KlaritiQKlaritiQ
Sign inGet my free audit readiness assessment
Evidence-Based ScoringDocument-verified · deletion approval-gated
Get my free audit readiness assessment
OverviewFeaturesPlatform ViewUnder the hoodTrust TiersHow It WorksLive PreviewWho It Helps
Evidence-Based Scoring

Every score backed by something you can point to in an audit.

NABL and RBI assessors don't accept a number on its own. They accept a timestamped, user-linked record of what it was based on. That's the standard this is built to.

Every reading behind your ARI carries its own trust level, document-verified, self-reported, or inferred, and stays linked to whatever decision cited it, so an edit later can't quietly change what an old conclusion was based on.

How to track compliance with RBI's new AI governance rules →

Demo video coming soon
Upload, extract, confirm
Upload a document and AI proposes facts, risks, and milestones from it, but nothing is written to your record until a person reviews and confirms it.
Trust level on every record
Each piece of evidence is tagged document-verified, self-reported, or inferred, so leadership sees how solid a reading actually is, not a flattened single truth.
Audit Defense Report export
Export a report where every line cites the evidence or decision it came from, built for handing to an assessor, not generated narrative dressed up as one.
Deletion is checked, not silent
Before evidence is deleted, you see exactly what it created downstream, risks, blockers, verified claims, so removing a record is an informed choice.
Straight from the platform

See exactly how it works, step by step, below.

KlaritiQ Evidence Locker showing a document being read by the extraction engine alongside already-confirmed evidence, each with its readiness impact, asserter, and trust levelKlaritiQ Evidence Locker showing a document being read by the extraction engine alongside already-confirmed evidence, each with its readiness impact, asserter, and trust level
KlaritiQ extraction review modal for a Data Retention & Deletion Policy, showing the AI-generated summary and key facts pending human approval, with Discard and Approve actionsKlaritiQ extraction review modal for a Data Retention & Deletion Policy, showing the AI-generated summary and key facts pending human approval, with Discard and Approve actions
KlaritiQ Evidence Locker after confirming a Data Retention & Deletion Policy document, now listed under DPDP Data Retention Policy Hardening with its own linked evidence and readiness figureKlaritiQ Evidence Locker after confirming a Data Retention & Deletion Policy document, now listed under DPDP Data Retention Policy Hardening with its own linked evidence and readiness figure

The Evidence Locker: every document, its extraction status, and the readiness impact it's already confirmed.

How it works
1
Upload a document
A policy, audit report, or vendor contract is uploaded and queued for extraction.
2
AI proposes, doesn't write
The extraction engine reads the document and proposes facts, risks, and milestones. Nothing is saved yet.
3
A person reviews and confirms
The extraction preview requires a human to approve before anything becomes a real evidence row. Reject and nothing is written.
4
Evidence stays linked
Confirmed evidence carries its trust level, its source document, and every decision or claim that later cites it, so the trail holds together.
Under the hood

The values below are a sample. The field names and structure are the same ones the evidence ledger actually writes.

evidence.json
{
  "id": "ev_9f2a1c",
  "type": "document",
  "assertedByType": "user",
  "trust": "high",
  "title": "DPDP Data Retention Policy v2.1 - Legal Review",
  "sourceRef": "document:8e21ac4f",
  "sourceEvidenceId": null,
  "createdAt": "2026-08-14T09:32:00Z",
  "deletedAt": null
}
Three trust tiers, never collapsed into one number

Every reading behind your score keeps its own trust level, visible, not flattened into a single undifferentiated truth.

Document-verified
Came from an uploaded, reviewed document. Carries the highest trust weight in the score.
Self-reported
Entered directly by someone on your team, with no document behind it. Flagged for periodic review.
Inferred
Derived from other evidence already on file rather than stated outright, weighted accordingly.
See it in the platform
Evidence trust level · Illustrative
Governance Sub-Score78/100
Document-verified · deletion approval-gated
Frequently asked
No. AI can only propose facts extracted from an uploaded document. A human has to review that proposal in the extraction preview and confirm it before it becomes a real evidence record.
Who this helps
The CTO
Sees which claims about the org's AI posture are document-verified versus self-reported, instead of one undifferentiated number.
The Compliance Head
Gets an Audit Defense Report where every line traces to a real evidence row, ready to hand to an assessor rather than assembled the night before.
The Board
Sees a record that survives scrutiny: evidence stays linked to the decisions and scores it supported, even as new evidence is added over time.
Built for Indian Enterprises
Pilot partners, coming soon
DPDP Act-aware
ISO 42001 · on roadmap
SOC 2 · on roadmap
Related products
Risk Management
A live AI risk register tied to your readiness score, not a spreadsheet updated before an audit.
Decision Records
Know who approved what, and why, permanently.
Milestone Evidence Gates
A milestone closes on verified evidence, never a checkbox flipped on someone's word.

Ready to see where you stand?