Why we’re building RFPBench.

The RFP process is broken in a symmetrical way. Issuers send out structured requests and get back unstructured chaos — responses scattered across email threads and mismatched files, compared by hand. Responders receive the same questions again and again, and answer them again and again, because their organization’s approved answers live in old spreadsheets and someone’s memory.

Both sides are doing knowledge work with the tools of document shuffling.

We’re building RFPBench to fix both sides at once: a platform where buyers create, distribute, and grade RFPs with standardized intake and consistent comparison — and where responders turn what their organization already knows into a knowledge base, draft with supervised AI, and finalize every answer with human review.

We have strong opinions about how AI belongs in this workflow. It should draft, not decide. It should show its sources on every answer. And when it doesn’t know, it should say so — plainly — instead of inventing something confident. RFPs end up in front of procurement teams, security reviewers, and auditors; software that touches them should be built for that scrutiny from day one.

We’re building in the open with our early users. If your team issues RFPs, responds to them, or both, we’d like your process to shape the product.