Insurance Certificate Compliance Check

Insurance Certificate Compliance Check

Compares Access Agreement insurance requirements to COI and endorsement evidence to produce a defensible compliance determination with actionable deficit remediation.

Prerequisites

1. Access Agreement (final) with insurance clause, exhibits, and amendments
2. COIs for all required lines: CGL, auto, WC/EL, umbrella, specialty
3. Endorsement copies: AI, completed-ops AI, primary/non-contributory, waiver of subrogation, umbrella schedule
4. Recipient legal names, required AI entities, project scope
5. Access dates, locations, post-completion carry-forward period
6. Governing law / jurisdiction
7. Delivery recipients and privilege instructions

Quick Start

1. Extract all insurance requirements from the agreement into a requirements matrix
2. Run COI baseline checks against each requirement
3. Verify endorsement-level evidence for every core element — COI text alone is non-binding
4. Assess risk and coverage-structure gaps
5. Apply jurisdiction-specific considerations
6. Output compliance determination with deficit list

Workflow

Step 1: Build Requirements Matrix

Extract from agreement into normalized rows:

Flag vague clauses (acceptable, sufficient, as required by owner) as Attorney Clarification Required.

Step 2: COI Baseline Checks

These are information-only — not proof of coverage:

1. Named insured matches recipient entity exactly (name + suffix + related entity logic)
2. Policy period covers access window
3. Required coverages and limits appear in correct fields
4. Occurrence vs claims-made confirmed
5. Aggregate designations align (per occurrence, per location, per project)
6. Producer/broker contact present
7. Red flags: expired/expiring policies, wrong entity, missing policy numbers

If incomplete, set Conditionally Compliant and list missing documents.

Step 3: Endorsement Verification

COI text alone is non-binding. Require endorsement-level evidence for each:

Step 4: Risk and Coverage-Structure Checks

1. Verify minimum limits vs exposed values per line and project
2. Compute combined limits where umbrella augments GL/auto/EL
3. Check for coverage-exclusion conflicts (geographic, type-of-work, access-specific)
4. Review deductibles/SIR if high-risk or large retention (from declarations/broker attestation)
5. Confirm insurer rating with explicit evidence

Step 5: Jurisdiction Adaptation

Apply [VERIFY] to all jurisdiction-specific conclusions:

• New York: Construction gravity-risk and labor-law exclusions [VERIFY]
• Texas: Anti-indemnity impact on AI enforcement [VERIFY]
• California: Active-negligence/AI enforceability constraints [VERIFY]
• Florida: Policy disclosure and insurer-info rights [VERIFY]

If governing law is uncertain, mark all jurisdictional conclusions as requires legal verification.

Step 6: Output

Result: [Compliant | Conditionally Compliant | Non-Compliant]
Jurisdiction: [state]
Coverage period tested: [start] to [end]
Review confidence: [High | Medium | Low]

Requirements Matrix:
- Requirement | Evidence | Status | Finding | Deficiency
- ...

Escalation: [In-house counsel / Risk manager / Coverage counsel]
Access decision: [Recommend allow / conditional / deny pending docs]

Use entity-name exactness and numeric precision ($1,000,000 each occurrence, CG 20 37).

Critical Guardrails

• COI statements are documentary indication, not contractual proof
• Never state "fully covered" without endorsement/policy support
• Certificate holder ≠ additional insured — never conflate
• Never infer AI scope from checkbox or Description of Operations alone
• For each deficit, provide precise broker instruction: exact endorsement form/edition and deadline
• Track status as confirmed, indicated-not-verified, or not shown
• This is a risk-control assessment, not a coverage legal opinion
• Include attorney review requirement before operational reliance
• If access allowed with unresolved deficits, document written exception and mitigation
• If anti-indemnity/statutory effects drive outcome, escalate immediately
• Use [VERIFY] on all uncited or uncertain authority claims