The Human Side of AI Readiness - How to Prepare Your Team and Business Logic
Mar 28, 2024Episode 123
Brick Thompson, Caleb Ochs
In a continuation of their AI-readiness discussion, Brick and Caleb outline the importance of the human element in preparing for generative AI analytics. They share strategies for ensuring that your company's culture is ready to harness the power of AI, from how to start improving data literacy to defining clear and consistent business logic.
Blue Margin increases enterprise value for PE-backed, mid-market companies by serving as their fractional data team. We advise on, build, and manage data platforms. Our strategy, proven with over 300 companies to-date, expands multiples through data transformation, as presented in our book, The Dashboard Effect.
Subscribe here to get more episodes of The Dashboard Effect podcast on your favorite podcast app.
In a continuation of their AI-readiness discussion, Brick and Caleb outline the importance of the human element in preparing for generative AI analytics. They share strategies for ensuring that your company's culture is ready to harness the power of AI, from how to start improving data literacy to defining clear and consistent business logic.
Blue Margin increases enterprise value for PE-backed, mid-market companies by serving as their fractional data team. We advise on, build, and manage data platforms. Our strategy, proven with over 300 companies to-date, expands multiples through data transformation, as presented in our book, The Dashboard Effect.
Subscribe here to get more episodes of The Dashboard Effect podcast on your favorite podcast app.
Welcome back to The Dashboard Effect Podcast. I'm Brick Thompson.
Caleb Ochs:
I'm Caleb Ochs.
Brick Thompson:
Caleb, we're going to continue our discussion from the last episode around getting ready for AI - generative AI analytics. And last time, we talked about consolidation of data and creating a semantic model or semantic layer for providing access. First of all, pulling data together and then providing access to that data that humans and various tools, including AI can understand. But there's more to it than that. Right?
Caleb Ochs:
Yeah, we kind of talked about the data side, the hardware side, if you will. Now, we're going to talk more about the softer side of things and more of the people element because people are always going to be a part of this. And they're going to be the ones using it. So we got to get them ready for it, too.
Brick Thompson:
That's right. And just from an organizational standpoint, if you've got good, consolidated clean data with a nice semantic model, but your organizational culture is not data-driven, then it may be a wasted effort. And so how do you get to being data driven? I mean, it's not that simple. Right? We see some companies that are there. But there's various levels of maturity. And very often people are not at the point where they can actually say they're data driven. And I think we saw a CIO survey recently, that a pretty low percentage that thought their organization was data driven. So to start with, I think what you have to do is actually assess where you are in terms of being data driven. And there's various tools you can use for that. But most people running a company or senior executives in a company will know, are their decisions driven by data? Are the people coming to them with recommendations backing up their recommendations with data? Or is it mostly seat of the pants, anecdotal, one-off experience that led to this decision. And once you have that idea, if you need to improve - which most companies will - you need to start engaging with your people to help them get there.
Caleb Ochs:
Yeah, yeah, enable them with the tools, the knowledge, all that stuff, to be able to feel comfortable bringing data when they're making a decision or bringing a recommendation.
Brick Thompson:
Right. So yeah, so as you said, tools and knowledge, so some education around, how do you talk about data? So just some data literacy. There's a famous book, I remember from an econ class back in college, "How to Lie With...," I think it was "How to Lie with Data" or "How To Lie With Graphs" ...anyway, you can make mistakes, even if you have good data, if you're not approaching it right. And then giving them the tools - like a Power BI - or teaching them how to use PowerPivot in Excel, or Tableau, or whatever your tools are. And making sure that they're comfortable using those to produce information that's going to help them make better decisions and make better recommendations. Right. So there's that whole equipping piece. And then as a senior leadership team, you have to start modeling and demanding that the company be data driven. So when someone comes to a manager comes to a meeting and says, "I have this recommendation, and it's based on the fact that I talked to this one customer, and they said that such and such..." You need to be able to start engaging and say, "Okay, so help me see the data that supports that, not just the anecdote. What's the data behind that?" and really driving and fostering that data driven mentality. Easier to say then do.
