Help & Reference

Log in with your credentials. The Dashboard is your home screen — it shows aggregate metrics for both compliance checks and policy verifications.

Go to Documents and upload at least one Regulation PDF (e.g. an IRDAI guideline) and one BRD (Business Requirements Document). Wait for each file to finish processing before continuing.

Go to Compliance and run a check: select your BRD, choose which regulations to compare it against, then click "Run Check". The system will extract, compare, and score the document.

Open the resulting Report to see the compliance score, per-regulation findings, and the Policy Reasoning Trace.

To verify a policy document (e.g. a customer-facing policy PDF), go to Policy Verification, upload the PDF, and wait for the scorecard to appear.

Use Rule Studio to manage the rules that drive policy verification. Activate, suspend, or create new rules without re-uploading any documents.

The top KPI row shows Total Checks, Average Score, Regulations Loaded, and Non-Compliant count for the compliance workflow.

The "Compliance Score Trend" area chart plots the overall score of each completed check in chronological order.

The "Status Distribution" donut shows the ratio of compliant, non-compliant, and insufficient-info results.

The "Recent Activity" table lists the last 5 compliance reports. Click any row to open the full report.

The "Policy Verification" section below shows the same type of metrics for uploaded policy PDFs: total verified, average pass rate, failed count, and grade distribution.

All data refreshes automatically every 30 seconds. To force a refresh, reload the page.

Click "Upload Document" and select a PDF file. Choose the document type: "Regulation" for regulatory guidelines, "BRD" for internal business documents.

After upload, the document enters the processing pipeline: text is extracted, chunked, embedded, and indexed into the vector and graph databases.

Wait for the status to show "ready" before using the document in a compliance check. Large PDFs may take 30–120 seconds.

Documents in "failed" status could not be processed — check that the file is a valid, non-corrupted PDF with extractable text (not a scanned image).

To delete a document, use the delete button on the document card. This also removes its entries from the knowledge base.

Step 1 — Select BRD: Choose the BRD you want to check. Only documents with status "ready" are listed.

Step 2 — Select Regulations: Choose one or more regulatory documents to compare against. You can select multiple.

Step 3 — Run Check: Click "Run Compliance Check". The system sends both documents to the multi-agent pipeline (Builder → Legislator → Judge) and scores each regulatory requirement.

When complete, you are redirected to the Report. A compliance check typically takes 30–90 seconds depending on document length.

Scores range from 0–100. A score above 70 is generally considered compliant, though this threshold depends on your organisation's standards.

The Reports list shows all completed checks with the BRD name, overall score, status, and date. Click any row to open the full report.

Inside a report, the header shows the overall score and status. Below it, each regulation's findings are listed in a collapsible panel.

Each finding includes an extracted clause, the applicable regulation reference, and the Judge's reasoning.

The "Policy Reasoning Trace" (PRT) section shows the full chain of reasoning the AI used — every step from extraction to verdict. This is the audit trail.

To export, click "Print / Save PDF" in the header. The page uses print-optimised CSS with the CheQ header and branding.

Type a question in plain English in the query box and press Enter or click Send. Example: "What are the capital adequacy requirements under IRDAI?"

The system retrieves the most relevant chunks from indexed regulations using hybrid vector + graph search, then synthesises an answer.

Citations appear below the answer — each citation links to the source document and the specific passage used.

For best results, ask specific questions about requirements, definitions, or thresholds. Vague questions ("tell me about IRDAI") produce broad summaries.

The query interface searches only documents that have been successfully ingested. If a document is missing from results, verify it has status "ready" in Documents.

Go to Policy Verification and click "Upload Policy PDF". Select one PDF (or up to 10 for batch mode). Only PDF files are accepted.

Pass the X-Org-Id header (or select the org in the upload form) to use the correct extraction schema. Supported: hdfc-ergo, prudent, aditya-birla.

The system runs a two-layer pipeline: (1) extract text → extract structured fields using the org's Schema Studio schema → evaluate business rules (Document QC layer); (2) evaluate regulatory ACU rules with MiniPRT reasoning (Regulatory Compliance layer).

Once status is "validated", click the document row to open the detail view. Two tabs are available: "Document QC" and "Regulatory Compliance".

"Document QC" tab shows extracted fields, standard rule results, and the business scorecard. Each rule result shows PASS / FAIL / WARN / SKIP with evidence.

"Regulatory Compliance" tab shows IRDAI regulation-derived rule results, each with a MiniPRT: evidence found in the document, page reference, reasoning chain, and COMPLIANT / NON-COMPLIANT / UNABLE_TO_DETERMINE verdict with confidence score.

If you change rules in Rule Studio, use the "Re-validate" button to re-run rule evaluation without re-uploading the PDF (~5 seconds).

Export a branded PDF report using the "Export PDF Report" button in the detail view header.

