Nova Intelligence

The premise is straightforward once you accept the boring parts. Big companies run their payroll, their supply chains, and their finances on SAP. Over the decades they have written mountains of custom code - much of it in ABAP, a language most twenty-something AI engineers have never heard of - to make SAP do the specific weird things their specific weird business needs. This code works. It also becomes a problem the moment anyone wants to change anything.

Now everyone wants to change something. SAP is pushing customers onto S/4HANA, its modern platform, and has set a deadline of 2030. That deadline used to be 2027, which sounds like relief until you realize it just means a bigger backlog. Analysts peg the migration market - implementation, upgrades, support - at north of $89 billion. Emma Qian, Nova's CEO, says she has met companies budgeting over a billion dollars just to migrate. One billion. To move software from one place to a slightly newer place.

The traditional way to do this is to hire an army of consultants who read your custom code line by line and decide what to keep, what to rewrite, and what to throw away. This is slow, expensive, and error-prone, which is another way of saying it is a perfect job for software. Nova's pitch is that agentic AI - not one model answering one prompt, but a system of AI agents that analyze, document, refactor, and generate code - can do a large chunk of that work.

Crucially, Nova doesn't just rewrite your custom code. It tries to map it onto SAP standard functionality wherever possible, so you end up with less custom code than you started with. In SAP-land this is called keeping a "clean core," and it is the difference between renting complexity forever and actually owning something simpler. That is the sales pitch, and enterprises with billion-dollar migration budgets are, it turns out, willing to listen.