Case Study · B2B PROCUREMENT

Racepoint: Multi-agent LLM procurement

Natural-language B2B procurement.

The story

What Racepoint needed, and what we built.

B2B procurement is full of multi-step friction, parse the quote request, configure the right product, calculate dynamic pricing, generate a quote. Founder Melyssa Plunkett-Gomez saw the opportunity to collapse this into a natural-language AI workflow. KUMO is building Racepoint as a multi-agent LLM SaaS.

What we delivered

Six areas of production work.

Multi-agent orchestration

Specialist agents for query parsing, product configuration, pricing, and quote generation, coordinated through structured handoffs.

Natural-language procurement UX

Procurement teams interact in plain language; agents handle the multi-step backend without exposing the orchestration.

Dynamic pricing engine

Inventory- and contract-aware pricing that adapts in real time to availability, terms, and customer-specific deals.

Automated quote generation

Quote document creation with the configuration, terms, and pricing context resolved, ready for review and send.

Per-agent eval suites

Evaluation harnesses per agent so regressions are caught in CI before they reach a customer.

Production AI architecture

Observability, fallback paths, and human-in-the-loop on edge cases, designed for production, not a demo loop.

Highlights

Project highlights.

Multi-agent Production AI orchestration
Natural-language Procurement UX
Dynamic pricing Real-time
In build Active

Tell us what you're solving for.

We'll listen first, ask the right questions, and follow up with a clear proposal.