The Rules Library tab lists all rules. Use the status filter to view "active", "draft", or "suspended" rules.

Each rule card shows the rule name, type (deterministic or semantic), severity, and current status. Use the toggle to activate or suspend a rule without deleting it.

To create a new rule, go to the "Create Rule" tab and type a plain-English description of what you want the rule to check. Example: "The sum insured must be at least ₹5 lakh."

The AI analyses your intent, asks clarifying questions if needed, then compiles the rule. Review the compiled rule before saving.

Use "Sandbox Test" on any rule to run it against recent verified documents and see how it would perform before activating it.

The "Rule Packs" tab lists pre-built YAML rule collections (e.g. IRDAI Life Basics, Premium Arithmetic, HDFC Ergo Custom). Click "Install" to add all rules in a pack.

After installing a pack or changing rule status, use "Re-validate" on existing policy verifications to see the updated results immediately.

Go to Schema Studio. The Schemas tab lists all extraction schemas in the system, filterable by organisation (HDFC Ergo, Prudent, Aditya Birla) and status (active / draft / deprecated).

Each schema card shows the schema ID, organisation, version, field count, and status. Expand a card to see all fields grouped by category (identity, financial, dates, terms, etc.).

To activate a schema so it is used for new uploads, expand the card and click "Activate". Only one schema per org can be active at a time.

To create a new schema with AI, click the "Create with AI" tab. Select the organisation, describe what you want to extract in plain English, and click "Generate with AI".

If the AI needs more information, it will ask clarifying questions. Answer them and click "Generate Schema". Review the generated fields and click "Save Schema" to store it as a draft.

Activate the new draft schema from the Schemas list. Future uploads for that org will use the new schema immediately.

To deprecate an old schema, expand the card and click "Deprecate". Deprecated schemas are retained for audit purposes but not used for new extractions.

The Knowledge Graph page shows live node and edge counts from the graph database (FalkorDB).

Every regulatory document that is ingested is parsed for entities (sections, clauses, definitions, requirements) and the relationships between them.

These relationships enhance the Query feature — when you ask a question, the system uses both vector similarity and graph traversal to find relevant passages.

Entity Search lets you look up specific concepts (e.g. "premium", "exclusion", "sum insured") and see which regulations reference them.

The graph is populated automatically during document ingestion. No manual action is required.

Connector Hub is for PAS API integrations only — Duck Creek, Oracle OIPA, TCS BaNCS, or any custom REST API that exposes product or policy data.

Click "+ Add Connector", choose a pre-built template (Duck Creek, Oracle OIPA, TCS BaNCS) or start from scratch.

Fill in the Endpoint URL, Organisation ID, and authentication credentials (API Key or Bearer token).

Use "Test Connection" on the connector card to verify the endpoint is reachable and returns a valid response.

To extract fields and rules from a policy or product filing PDF, use Pack Studio → "From Document" tab instead — that is the correct ingestion path for document-based content.

Every ACU (Atomic Compliance Unit) has an intrinsic severity — CRITICAL, HIGH, MEDIUM, or LOW — assigned by the AI during extraction, aligned with IRDAI's weighted risk formula.

When a compliance check or policy verification runs, each finding's severity is computed from a deterministic matrix: a CRITICAL ACU that is NON-COMPLIANT becomes a CRITICAL finding; a CRITICAL ACU that has INSUFFICIENT_INFO becomes a HIGH finding; any COMPLIANT finding is INFO.

The Risk Breakdown summary (e.g., "2 CRITICAL · 1 HIGH") appears at the top of every compliance report and policy verification result. Use this to triage the most important issues first.

Use the severity filter bar on report detail pages to focus on CRITICAL or HIGH findings only — this is the fastest path to the highest-risk gaps.

Dashboard → Risk Distribution card shows your aggregate severity profile across all recent checks.

You can manually override an ACU's intrinsic severity in the ACU Explorer (Compliance → ACU Explorer tab → edit severity field). Changes take effect on future checks.

When you upload a new or updated regulation PDF, CheQ automatically compares the extracted ACUs against existing ones using SHA-256 content hashing and semantic similarity.

Changes are classified by the MURCIA taxonomy: ADDED (new obligation), MODIFIED (rewording detected — cosine similarity > 0.75 but content differs), SUPERSEDED (same rule code, new version replaces old), or UNCHANGED (identical).

The Dashboard shows an amber "Regulation Updates" card when pending changes exist, listing how many BRDs were checked against now-stale regulation versions.

Navigate to Compliance → Regulation Updates tab to review all detected changes. Each event shows the change type, similarity score, affected BRD count, and a summary.

Click "Acknowledge" to mark a change as reviewed, or "Resolve" once you have re-run the relevant compliance checks against the updated regulation.

Filter events by status (Pending / Acknowledged / Resolved) to manage your regulatory change backlog.