If you’ve worked at a large company, you’ve probably heard some variation of “when we finally migrate off SAP, things will get so much better.” That sentence has been circulating for decades. The migration rarely happens.
It’s not laziness. It’s not lack of budget. And it’s not that SAP is good.
What Really Locks You In
The reason systems like SAP persist isn’t software quality. It’s the cost of accumulated context.
A large company that has used SAP for fifteen years has inside it: business-specific customizations that no consultant documented properly, workflows that make sense for historical reasons whose decision-makers have long since left, integrations with other legacy systems built one at a time over years, and data — decades of transactions, contracts, supplier history.
Migrating isn’t installing new software. It’s recreating fifteen years of organizational context, testing that every edge case still works, and training hundreds or thousands of people. The real cost of a major migration is measured in hundreds of millions of dollars and years of risk.
Why This Is an AI Problem
AI’s promise for enterprises almost always runs into this.
“Let’s use AI to automate process X” — but process X lives in SAP, on screens that only work with specific clicks in the right sequence, exporting reports in formats that need manual processing. The AI agent can’t navigate this reliably.
The alternative is to replace SAP first — and then we’re back to the migration problem.
The Solution That’s Emerging
What’s becoming viable now is different from both options. It’s not replacement, and it’s not direct integration with the terrible screens.
It’s an AI layer that wraps the existing system. SAP continues as the system of record — but human interaction is mediated by agents that understand natural language, know how to navigate the workflows underneath, and expose a modern interface.
The good metaphor: the bridge becomes the highway. SAP keeps existing, but it stops being where work happens. Work starts happening in the AI layer on top.
What This Changes
For product builders, this shift creates a window of opportunity that didn’t exist before.
For years, the only way to make money in the SAP ecosystem was to be a large consultancy or a niche player doing specialized integration. Both had very high barriers to entry.
The AI layer changes this. You can build a product that wraps SAP at a fraction of the cost it would take to integrate directly. And the customer gets ROI without facing a migration.
The context accumulation that protected SAP for decades may paradoxically become the reason a new layer has so much value. Because the more historical context the system has, the richer the material the AI can work with.