The question isn't what AI can build. It's which problems are most valuable to solve.

That was the frame our CTO, Laurent Sorber, opened with recently at Vlerick Business School's Supply Chain Transformation program. He was there to make a case that's become increasingly clear to us through our work with operations leaders: task automation is real, but it won't move the P&L.

The highest-leverage application of agentic AI in supply chains is Enterprise Orchestration.

Here's the argument.

Task automation is real. But it won't move the P&L.

Most AI work in supply chains today is task automation. Forecasting tweaks. PO approvals. Exception handling. The technology works. The demos land. The pilots succeed.

And yet, when we sit with operations leaders, we hear the same thing: it isn't showing up in the numbers.

The math explains why. Enterprises run on hundreds, often thousands, of tasks. Automating any single one gives you a small efficiency gain. To move the bottom line, you'd need to industrialize that effort across the organisation. And a task-by-task rollout takes years. By the time you finish, the business has moved on.

This isn't a popular thing to say in 2026, when most of the AI conversation is about automation. But the math is the math. We're not against task automation. We build it. It's just not where the biggest value sits.

So where is?

Not in the tasks. In the space between them.

Supply chains are organised in silos: demand, inbound, production, outbound. Each one has its own goals, its own rules, and its own limited field of view. It sees its own situation well, and maybe a sliver of its neighbours'. Beyond that, it's flying blind.

Even when a planner wants to factor in upstream or downstream effects, they usually can't. The visibility isn't there. The decision cycle is too slow. So everyone optimizes locally, and the system as a whole quietly underperforms.

In our experience, this is where the real value leaks. Not inside the tasks. In the handoffs and coordination gaps between them.

What Enterprise Orchestration is

Enterprise Orchestration is a single system of intelligence that spans the full chain. It optimizes for system-wide KPIs, not departmental ones. And it coordinates decisions faster and better than any human planning cycle can.

A few things follow:

• It optimizes across silos, not within them.

• It runs at a cadence no human planning team can match.

• It makes second-order effects visible, and acts on them.

It treats the supply chain as one decision-making system, instead of a relay race between specialists.

Why now?

Operations researchers have wanted system-wide optimization for decades. So what changed?

Three things had to be true at the same time:

1. Context is capturable. AI can now ingest the messy, unstructured cross-system context that earlier tools couldn't see. ERPs, emails, supplier portals, shop-floor signals,… All of it can be folded into one operational picture.

2. Optimization is tractable. Better solvers, cheaper compute, and AI-driven heuristics mean decisions that used to require overnight batch runs now happen in minutes, against models complex enough to reflect the real business.

3. Custom software is affordable. Building software tailored to one enterprise's specific reality used to be prohibitive. With agentic AI in the development loop, that has flipped. The "we'll have to live with off-the-shelf" constraint is gone.

In 2026, for the first time, all three are true at once.

The right question

It's tempting to chase tools. To stack pilots. To measure progress by the number of AI initiatives running.

We don't think that's the right scoreboard. The companies that pull ahead won't be the ones with the most AI tools. They'll be the ones that ask the right question first:

Where is value leaking the most? And what would it take to stop it?

If the honest answer is a hundred small task inefficiencies, automate them. Patiently. But if the answer is the coordination gaps between your silos, no amount of point automation will close them. That's a different problem, and it deserves a different response.

We believe that response is Enterprise Orchestration.

Curious what it could mean for your enterprise? Let's talk.

For a comprehensive look at how Enterprise Orchestration works across industries, see our guide to Enterprise Orchestration.