Here it is. If you're an AI professional in the world of business, the one thing you cannot change is the AI. Yeah. If you're an AI professional in the world of business, the one thing you cannot change is the AI. Now, AI professionals being unable to change the AI might sound like a contradiction, but it's 100% the truth. There is a very small population of human beings working in AI who can change the AI, but they're all working at places like OpenAI or Anthropic. The researchers who are building things like GPT-5 or Claude Sonnet at those companies, yeah, they can change the AI, but the rest of us cannot afford to, even if we knew how, it costs hundreds of millions of dollars to develop and train something like GPT-5 and when they're done with it and they give it to us, it is absolutely the same thing for all of us.

(06:22): To really drive that home when Microsoft's CEO, Satya Nadella, sits down with GPT-5 or Claude Sonnet 4.5 for a chat session, he's sitting down with exactly the same thing you are. Yeah, he logs in as him and you log in as you, but eventually behind the scenes, the exact same backend LLM voodoo is running on your behalf as running on his behalf, maybe even on the same farm of GPUs in some hydroelectric power data center. GPT-5 or whatever model it is you're using is identical for him as for you and for everyone else, even Satya Nadella can't change it. Excel's 40th anniversary party was held here in Seattle at Microsoft campus this past week, and even though I don't work at Microsoft anymore, everyone who has ever worked on Excel was invited. So of course I went and I saw a bunch of old friends and former colleagues, and guess what?

(07:18): The ones who were still there working on Excel, most of them are working on AI features in Excel. Shocker, I know. I talked to an old friend and engineer, for instance, who first came up with the new Copilot function in Excel. I also talked to multiple product managers who've been working hard on the new agent mode in Excel. Now, these are all professional software engineers with access to the deepest resources imaginable. They work at Microsoft and they're building software for hundreds of millions or maybe even billions of people. And do you know what they have in common with the rest of us? That's right. They can't change the AI either, behind the scenes, agent mode, Copilot function, all those sorts of things. They're still just using the same LLM models that we have access to, and they're no better or no different for them than they are for us.

(08:06): I keep saying this like it's good news and maybe you're confused as to why it's good news. I mean, can't is usually a bad thing, but in this case, can't change the AI also means you don't have to change the AI in order to be successful with it. You're not expected to, and the bar for success with AI just becomes so much more within reach. Once you internalize that, you're not expected to be an AI researcher, none of us are. Okay, but if you can't change the AI, you have to be able to do something with it, right? When I talk about success with AI here, I mean the game changing feels like magic success, and you don't get that when you're just using off the shelf chatgpt.com.

(08:50): So even though you have zero power to change the core AI itself, you do have tremendous power to shape the kind of results you get from it, results that go far beyond the off-the-shelf single-user chat experience. And you do that by feeding data to the AI. Every customized AI solution, every game-changing productivity boost, every time your business is going to gain what feels like a superpower, it's going to be because you fed it the right data, whether it's numbers or text or both, you fed the right data into the right AI at the right time, or alternatively because you set things up in a way that the AI can decide for itself what it needs in the moment and then go fetch it.

(09:34): In the same way that you might stand at a whiteboard and draw boxes representing data sources that you want to bring together in a Power BI model today, tomorrow, you're going to be drawing similar sketches, boxes representing sources of data or information, arrows flowing between those boxes into another box representing the AI, and then arrows flowing back out of the AI representing answers or responses which are themselves, those answers or responses from the AI, which are themselves a form of data.

(10:05): Now, where do those arrows lead when they leave the AI box? Sometimes they lead back to one of the original sources of information. It gets stored there in some new way. Sometimes those arrows will flow back to a chat interface, so the user who's using a chat interface sees the answer. Sometimes they'll flow to a system resulting in an email or like a Teams message or something like that. Now, that kind of flow is very much right up our alley as data people. In fact, the way I keep using the words data and information interchangeably is itself very exciting because even free-form text is now something we can treat as data thanks to AI, whether as inputs to or outputs from AI, free-form text can now be analyzed, calculated over, reasoned over, and then stored, the results of it anyway, stored in ways that are going to feel very, very familiar to people like us, people who spent time thinking about doing these sorts of things with structured data like databases, spreadsheets, and tables.

(11:05): Well, it's not going to be exactly the same, but it pretty quickly starts to feel 90% similar. So my dear data brethren, whether business leaders or data practitioners or hybrids, this means you can be the AI cavalry, and that's a good thing because guess what, there isn't another cavalry coming. I've long talked about how the data gene is essentially just a personality trait that says, "Oh, cool," the first time you see a nifty spreadsheet rather than, "Oh, gross," like what most of humanity says, and for those other people who do regard spreadsheets as gross and scary and boring, even if they do get excited about AI, like for selfish career reasons, when it's time for them to do anything practical with it, their same allergic reactions are going to kick in and stop them from getting anywhere. On that last point, by the way, we're already seeing that snake oil dynamic in full effect.

(12:00): AI is an absolute once in a lifetime gift to the opportunistic fake it until you make it types. And I'm not talking about the good kind of fake it until you make it the kind where you do actually make it. I'm more talking about the fake it till you take it crowd as in take the money and who cares if you make it funny story from the trenches last week, we see ourselves, P3 Adaptive, as an AI and data firm for the mid-market. We love the mid-market for its agility relative to enterprise orgs, but also because of the authenticity and accountability that necessarily develops on smaller teams. And frankly, it's also just so much more satisfying, giving superpowers to people who thought they were priced out of the game. We love the whole Robinhood vibe. It gets us out of bed in the morning, but we do still have a handful of enterprise clients and over the past year we've watched one of those enterprise clients bring in a series of high-priced AI consultants.

(12:55): They've spent obscene amounts of money on those AI consultants because frankly they can afford to. And we, P3, haven't been the least bit bothered by that. And we have been developing our own AI capabilities, again with a lens towards the mid-market, and it would've seemed to us anyway arrogant for us to go to our client or enterprise client and say, "Hey, why didn't you hire us for that work instead?" But we recently started to get wind that none of those high-priced AI consultants have managed to deliver anything. And as we heard more and more, we started going, "Oh wait, it sounds like we might already actually be doing better than that." Which even to me at the time didn't make sense. I get the impression that the spending on these projects has already run into seven figures at this company. So even I, skeptic of big consulting firms that I am, I just kind of believe by default that those people must know what they're doing if they're paid that much.

(13:45): But even while we were giving those other firms the benefit of the doubt that they haven't earned, we know that where there's smoke, there's often fire. So we finally raised our hand and said, "Hey, maybe we should meet and walk you through our own capabilities with AI." So we had a meeting, showed what we could do, and our sponsor at the company, a VP at this enterprise org was almost like what took you so long to speak up? Now it's still early in the maybe P3 helps them with their AI project story, but he did seem to confirm that we are indeed well ahead of the self-proclaimed hotshot firms they've been using so far. He said to us, and this is almost a direct quote, "I think it's you, P3, who is going to be able to solve this for us because you're so dialed in to the underlying data and systems that it relies upon."

(14:30): So just like I'm here today telling you that you're the AI cavalry, he was telling us his go-to data people that we are the AI data cavalry. Yeah, we already believe that we were, but it turns out even we weren't believing it as much as we should be. So there's a good chance we're going to be in the AI game for that enterprise client here pretty soon because we're data people who've grown into AI people, and if we can play on that stage, we can play on any stage. So come on in, fellow data geners, the water's fine here in AI land.