Subscribe
Share
Share
Embed
AI First with Adam and Andy: Inspiring Business Leaders to Make AI First Moves is a dynamic podcast focused on the unprecedented potential of AI and how business leaders can harness it to transform their companies. Each episode dives into real-world examples of AI deployments, the "holy shit" moments where AI changes everything, and the steps leaders need to take to stay ahead. It’s bold, actionable, and emphasizes the exponential acceleration of AI, inspiring CEOs to make AI-first moves before they fall behind.
Andy Sack (00:05.55)
This is AI First with Adam and Andy, the show that takes you straight to the front lines of AI innovation and business. I'm Andy Sack and alongside my co-host, Adam Brotman. Each episode, we bring you candid conversations with business leaders transforming their businesses with AI. No fluff, just real talk, actionable use cases and insights for you.
Welcome, Adam. We're doing this 10-minute episode on Monday, not Sunday. But good to record another episode with you.
Adam Brotman (00:40.249)
Yeah, yeah, these are fun, these little bite-size moments of AI first insights, I guess.
Andy Sack (00:48.95)
Yeah, so I think today we want to talk about something that we've been seeing in the market. And when I say that, it's come up in conversations with different either AI councils or executives at businesses who are starting their AI first transformation. And a lot of times we hear, when talking about that, we'll hear,
that they want to get their data in order or set up a data lake. they expect to, once that's done, to be moving forward with more significant AI-first transformations. What's your observation of that? And how do you respond to that, Adam?
Adam Brotman (01:34.619)
So it's an interesting topic because on the one hand, it's right and on the other hand, it's not right. And so how can both be true at the same time? So you've got, first of all, the definition of data is changing when we talk about AI. It used to be data was all about structured data, for the most part.
And it was like, are we? We need to put our data in a data lake. And it's all of like a bunch of, you know, structured data that you could do analysis on. And if that's, if that structure data is not centralized and accessible and cleaned up and correct, then how can you possibly do analysis and get insights from that data from business intelligence teams and data science teams? And so.
there's a natural sort of muscle memory from lot of organizations, particularly CTOs to say, we're not ready for the AI because the eye and AI, the information, the intelligence isn't cleaned up and ready to go. And yet the thing that's happened with these generative AI systems is that they are very good at looking at both structured and unstructured data. Unstructured data being like transcripts of calls and documents and
presentations and the like. I think that the fact that AI can get after both of those things, there's a, we see a lot of times these CTOs wanting to be like, hold up, we got to get our data cleaned up and then we got to connect the data to the AI and that's going to create security and other issues. And all of a sudden you're now like airplanes waiting on the ATAR Mac to get going on your AI.
journey. And what I typically say in those situations is I get the need and the desire to clean up and centralize your data. Because you're right, like there and we can talk in a second about all the things you can do better if that's the case with AI with generative AI. But if you're going to wait six months or a year to get going on your AI journey until you can get your your data organized and cleaned up, like you're missing the boat because
Adam Brotman (03:54.096)
You can do so much just by grabbing the last six months or last 90 days of data, just doing a quick query, grabbing it, and then uploading it and using it as part of your source engineered smart prompt where you're solving some problem, tackling some issue as a business. And you can be solving that problem, getting those insights today, not waiting six months. And six months is a lifetime right now. So that's...
but at the same time, I get why there's muscle memory to do that. I get why, because there is a lot of value you can get if you did have all of your data, like organized and connected right into your AI system, so to speak that you were using. But I think it's a little bit of a throwback or a holdover from these, these AI initiatives being sort of given to tech folks and the tech folks is do a tech implementation.
playbook on it, whereas there's a new playbook for this stuff. I think the part of the new playbook is experimenting and not waiting for perfect to be the enemy of good enough.
Andy Sack (05:01.89)
Yeah, just to ground what you're saying, think, in two concrete customer examples. One was,
we've had a customer call and the AI implementation had been handed off to the IT manager. And the IT manager was basically saying that they wanted to get all their data into a data lake first and that the real...
opportunity was to gentify the data in relationship to the customer and that they would expect to do that in mid to end of 26, maybe 27. And I think you and I looked at each other on across the zoom call. And we've heard that story before. And, and, and our advice was, I think it's for this reason that we go this is that AI is not a standard software deployment
at enterprise. should probably not be handed off to the IT department.
