Everyone's Using AI. Nobody's Using It Where It Counts.

It's 2026. Your team is using AI every day. Marketing drafts content with it. Sales preps for calls with it. Finance summarises board reports with it.

That saves time. Real time. If your people are getting work out the door faster and spending less of their day on busywork, that's a win. Take it.

But don't confuse activity with impact.

More than 75% of employees say they use AI at work. And nearly all of them are using it for the same thing: making communication faster. Drafting. Summarising. Research. The stuff that lives in documents, emails, and slide decks.

Almost nobody is pointing it at the decisions that actually determine whether you hit your numbers.

You solved the easy problem first

Copilot seats. Approved tools lists. Governance frameworks. AI was rolled out to every department. Your board sees "AI adoption is up across the organisation," and everyone nods along.

52% of employees are drafting and summarising content. 57% are doing research. Teams are moving faster.

And that's just the official number. Most organisations know there's a whole layer of unauthorised AI use happening beneath the surface. People are using ChatGPT, Gemini, or whatever else they found on their own because the approved tools didn't move fast enough. But here's the thing: the unauthorised stuff looks exactly like the authorised stuff. More chatbots. More research tools. Maybe a bit of automation. Nobody's sneaking around building demand forecasting models in secret. The shadow AI is just as unambitious as the approved AI.

So whether it's sanctioned or not, the picture is the same.

But what's the financial impact? Not time saved. Not adoption rates. The actual impact on your margins, your procurement costs, and your inventory carrying costs. Can you point to a line item on your P&L and say, "AI did that"?

Most leaders can't. Not because the productivity stuff is useless. It's because companies treated it as the destination instead of the starting point. They got the quick win and stopped.

The decisions that actually cost you money

Think about where your business really makes or loses margin. It's not in how fast someone writes an email. It's in purchasing. Demand forecasting. Inventory management. Pricing.

A $400 million company might spend $200 million a year on purchasing. If 2% of that is waste (and there's almost always at least that much hiding in the data), that's $4 million. Every year. Just sitting there.

So where's the AI? Helping marketing write LinkedIn posts.

That's not the marketing team's fault. They found something that made them better at their job and they used it. Good for them.

The real question is why nobody aimed AI at purchasing decisions. Or forecasting. Or inventory.

Part of it is that the productivity stuff was easy. Easy to deploy. Easy to feel good about. Nobody had to be wrong. You hand someone Copilot and their work improves. Everybody's happy.

But there's another reason people don't talk about enough: until very recently, you couldn't do it. The tools to go into your ERP data and find real operational waste didn't exist. Your ERP vendor wasn't building it. Nobody else was either. So what were you supposed to do?

That's a fair excuse. It was a fair excuse. But it's expiring fast.

The tools exist now. The ability to go into your actual data, find the waste, and put a dollar figure on it is real. Which means the only thing standing between you and that money is the willingness to look.

And that part is still hard. Going into your data and finding waste means admitting that decisions have been wrong. It means telling your VP of supply chain that the spreadsheet he's been refining for five years might be leaving millions on the table. People don't want to hear that.

But "it's uncomfortable" is a very expensive reason to do nothing.

Your ERP vendor isn't going to fix this

Most companies run SAP, Oracle, or something similar. And those vendors have been adding AI features for years. New releases. New roadmaps. Buzzwords in every quarterly update.

But pay attention to what that AI actually does. It makes the ERP run smoother. Faster queries. Cleaner data processing. Nicer dashboards.

It doesn't tell you you're overpaying a supplier by 12%. It doesn't flag dead inventory eating your margin in a warehouse nobody's checked in six months. It doesn't dig into your purchasing data and say: Here's where you're losing money, fix this first.

ERP AI is built to improve the platform. Not your business. If you've been waiting for your ERP vendor to solve this, you've been waiting for someone who has no reason to tell you their system falls short.

"Roughly" is not a strategy

Ask any leader where their waste is, and you'll get the same answer: "I know. Roughly."

That word should scare every CFO reading this.

"Roughly" means million-dollar decisions based on gut feel. You know you're probably carrying too much inventory in some categories. You know purchasing patterns haven't been seriously looked at in years. But you can't put a number on it. And without a number, you can't prioritise it.

So it sits on the roadmap. Behind the next ERP upgrade. Behind the next Copilot rollout. Behind whatever's easy.

Every quarter, you're in front of your board talking about margin pressure. Supply chain costs. Volatility. And you've never once been able to say: "We went into our own data, found the waste, and fixed it." Not once.

Start with the money

The AI your teams use today has value. Keep it. But stop calling it a strategy.

A real AI strategy starts with your P&L. It starts with someone going into your actual data and showing you exactly where you're losing money. Not a demo with fake numbers. Not an industry benchmark. Your data. Your dollars. Specific and clear.

If you measured your total AI investment against its actual impact on your bottom line, what would you find?

If you don't know, you have your answer.

Your teams nailed the productivity side. Now point AI where it counts. Because the margins aren't going to find themselves.

That's what Divi does. It plugs into your ERP, rips through your data, and tells you exactly where you're losing money and what to fix. No dashboards that tell you nothing. No buzzwords. Just the waste, the dollars, and what to do about it.