Caleb Ochs:
Yeah, for sure. I mean, it's easy to just rely on intuition. I also think that, you know, just simple things, like, if you're not bringing data, or a dashboard, or report to a meeting already. If it's not part of your regularly scheduled meeting agendas - make it one. And then everything you said. You're just starting somewhere. But yeah, push some reports up on TVs around the office or on your company intranet. Stuff like that. Get people used to seeing it and used to looking at it. Used to really relying on it, and then, you know, it'll start becoming natural, and people will actually feel weird not having it if they're making a decision.
Brick Thompson:
Right. That's good. Yeah, I like all those. And then along with that softer component, you still have to take that consolidated and model data and start figuring out alright, what do we care about in here? And figuring out what are the business goals? What are the most important business goals? And what are the KPIs and sets of data that support those? And then start developing that and that's harder than it seems, too.
Caleb Ochs:
Yeah, you got you have to define what those mean.
Brick Thompson:
And we've seen that over and over. We'll end up You can go sit down and prioritize (that's a challenge in its own). Once you have the things prioritized and you're with a meeting with the finance department and the sales talking about how do these things come together? Then department, and they think of revenue completely differently. you're into business logic definition. And that's always And so, you know, finance will say, "Okay, we got this great kind of a abrupt experience for some people because they've just kind of gone through the last few years thinking that everybody was thinking about things the same. And then you start uncovering like, "Oh yeah, no. People do not think about something as simple as revenue the same way across our company. So we have to align on that exactly what that means." So that's a big step, and it's super important one. Yeah. report we just built, and here's the numbers." And sales will say it's wrong, or vice versa.
Caleb Ochs:
Yeah. Right. Exactly.
Brick Thompson:
And so you have to sort that out, because you're gonna want your systems to be speaking a single version of the truth, or people won't trust the data. And so you can have two different definitions of revenue. That's totally fine if that supports the business and getting towards the strategic objectives. But you need to call them something different.
Caleb Ochs:
Right? Right. Yeah. Call them something different. That'll do it. Then people will understand Sales Revenue is different than Financial Revenue or whatever, you know. So yeah, it's really important.
Brick Thompson:
Yeah, so once you've defined those, you have to start building the measures and KPIs around that. And then pressure testing. So you're gonna put them in the hands of trusted users and get feedback and find where the holes are. Find where there's edge cases that actually deliver something that you didn't expect, and get those corrected. And that's sort of an ongoing process. You know, as the business changes, those rules will change. If you change your ERP system, I guarantee you're gonna go through new gyrations on this. If you're integrating a new system, let's say you buy another company, you're integrating their stuff, you need to understand how do they define it, compared to how do we define it? How do we make those matchups, so that we're not adding apples and oranges and thinking we're adding apples and apples? There's a lot of stuff there. So even though we'll hopefully - soonish - have generative AI that you can ask a natural language query to get an answer. You've got to have all this stuff sorted out.
Caleb Ochs:
Yeah, I kind of think of this, defining the metrics and the logic behind them is almost like your start to internal, generative AI prompt engineer guide. Yeah. So if you need to see this, you would consult that list, and then you would know what to ask to get what you're expecting. Right. You need to know what you're after, so the thing can go get it.
Brick Thompson:
I think that's right. Yeah, I'm sure that as we're implementing these generative AI systems, there'll be ways to generalize and have synonyms and so on, but there's going to be, there's going to have to be training around that, too. And so just defining those KPIs and measures to start with is going to be critical.
Caleb Ochs:
Right, right. Exactly.
Brick Thompson:
And that's all stuff you can do now. That's valuable right now.
Caleb Ochs:
Yeah. Valuable right now, and it sets you up for the future. Well said.
Brick Thompson:
There's no reason to wait until GPT-5 is actually doing some of the stuff we want or Microsoft Copilot, or whatever system comes out that we say, "Ooh, this is actually pretty good generative analytics." We haven't quite seen that yet. It's getting there. A little bit of bleeding edge still, but, but it's going to get there. And if you've got a lot of this, you've got the plumbing that we talked about in the last episode, you got the softer things now, then you're ready to take advantage of it. And you're getting value all along the way. Doing an infomercial here. Do it now. It will pay off. All right. I think that's all I had. Anything else from you?