And instead, you should follow the playbook that is in our book in AI First, in which it involves actually getting all employees trained and up to speed, giving power to the executives and forming an AI council to deploy it. And I counter the example that I just used with another customer example, which this was a
Andy Sack (06:38.318)
client of ours that runs a chain of cafes and they had a business problem of declining sales and cold beverages. And I think you said to the executive, hey, just send me the last year's data in a spreadsheet, which they did.
without asking for it from BI, they just sent it to you in an Excel spreadsheet. You then took it and put it through ChatGPT and asked it to analyze the data, managed to get some real insights, then transferred that into a...
into an eight slide deck that you sent to me as well as to the executives and everyone was blown away at the speed and the efficacy of those insights. And that's the example, those are the two examples. One which is wait till mid 26, 27 and I'll have a complete picture, really thorough picture of insight, AI insight versus, hey, send the stuff to me in an Excel spreadsheet, let me run through it today and I'll have it to you back in two hours.
Adam Brotman (07:48.26)
Yeah.
Andy Sack (07:49.504)
speed of iteration that's the real unlock in the iteration for Gen.ai and I think you and I are saying to all executives that are listening to this podcast, do not wait, start using the systems now.
Adam Brotman (08:03.502)
Yeah, it's funny because on this topic, I'll be honest, I'm a little bit, sometimes sheepish because I really respect these technology executives and their knowledge for how to build software and how to run technology departments. I mean, I couldn't run a technology department probably, but, but, but,
I, when you started giving those examples, Andy, for just a second, when you gave the example about the one that said, let's wait until we get some data organized. I actually, as you were saying it, I, I could think of four clients and we don't need to go through all four of them that you and I have talked to in the last two months where I couldn't tell which story you're about to give because, because I was thinking about another client we had where we were talking to them and they, they actually said, we really want to get out and become an AI first company.
Andy Sack (08:47.918)
Totally.
Adam Brotman (08:57.308)
And when you get on the phone with our team that's working on it, and this is a different example, and they were like, we're working with this vendor and they're trying to like get all this unstructured data, I think, organized in some ways. And it had to do with like, it was like an investment company. the, and I, and yet at the beginning of the call, the principle that we were working with was saying, here's what I wonder if AI can help me with. And there were all things that you and I were just nodding, like, yes, it can, it can do that, it can help with that.
And then we got on the call with their AI people and they're like, totally with the good intention were saying, we can't do this until we can't really do this until these stuff gets the AI system gets integrated. And I think the problem, I'm just going to say it, just winging it here a little bit, but like a little off script. But I do think the problem is that these AI implementations, if they're given to the tech department to go run, they
They want to, it's interesting. They want to go all the way to bright by doing an integration. Like they think about integrating their data. And I get that because if you're building a technology system that needs to just sort of run on its own and be scalable and have like the least amount of support, like from humans that are not technology or technology technologically oriented, you're going to think it all has to be wired up and plumbed up in a certain way. But that's not how this technology works. Like,
this tech, like the example you gave of the cold beverage decline analysis, like there was another example of a restaurant chain that was trying to optimize their discounts and their ad spend. And they were like, my data is all over the place. I'm like, well, don't you have like 90 days worth of like just X, Y, and Z data laying around? Yes, we do. Fine. Zip it up into a spreadsheet, send it over to me, and let's run it through a thinking model. In this case, Jot2BT5 Pro. And like,
It was incredible how well it did. there is this, I think this misunderstanding between the need to have the data integrated and cleaned up and organized, which is normally true with structured data and systems that you're trying to like not have to worry about everyday users figuring out how to do it. In this case, this is tech.
Adam Brotman (11:18.138)
where the data can actually just be uploaded by the CEO or the CFO or the CMO. And that's not normal. And I get why it throws companies off. But once you start realizing how simple this is to actually upload the data and use some of these models, I think these companies will stop trying to organize and clean up all their data as the precursor to getting anything done. I still think they should do it and integrate because we'll talk in another episode about where these
systems are going when they can start to become autonomous and agentic and just start doing work for you and you're going to want them to have access to clean centralized data in an integrated way. But that's missing all this amazing stuff you could just get off the snide and start doing that.
Andy Sack (12:03.18)
Yeah, so in summary, don't put your AI.
first transformation in charge. Don't put the CTO in charge of the AI first transformation. Build a data lake, but don't wait for the data lake. Get started today using Gen.AI and use Excel spreadsheets to provide the context and get your executives and teams moving. I think that's our advice for the day. Thank you. Thank you all for listening to AI First with Adam and Andy. For more resources on how to become
AI First, can visit our website, forum3.com, download case studies, research briefings, and join our community, our AI First community hub, a curated hub network for leaders turning AI hype into action. We truly believe you can't over invest in your AI learning. Onward.