Head of Ops, growth-stage SaaS

Marcus Chen runs ops at a 350-person SaaS company and was one of the first customers to put Maya on a non-AP workflow. We asked him what he's learned.

You started with vendor onboarding, not AP. Why?
Our AP was already passable. Vendor onboarding was the bottleneck. Every new contract triggered three days of back-and-forth across procurement, finance, and IT. We had vendors waiting two weeks to get paid because their record wasn't set up yet.

What changed in the first month?
The first week, nothing visibly. Maya was learning our setup. Then around day ten, new vendor requests started clearing the same day. By the end of the month, my procurement lead asked me what she should be working on instead. That's a good problem.

What did you not expect?
How much of the slow part was just waiting on internal sign-offs we didn't actually need. Maya forced us to look at every step and ask if a human really had to approve it. Most didn't. The platform was just a mirror.

What do you watch for now?
The exception list. If it stays short, we're fine. If it grows, something upstream changed — a new vendor type, a contract change, a policy update. Maya doesn't fail loudly. It just routes more things to humans, and that's the signal.

Advice to ops leads thinking about agents?
Don't pick the workflow that hurts most. Pick the one that's most ready. The one with the cleanest data and the simplest exceptions. Get a win, then the rest is easier.