Case study
Draftroom — AI Proposal Studio
A generation studio that assembles first-draft proposals from a firm's past work — on-voice, sourced, and 70% faster.
01The problem
Atlas wrote every proposal from scratch even though 60% of the content — capability statements, case summaries, team bios — already existed across hundreds of past documents. Partners burned evenings on assembly work, and the best material was whatever the author happened to remember.
02What we built
We indexed ten years of past proposals and case studies into a retrieval layer, then built a drafting studio: consultants outline the pitch, the system assembles candidate sections from real past work with provenance shown inline, and a style layer keeps everything in the firm's voice. Nothing enters a draft without a visible source; partners edit in-place with tracked AI suggestions.
03The outcome
First-draft time fell 70%. Win rate on competitive bids rose four points in the first six months — partners spend review time sharpening arguments instead of hunting for boilerplate.
