Why Your ERP Strategy Is Your AI Strategy

What One Company’s Years of ERP Discipline Made Possible Almost Overnight

When AI deployments go well, there’s usually something in the background that made it possible. Often, that something is a modern ERP platform, a common data structure across business units, and years of quiet infrastructure work that nobody celebrated at the time. You can pursue AI without that foundation, and plenty of companies do. But the ones that move fastest and scale most easily, almost always built it first.

Here’s a story of exactly that kind of company. A healthcare distributor with multiple brands and a shared services model — meaning IT, finance, HR, and legal all operate at the parent level, supporting the commercial entities underneath — they’ve been a Velosio client for over ten years, and they did the infrastructure work before it was obvious why it would matter. The AI wins they’re seeing now aren’t a surprise to anyone who watched that journey up close.

From Three Minutes to Three Seconds: How ERP Data Enabled AI-Driven Customer Service

A customer service rep gets a call. The customer wants to know if a product is in stock, the status of an open order, and whether they can get a copy of a recent invoice. Before, the rep would navigate through multiple screens in the ERP system, run a report, and put the customer on hold while they figured out where to find what they needed. The whole thing might take three to five minutes per call, across a high-volume team where turnover is common and many people are relatively new to the job.

Now they type the question into a chat interface in Microsoft Teams. The agent responds in seconds.

That’s the visible part of what this organization built, and it’s genuinely impressive. The more interesting question is what had to be true before it was possible, and why the same capability that took months to build for one business unit took a fraction of that time to roll out to the next.

Where It Started: The ERP Data Challenge Behind AI Adoption

The impetus for the agents wasn’t a top-down AI initiative. It started with a practical problem in the customer service operation. The CS teams spend a significant part of their day fielding questions that require pulling information from Dynamics 365 Finance and Supply Chain Management — order status, tracking, inventory availability, and invoice retrieval. The information exists, but getting to it, assembling it, and presenting it to the customer takes time.

There’s also meaningful turnover on these teams, which means a constant cycle of new employees learning their way around a system that isn’t always intuitive. Meanwhile, the sales team was generating its own flow of questions for customer service — partly because sales reps don’t have F&SCM licenses and had no direct path to the data, so they routed their questions through the people who did.

The solution Velosio helped them build uses Microsoft Copilot Studio to construct agents that sit inside Teams. Anyone who needs order or inventory information — CS rep, sales rep, new hire in their second week — types a question in plain language and gets an answer drawn directly from the underlying ERP data: no navigation, no screens, no hold music.

What Made It Work: ERP Consolidation as the Hidden AI Enabler

The agents are only as useful as the data they can access. A question about order status requires clean, accessible, current order data. An invoice retrieval requires that the document actually live where the agent can find it. None of that is a given in organizations running fragmented systems — where an acquisition from three years ago is still on a different ERP, or where certain data lives in a standalone application that’s never been properly integrated.

The organization dedicated several years to ensuring none of that was true. Over time, they completed four ERP conversions onto Dynamics 365 Finance, consolidated everything onto a single CRM platform, and implemented warehouse management technology across multiple distribution centers. They increased the cloud application usage from 17% to 68%. Every acquisition was migrated onto the same platform. When building the agents, the data environment was unified, structured, and consistently maintained across the business. The agent didn’t need to search elsewhere; everything was in the same place.

The Detail That Changes The Math: Why Standardized ERP Accelerates AI Scale

When the first agent launched, it worked — and more than that, it got used. Adoption was strong, the team responded well, and the efficiency gains were real. Average handle time improved. Onboarding now includes the agents as a primary training tool, meaningfully shortening ramp time for people still learning the systems.

Then came the second business unit, and this is where the story gets interesting for anyone managing a multi-entity organization. Building the second agent wasn’t really building at all. It was copying the first one and making minor modifications — same logic, same underlying platform, same data structure. No new integrations to wire up, no translation layer to build between systems that had never talked to each other. What took months the first time took a fraction of that the second time around.

For companies running multiple brands or business units, particularly those that have grown through acquisition, that kind of replication is often the difference between AI as a real operational capability and AI as a perpetual pilot. The fast rollout was a direct result of the consolidation work that came before it.

Velosio partnered with the organization’s team over several months to further extend and enhance the agents, adding capabilities and expanding their scope. The agents now serve 300 to 400 users and handle thousands of interactions per month. What previously took two to five minutes to look up manually now takes seconds — multiplied across a full customer service operation running all day, that adds up fast.

What This Means If You’re Not There Yet: Assessing Your ERP Readiness for AI

Most executives who read AI success stories have a version of the same reaction: interesting, but it wouldn’t work the way our systems are set up. They’re usually identifying something real. The reason these agents work isn’t primarily about the sophistication of the underlying model — it’s about having a single, consistent data source that the model can actually rely on. Organizations that haven’t gotten there yet aren’t wrong to be skeptical. They’re looking at an honest reflection of where they stand.

Worth noting: several years ago, this company would have said the same thing. The foundation that makes the agents possible was built deliberately, over time, through a commitment to platform consolidation that didn’t have an obvious AI payoff when it started. They brought every acquisition onto the same system. They saw the migrations through. That discipline is what created the conditions for the agents to work, and for the second rollout to be so much faster than the first.

The advice from their team to executives in similar positions: you don’t necessarily have to wait until every migration is complete before doing anything with AI. Identify the biggest pain points, find the most manual processes, and get some wins on the board. But be honest about which use cases your current foundation can genuinely support, because the organizations moving fastest aren’t the ones with the most sophisticated tools. They’re the ones who take the platform work seriously.

What Comes Next: How ERP Foundations Unlock the Future AI Use Cases

This organization’s efficiency gains are real and measurable, and they’re also just the beginning. The same foundation that makes a customer service agent work on day one is what makes the next use case faster to build, and the one after that faster still. For them, that means expanding into CRM data, exploring what AI can do for revenue and market growth, and pursuing capabilities that wouldn’t have been practical before. The infrastructure work that looked like table stakes a few years ago is proving considerably more valuable than that. That’s the investment that tends to compound

Velosio has worked with organizations like this at every stage of the ERP and AI journey. If you want an honest look at where your foundation stands and what’s realistic to build on top of it, Take our AI Maturity Assessment or reach out to our team.