Skip to content

The platform

How Papertrail works, end to end.

A candidate goes in at one end and a signed, evidence-backed report comes out the other. No implementation project, no separate portal per check type, no black box in between — every stage below is something you can open and see for yourself.

  • 50+checks available
  • <24htypical turnaround
  • 20+sources checked per case
case/PT-24817/dispatch0/4 answered
Papertrail core
Court recordsNo case foundquerying…
Credit bureauNo defaultquerying…
University registrarSerial matchedquerying…
Global watchlists0 of 214 hitquerying…

Every check dispatches at once. A slow registrar delays its own answer, never the case.

The journey

Five stages, one case file.

Every case moves through the same five stages regardless of which checks are in it. Pick a stage to see what actually happens inside it.

Following the hand-off — pick a stage to hold it

The reliability layer

What catches it when the AI is wrong.

An AI-native product that never says so is the enterprise objection this section exists to answer. Here is what actually stands between an automated read and a client-facing verdict.

A wrong automated read enters here. It has to survive all four gates below before it can reach a client's report.

Walking the gates — pick one to hold it

FAQ

The platform, answered

  • You can create an account and run your first case from the dashboard the same day, with no integration required. If you want candidates to flow in automatically from your ATS or HRMS, our pre-built connectors typically go live within a day; a custom API integration usually takes about a week.

  • Alongside it. Papertrail runs the checks, reconciles the evidence and surfaces discrepancies — your team still makes every hiring decision. Analysts on our side handle the parts that need a person: reviewing a flagged document, calling an employer who won't respond to a portal request, reading a court record that didn't digitise cleanly.

  • In Indian regions, for the life of the case and the retention window you configure. Sub-processors are documented, documents are reachable only through signed, expiring URLs, and the retention schedule purges data on your timetable — with the purge itself written to the audit log.

  • It doesn't publish. Every automated read below its confidence threshold escalates to a trained analyst, who resolves it against the primary source before it reaches your report. Nothing reaches a client-facing verdict on a machine guess alone.

  • Yes. Packages are built per role or location and can be edited at any point — add a court record search for a finance hire, drop education verification for a role that doesn't need it. Nothing about the platform assumes a fixed package.

See it on your own cases

Walk through your first case with us.

Bring a real role and we'll show you the package, the turnaround, and the evidence trail it produces — on your data, not a demo script.