ERP's next evolution isn't automation, it's creativity

Think back to how you learned mathematics at school. The teacher showed you the method, you repeated it back, and eventually you could do integrals. The knowledge was fixed. Someone programmed you, in effect, to follow a process that already existed.

Art class worked differently. The teacher might give you a topic, a bowl of fruit, perhaps a banana, but they never told you which colours to use or how to bring out what you were feeling. You explored. You went back and forth with yourself until something emerged that reflected, at an artistic level, the meaning you found in the subject.

AI sits at the intersection of those two modes of learning. On one hand, it connects to the numbers, the systems and the data, and it can access all of it with the rigour of the mathematician. On the other hand, what you then do with that information depends on your day-to-day needs and the problems you are trying to solve. That part is artistic. It becomes a creative exchange between you and the AI, between you and the solution.

This represents a profound shift for enterprise software, and for ERP in particular.

What happens to fixed business processes with AI?

For 40 years, implementing ERP followed a familiar loop. The customer talked to the consultant, the consultant talked to the system, and then the system and the customer played the scenario back to check it matched expectations. Usually it fell short, so the loop ran a few more times, and eventually you were good to go. The result was a fixed business process, locked in place.

This is process-centric design by nature. ERP systems enforce structured workflows, and that structure is what ensures consistency across the business. The trade-off is that it also limits operational flexibility. 

Hundreds of thousands of dollars, often millions, have been spent drawing out standardised business processes over those decades. That entire model is now changing. The data itself stays standardised, and it must, because trust in the numbers underpins everything. But the business processes you build on top of that data, and the problems you choose to solve, have become creative acts.

There still needs to be structure behind it. Security, approvals, and the governance  limits of what your role and responsibilities allow you to do all remain in place. The difference is that within those guardrails, you are working with tools that let you create new processes and look at data in new ways, rather than following a sequence someone mapped out years ago.

What does an AI-powered system of action actually look like?

Microsoft has a useful way of framing this: ERP is moving from systems of record to systems of action.

A system of record is a process you follow so that information gets stored in a way everyone agrees is accurate, auditable and reliable. Double-entry bookkeeping, inventory tracking, subledgers, accounts payable and receivable. All of it was built so that we could trust the number, and then act on it ourselves. The system recorded; humans acted.

Now the system can act on that information too. How you want it to act becomes a conversation in the moment. And where traditional workflows confine each process to its own lane, AI can span multiple functions and systems at once, drawing them together to deliver unified, actionable outcomes. 

Some of that action is pure automation. Think of it as the Andy Warhol end of the spectrum. Once the print is designed, you can run off a million of them. These are the things your business needs to do every day, so you create an agent, tell it to run daily, and you never need to look at it again.

Other cases sit at the Pablo Picasso end. They are peculiar, challenging and random, the unique problems nobody could have anticipated in advance. These are the moments where you experiment, throw a bunch of ideas forward, test, iterate and try new things on the fly, with the AI as your creative partner.

You could call this process creativity. Work stops being linear. The intelligence is right there in the platform, and you are free to explore what it can do for the specific problem in front of you today, which may look entirely different from the problem in front of you tomorrow. 

So here is my challenge to anyone evaluating AI in their enterprise systems. Stop asking where the solution is and where the process is, because AI works as neither in the traditional sense. Approach it the way you approached that bowl of fruit in art class. The data gives you the discipline of mathematics. What you create with it is up to you.