From BI to AI: What My Years in BI Have Taught Me About What Comes Next

I still remember the excitement around some of the early BI projects I worked on.

We would spend weeks, sometimes months, bringing data together, cleaning it up, defining the metrics, building the reports, and finally putting a dashboard in front of a business leader.

And then came the moment we were all waiting for.

The business leader would look at the dashboard and say: “This is great. But why is this happening?”

That question has stayed with me. Because in many ways, it captures the journey we’ve been on in BI for years. We got very good at showing people what happened. Then we got better at helping them understand why. Now AI is asking us to think about something much bigger: what if the system could help us figure out what happens next, and actually do something about it?

That, to me, is the real story of the move from BI to AI.

I’ve spent a significant part of my career in the BI and data world, and I’ve seen the progression firsthand.

We started with reports. Then dashboards became the thing everyone wanted. Then came self-service BI, where everyone wanted to slice, dice, and drill down into their data without depending on IT. Then came advanced analytics, cloud data platforms, data lakes, and increasingly sophisticated ways of bringing information together.

Every generation promised to bring the business closer to the data, and every generation did. But there was one thing that never really changed: the human was still responsible for connecting the dots.

A dashboard could tell a sales leader that revenue was down. The leader then had to figure out why. Perhaps a large customer had reduced orders. Perhaps pricing had changed. Perhaps a competitor had entered the market. Perhaps the problem wasn’t sales at all, and supply constraints were preventing orders from being fulfilled.

The dashboard gave us the starting point. The thinking happened somewhere else, usually in a person’s head, sometimes in another spreadsheet, and occasionally in a meeting that lasted much longer than anyone intended.

That was BI, and honestly, it did a remarkable job. But the world around it has changed.

This is where AI has caught my attention. Not because it gives us another way to build dashboards, and not because we can now ask a question in natural language instead of clicking filters. Those things are useful, but they’re not the real change.

The real change is that we’re beginning to move the machine closer to the decision itself.

Think of it as a shift in verbs. BI mostly surfaces information. AI can actually reason over that information: connect dots, weigh options, explain trade-offs. And agentic AI takes it a step further. It can act on that reasoning, not just describe what someone should do.

Map that out and it reads almost like a maturity curve: descriptive, diagnostic, predictive, prescriptive, and now agentic. Each stage pulls the enterprise a little further from “here’s what happened” and a little closer to “here’s what we should do, and it’s already being done.”

Imagine a business leader asking, “Why are we losing customers in this segment?” Instead of simply returning another chart, an intelligent system could look across customer history, transactions, product usage, pricing, service interactions, and other relevant information. It could connect the dots. Then the next question might be, “Which customers are most likely to leave?” And then, “What can we do about them?”

That’s where things start getting really interesting. We’ve moved from reporting to reasoning, and now, with agentic AI, we’re beginning to move from reasoning to action.

This is where my BI background probably makes me a little more cautious than some of the current AI conversations.

AI is only as good as the world it can see. I’ve spent enough time around enterprise data to know that the answer to a business question is rarely sitting neatly in one database waiting to be discovered. The customer may be in CRM. The order may be in ERP. The service history may be somewhere else. Pricing may live in another system. Important business rules may exist in a document. And some of the most valuable context may still be sitting with an experienced employee who has been doing the job for fifteen years.

So when someone says, “Let’s put AI on top of our data,” my immediate question is: which data? And does the AI actually understand what that data means?

This is why I don’t believe AI makes the work we did in BI irrelevant. It makes that work more important. The data foundation matters. The definitions matter. The integration matters. The governance matters. And context matters, perhaps more than anything else.

At Sage IT, this is the part of the conversation I find particularly exciting. We often talk about Data, Integration, and AI as separate capabilities. I don’t think they should be.

Suppose an AI agent identifies that an important customer is at risk of leaving. Knowing that is useful, but what happens next? The agent needs access to the customer’s history. It needs to understand the contract, know current pricing, possibly check inventory, and understand open service issues. And if we want it to actually do something, it needs to be connected to the systems where that action takes place.

That’s where integration changes its role. For years, we thought about integration largely as connecting systems to each other, getting Salesforce to talk to an ERP, basically. Today, I think we need to start thinking about integration as connecting intelligence to the business. That’s a meaningful shift, because it means AI can’t create real operational value floating on its own. It has to be wired into the actual systems the business runs on.

So the picture I see emerging is simple: data gives AI something to work with, integration gives AI context and access, AI gives the enterprise intelligence, and agents turn that intelligence into action.

People ask me this quite often, in different forms. “Is AI going to replace BI?”

My answer is no. At least, I don’t think that’s the right way to look at it. The dashboard isn’t the enemy, and I suspect dashboards will be around for a very long time. What will change is our relationship with them.

Once AI is wired into the systems, the dashboard stops being a static wall of charts and starts becoming a conversation, and a decision-making interface, not just a place to look things up. Instead of opening six reports to understand why something happened, I might simply ask, “What’s driving the revenue variance?” Then, “Show me the customers contributing to it.” Then, “What are my options?” And eventually, “Take the actions that are within policy and show me what needs my approval.”

That’s not the end of BI. To me, that’s BI growing up.

A word of caution here: the move from BI to AI shouldn’t start with “where can we bolt on some GenAI?” That’s backwards. It should start with the actual business decisions people are trying to make and the outcomes they’re trying to drive.

My advice would be surprisingly simple. Don’t start with AI. Start with a business problem. Find the decisions that consume too much time. Find the processes where people spend hours gathering information before they can actually decide anything. Find the places where the organization has plenty of data but still doesn’t have enough context. Then ask: what would happen if we could make that decision faster, better, and more consistently?

If you’re mapping out where to begin, five priorities tend to matter most:

Those questions, and that sequence, matter far more than which AI model or which chatbot to deploy.

Perhaps the reason I’m optimistic about this transition is that I’ve seen this movie before. Every major shift in BI initially looked like it would make the previous generation irrelevant. It didn’t. The new capability simply built on what came before, and I think AI will do the same. The data warehouse doesn’t disappear. The dashboard doesn’t disappear. The integration layer doesn’t disappear. The governance doesn’t disappear. They become part of something bigger.

Zoom out far enough and the shape of this journey becomes clear. Systems of Record told us what happened. BI and Analytics helped us understand it. AI is opening the door to systems of intelligence, helping us predict and reason. And agentic AI is starting to build systems of action.

The journey is: Record, Insight, Intelligence, Action. And underneath it sits something very familiar to anyone who has spent years in data: trusted data. That hasn’t changed. What has changed is what we can do with it.

I’ve spent much of my career helping organizations get better at understanding their data. What excites me about this next chapter is the possibility that we can help them do something more. Not just understand the business better, but help the business respond better.

That, for me, is what the journey from BI to AI is really about.