Compare proposals apples-to-apples.
Normalize units, currency and domain terminology. Make smart assumptions and extrapolations. Extract from prose into tables. All in minutes, not days.
Ingest
AI agents handle the intake. Forward the emails with their attachments — or let an agent watch your corporate shared storage (Google Drive, Dropbox, Box, SharePoint). PDF, DOCX, XLSX, scans, multi-language; the same ingestion engine that powers Manage.
Once ingested, proposals are parsed into structured comparison data: unit pricing, FTE hours by role, monitoring plan, SAE and pharmacovigilance coverage, data management, statistics, start-up and close-out — plus the assumption footprint that lives in prose.
Compare
Outputs a comparison memo across every proposal.
- Overall assessment memo covering every vendor side-by-side
- Domain memos: Monitoring, SAE / Safety, Data Management, Statistics, Regulatory, Start-up
- Every memo surfaces what’s different, why it might matter, and what to ask the vendor
Iterate on assumptions
This is how the tool is actually used day-to-day. Change an assumption — “assume enrollment takes 14 months instead of 9” — and every proposal gets a revised total cost computed against its own pricing logic. Swap enrollment rates, site counts, country mix, dropout rates. Each iteration is a new memo.
Iteration is cheap, which means teams actually do it. Stress-test every proposal against the scenarios that have historically derailed your trials, before you sign.
Clarifying-question generator
For each proposal, a ranked list of questions to send back — grounded in specific line references of that vendor’s own document. For example:
“Page 42, CRO A assumed 28 monitoring visits per site — the protocol’s schedule suggests closer to 34. Can you clarify the underlying assumption and the cost implication?”
Ask Vera
Open-ended natural-language questions, answered across every proposal in the set:
- “Which proposal is cheapest if enrollment slips 20%?”
- “Compare pharmacovigilance coverage for Latin America across all three proposals.”
- “Where does CRO A’s site count diverge from the protocol’s target?”