AI Extraction

When you upload a certificate, Inori's AI vision engine reads the document and extracts structured data automatically. This eliminates manual data entry and accelerates compliance review.

What the AI Extracts

The extraction engine recognizes 89 standard ACORD 25 fields, including:

• Insured information — Named insured, address, contact details.
• Producer details — Agency name, phone, fax, email.
• Coverage lines — General Liability, Automobile Liability, Workers Compensation, Umbrella/Excess. Additional coverages (Professional Liability, Cyber, etc.) are captured in the Other Coverages section.
• Policy details — Policy numbers, effective dates, expiration dates.
• Limits — Each Occurrence, General Aggregate, Products/Completed Ops, Personal & Advertising Injury, and more.
• Provisions — Additional Insured, Waiver of Subrogation, Primary & Non-Contributory indicators.
• Certificate Holder — Name and address as listed on the certificate.

Confidence Scores

Each extracted field receives a confidence score from 0 to 100:

• 90-100 — High confidence. The field is auto-accepted.
• 70-89 — Medium confidence. Flagged for review with a yellow indicator.
• Below 70 — Low confidence. Flagged with a red indicator and requires manual verification.

How It Works

1. The uploaded PDF is sent to the AI vision model.
2. The model identifies form structure and extracts field values.
3. Extracted data is compared against the project's requirement template.
4. Compliance gaps are flagged automatically.

The entire process typically completes in 5-15 seconds.