Spendflo case study
Agentic AI procurement
Spendflo is an AI-native procure-to-pay platform covering intake, vendor onboarding, third-party risk, purchase orders, contracts, invoices and payments. turingLabs runs design for Spendflo as a fractional team: a design lead plus an embedded designer working inside its product and engineering rhythm.
When Spendflo moved to agentic procurement, its co-founder says, the product structure was a mess and design only saw decisions after they were made. turingLabs designed the agentic source-to-pay harness and built a design system the team sticks to.
Agents people can set up and trust
The agent work covers two connected surfaces. Agent Studio is where procurement experts write, configure, test and deploy their own agents without engineering help. The response system is where people work with those agents: about 35 response types across three groups, readable, interactive and actionable. Around them sit versioning, an audit trail of every run, and permissions that follow the roles a company already has.
If you’re a founder who wants design to actually move the business, work with these guys.
When Spendflo moved to agentic procurement, our product structure was a mess, and design only saw decisions after they were made. Ankit and the turingLabs team changed that. They designed our entire agentic source-to-pay harness, built a design system we actually stick to, and treated it like their own product from day one. I never had to write a big scope doc or explain anything twice. Today they know as much as our internal team, and half the time they’re ahead of us!
They’ve built AI systems that let them iterate at near product fidelity, so most of it is tested and refined before engineering writes a line. The real test was launching Flo, our AI agent for procurement. We picked a date, worked backwards, and made it.
Design was a big reason we hit it, and I wouldn’t start our next big build without them.
Siddharth Sridharan, Co-founder, Spendflo
More work
