Frequently asked questions
What the agent does, who makes the decision, where your data lives, and how to start.
The agent
Klarefi is an AI intake agent. It reads the documents an applicant submits, checks facts against your requirements, and asks for missing items in a chat during the intake. Your team gets a case file with cited facts, check results, and any unresolved issues, and makes the decision.
AI intake automation uses an AI agent for the first pass on an application, insurance submission, or onboarding file. The agent collects the answers and documents, reads them, checks them against your requirements, and asks the applicant for anything missing. The file your team reviews already holds the facts, their sources, and any open questions.
MGAs handling broker submissions, factoring companies onboarding new clients, and mortgage lenders processing applications.
A fact is one piece of information your team needs to decide, such as a sum insured, a company registration number, or a gross salary. You list the facts when you set up the agent. For each one, the agent finds the value in the documents and keeps the exact quote and page it came from.
PDFs, scans, photos, emails, and attachments. One application can hold a single document or dozens, and the agent reads all of them.
Yes. An applicant can hand the intake link to ChatGPT, Claude, or another assistant. It fills in the form, answers follow-up questions, and uploads documents through the same API the hosted page uses. Your own systems can prefill what you already know through the REST API, the TypeScript SDK, or MCP. Assistants can only complete the intake. Every decision stays with your team.
Your team and your systems
Your team. The agent completes the file and shows its evidence. Your people approve, decline, or correct it. Every correction is recorded in the case's audit trail.
No. Klarefi runs the intake in front of the systems you already use, such as your policy administration system, factoring platform, or loan origination system. Case files go to them through the API and webhooks.
No. You set up the agent by describing your intake: the form, the facts you need, and the checks to run. There is nothing to train, and you can start without historical files.
In the EU by default. Applicant documents and case content are not used to train shared models. Retention, access control, and audit logs are included, and the commitments for your deployment are set out in your agreement.
Getting started
With a free intake diagnosis. You tell us how your intake runs today, where files stall, and how many you handle. From there you can build your agent for free, or plan a pilot on your own files.
That is what the pilot is for. The agent runs on your real documents against acceptance criteria you agree before it starts, and you decide on the results.
You pay per processed intake: one application that reaches your team ready to review. There are no seat licences, and your first live intakes are free. The rate depends on your volume, your workflow, and where Klarefi runs, so we set it with you after the intake diagnosis.
See a submission go from upload to review.
Build your agent for free and run a test submission with your own documents.
No credit card required