Try Corpus on your next real pursuit
$5,000 fixed, about 2–3 weeks. We load a copy of your recent project PDFs, run the RFQ you are chasing through Corpus, and you judge the output on your own pursuit. Credited toward setup if you continue.
Free, 30 minutes, with Yermek or Pavel. We pick the RFQ together and agree which PDFs to copy.
The $5,000 is credited toward full deployment.
Four things, and not much of your time.
- A slice of your project PDFs, for example the last 2–3 years (we work from a copy)
- The RFQ you are pursuing now
- One 30-minute kickoff
- One review call at the end
Real output on a real pursuit.
- A go / no-go scorecard: suggested scores linked to their sources; fit, relationship and capacity stay your call
- An SOQ draft, every sentence cited to the page
- A fee build, with the fee math done by code
- A hands-on session for your team on your own archive
Your PDFs, handled like project records.
- We work from a copy. Your originals are not touched, and nothing connects to your systems.
- The copy is processed on our own GPU server or on a GPU box on your network, your choice, with an open-weight model. Nothing goes to a public AI API.
- Your files are never used to train a model.
- Only the people you name get access, and role-based access applies from day one.
- When the pilot ends, we delete the copy within 7 days and confirm it in writing. If you continue, we load from your live archive during setup.
Checks your team can run itself.
- Open any 10 citations. Each should land on the page that says it.
- Ask 3 questions you know the archive can’t answer. Corpus should say “Nothing on file”.
- Check the fee build line by line against your own spreadsheet.
- Search a restricted project as someone who can’t open it. Nothing should come back.
- At the review call: did the go / no-go match what your principals would have decided, and how much of the SOQ draft survived your team’s edit?
If the answers are no, you stop there. The output is still yours.
About three weeks, step by step.
Load and index
Kickoff call. We load the copy of your PDFs and index them with role-based access applied.
Run the pursuit
We run your live RFQ through Corpus: go / no-go, SOQ draft, fee build. Your team tries cited search on its own archive.
Review and decide
Review call. You judge the output on your own pursuit and decide whether to continue.
No integrations. No production rollout.
The pilot works from a copy of your PDFs. Connecting to your network drive, SharePoint or ProjectWise, delivery in Microsoft Teams and rollout to the whole firm are the full deployment.
Continue, and the $5,000 is credited.
Full deployment: setup from $21,000, then $1,500–$2,000/mo. Deployed on your archive where it lives, live in 6–8 weeks, no per-user fees. Or stop after the review call; the output is yours to judge.
→ See full deployment detailsWhat firms ask before they start.
No RFQ open right now?
Bring the last one you bid or passed on. We run it as if it had just arrived, and you compare Corpus with what you actually did.
How much of our time does it take?
A 30-minute kickoff, someone to send the PDFs, and a review call at the end. Most of the work is ours.
What if the archive doesn’t have the answer?
Corpus says “Nothing on file”. You see where your archive is thin, which is useful on its own.
Can we include scanned drawings?
In the full deployment, scanned sheets are OCR’d as part of setup. For the pilot, text-searchable PDFs work best, so start with those. If your older work is mostly scans, tell us on the call and we will plan the pilot around it.
Who runs the pilot?
Yermek Ibray and Pavel Dolgov, the founders. You won’t be handed off.
Pick the RFQ. We will do the rest.
Book a free 30-minute pilot call and bring the pursuit you want to test.