AI search for insurance policies, claims files and underwriting guidelines
Corpus is AI search over your own policy forms, endorsements, claims files and underwriting guidelines. Ask in plain words and get a short answer cited to the document and page. If nothing on file supports it, it is designed to say “Nothing on file.” Coverage, claim and underwriting decisions stay with your licensed people.
The answer is on file. Finding it is the work.
One question about one policy can mean opening the form, three endorsements and a claims file that is half scans.
- “Which endorsement on this policy changes the water damage exclusion?”
- “Have we handled a claim like this one before, and what did the adjuster note?”
- “What do our underwriting guidelines say about vacant buildings?”
- Policy forms, endorsements, claim notes, adjuster reports and guideline manuals live in different places and editions.
Answers from your own files, cited to the page.
- We load a copy of your policy forms and endorsements, claims files, adjuster reports and underwriting guidelines, saved as PDF. Corpus is retrieval-augmented generation (RAG): it answers from those files.
- Current You ask in plain words. Corpus searches by exact words and by meaning, then writes a short answer with citations to the document and page. Open a citation and the page opens with the passage highlighted, so you can check it.
- Current If nothing on file supports an answer, it is designed to say “Nothing on file” instead of guessing. An empty result does not prove that no such record exists.
- Current Access by role: files a role cannot open are removed before search runs, so answers never draw on them. In the demo, roles are illustrative and anyone can switch them. Set up for you Tying roles to your named people is part of full-deployment setup.
- Current Version history for drafts: each saved version shows who saved it, Compare shows what changed, and you can restore an earlier one. In the demo this covers fee, SOQ and contract drafts.
- Corpus reads PDF files, including scanned pages read by OCR. Set up for you OCR of your own scans is planned on the call and done during setup.
- Set up for you Corpus is cloud-agnostic: it runs on AWS, Azure, Google Cloud, or your own servers. It uses an open-weight model and no public AI API.
Example questions, not results
These questions are examples written for this page. They are not results: Corpus has no customers in this industry yet, and the live demo is built on engineering documents.
- Which endorsements on this policy change the definition of “occurrence”?
- What do our underwriting guidelines say about roofs older than 20 years?
- In this claims file, what did the independent adjuster report about the cause of loss?
- Which past claims files mention mold remediation under this policy form?
Corpus finds and cites. People decide.
- A licensed person reads the cited policy language before taking any coverage position.
- Whether the cited form or guideline is the edition in force for that policy period.
- State rules and filings. Corpus answers from your files, not from current regulations.
- Reserves, payments and underwriting decisions, which stay with your people.
Corpus is not insurance or legal advice. It does not decide coverage, set reserves or approve a risk.
Start with a free call. Then one real project.
- A free 30-minute call first. We talk about your files and the questions you need answered.
- Then a $5,000 pilot on one real project or pursuit, at a fixed price.
- We work from a copy of that project’s PDFs. Your originals are not touched.
- You ask your own questions, open the citations and judge the answers yourself.
- The $5,000 is credited toward setup if you sign a deployment within 60 days of the pilot's end.
- Our commitment In the pilot, your files are never used to train or fine-tune a model. In full deployment, fine-tuning on your files happens only if you agree in writing.
- Our commitment When the pilot ends, we delete the copy within 7 days and confirm it in writing.
See how citations and refusals work.
The live demo is built for engineering firms: public engineering documents filed under fictional sample firms. Its search also reaches a public-domain collection of U.S. court opinions. There is no demo built on another industry’s files. It shows what works the same in any industry: page citations, “Nothing on file” and roles.
→ Corpus by industryFrequently asked questions
Can Corpus decide whether a claim is covered?
No. It finds the policy language and the claim documents and cites the pages. Your adjusters and coverage counsel decide.
Which policy period does an answer use?
The one in the documents it cites. Each citation shows the file and the page, so your reviewer checks the edition and the dates. Corpus does not show an effective date on its own today.
Can claims files be kept away from some staff?
Yes, by role. Files a role cannot open are removed before search runs, so answers never draw on them. In the demo, roles are illustrative and anyone can switch them. Set up for you Tying roles to your named people is part of full-deployment setup.
Do you have customers in this industry?
Not yet. Corpus has no customers in this industry. Engineering firms are our first, worked-out example, and the live demo uses fictional engineering firms.
Where do our files go?
Set up for you Corpus is cloud-agnostic. It runs on AWS, Azure, Google Cloud, or your own servers. It uses an open-weight model, and nothing goes to a public AI API. In the pilot, your files are never used to train or fine-tune a model. In full deployment, fine-tuning on your files happens only if you agree in writing. We don't have a SOC 2 report, which is one reason you can choose your own servers.
Have a folder full of questions?
Book a free 30-minute call. We talk about your files, the questions you need answered, and whether a pilot makes sense.