How to Extract Data From an ACORD Form

To extract data from an ACORD form, run the PDF or scan through software that reads the form with OCR, identifies the applicant, the named insured, the policy numbers, the coverage limits and effective dates, and the certificate holder with AI, and outputs structured fields your agency management system or policy database can use. A dedicated ACORD forms OCR tool does exactly this for ACORD 25 certificates and 125 and 126 submissions. A CSR or underwriter no longer has to read an ACORD 125, 126, or 25 line by line and re-key the applicant name, FEIN, classification codes, limits, and producer details into the system. Upload an ACORD form above and the tool pulls the header, the coverage sections, the limits columns, and the holder information into clean rows you can review and export in minutes. This guide covers which fields to extract from the common ACORD forms, how extraction handles their variable layouts and handwriting, and how to route the data straight into your submission or COI workflow.

Last updated June 2026.

What data should you extract from an ACORD form?

Extract the fields your submission, rating, or certificate process actually uses, which differ by form. From an ACORD 125 (Commercial Insurance Application) you pull the applicant name and mailing address, FEIN, NAICS and SIC codes, business type, the proposed effective and expiration dates, the lines of business requested, prior carrier and loss history. From an ACORD 126 (Commercial General Liability Section) you pull the classification codes, premises and operations exposures, rating bases, and products/completed-operations details. From an ACORD 25 (Certificate of Liability Insurance) you pull the producer, the insured, each insurer and policy number, the coverage types with their per-occurrence and aggregate limits, the effective and expiration dates, and the certificate holder. The goal is a clean, structured record of every field the form captures, not a flat dump of text.

A typical ACORD 125 carries roughly 60 to 70 mapped fields across the applicant, prior-carrier, and loss sections, so manual entry is slow and error-prone. Pulling those fields automatically is where the time savings come from.

How does AI extract data from an ACORD form?

AI extraction reads an ACORD form the way a person does: it locates each labeled box, understands what the value means in context, and maps it to a named field. OCR converts the pixels of a scanned or photographed form into characters, then a document model trained on ACORD layouts identifies which characters belong to the applicant name, which to a policy number, and which to a coverage limit. Because the model understands position and label rather than matching a fixed template, it handles the small layout differences between form versions and between carriers.

The practical workflow looks like this:

• Upload the ACORD form as a PDF, image, or scan, in batches if you process many submissions a day.
• Recognize and classify the form so the software knows whether it is a 125, 126, 25, or another ACORD type.
• Extract the header, coverage, limits, and holder fields into a structured set of rows.
• Review the fields the model flags with low confidence, which is usually a handwritten entry or a poor scan.
• Export the clean data to Excel, CSV, JSON, or straight into your agency management system through an API.

How does extraction handle variable ACORD layouts and handwriting?

Extraction handles variable layouts by reading fields by their labels and positions rather than by a single fixed template, so a 2013 ACORD 125 and a 2024 version map to the same output fields even though the boxes moved. Handwriting is the harder part: AI models recognize hand-printed entries far better than legacy OCR, but a sloppy signature or a cramped handwritten limit will still drop below the confidence threshold. The reliable pattern is confidence-based routing, where high-confidence fields flow straight through and the 3 to 5 percent that fall below threshold queue for a quick human check, so you get automation speed on the bulk of each form and accuracy on the rest.

What is an ACORD form?

An ACORD form is a standardized insurance document published by ACORD, the nonprofit standards body for the insurance industry, used to exchange consistent information between agents, brokers, carriers, and underwriters. Because every carrier accepts the same numbered forms, an ACORD 25 from one contractor looks like an ACORD 25 from any other, which is exactly what makes the data predictable enough to extract at scale. ACORD maintains hundreds of forms across commercial, personal, and life lines.

What is the difference between ACORD 125 and ACORD 126?

The ACORD 125 is the master commercial insurance application that captures general business and policy-request information, while the ACORD 126 is the Commercial General Liability supplement that adds detailed liability exposures and coverage selections. The 125 alone is not a complete submission; agents pair it with the 126 (and often the 140 for property) when general liability coverage is requested. For extraction, that means a full commercial submission usually arrives as a packet, so the software has to classify and extract each form in the bundle separately and then stitch the fields back together.

What is an ACORD 25 form?

An ACORD 25 is a Certificate of Liability Insurance that provides proof of coverage to a third party, such as a landlord or general contractor, for a contractual requirement. It lists the producer, the insured, each insurer and policy number, the coverage types with their limits, the policy effective and expiration dates, and the certificate holder. The ACORD 25 does not amend coverage and is not an insurance contract; it reflects the policy as of the date issued. For COI tracking teams, the high-value extracted fields are the policy numbers, the limits, the expiration dates, and the certificate holder, since those drive compliance checks and renewal reminders.

How do you read an ACORD 25 form?

Read an ACORD 25 from the top down: the date and producer block identify who issued the certificate and who to call with questions, the insured block names the covered business, and the insurer block lists each carrier providing coverage. Below that, each coverage row (general liability, automobile, umbrella, workers' compensation) shows the policy number, effective and expiration dates, and the per-occurrence and aggregate limits. The Description of Operations box near the bottom is where producers add additional insured language, project locations, and endorsements, so review it closely. The certificate holder block at the bottom names the party receiving the certificate.

Are ACORD forms standardized?

Yes, ACORD forms are standardized. ACORD develops common forms and data standards so agents, brokers, carriers, and underwriters all exchange the same fields in the same places, which removes ambiguity and keeps submissions consistent across the industry. That standardization is what makes ACORD forms a strong fit for automated extraction: the field set is known in advance, so a document model can map it reliably even when a particular carrier's print run shifts the layout slightly.

Can you extract data from a scanned or handwritten ACORD form?

Yes. Modern AI extraction works on scanned PDFs, photographs taken on a phone, and forms with hand-printed entries, not just clean digital PDFs. Scan quality still matters: a crisp 300 DPI scan extracts more cleanly than a skewed phone photo, and dense handwriting is the most common reason a field gets flagged for review. The workflow accounts for this by surfacing low-confidence fields for a fast human check rather than silently guessing, so a stack of mixed-quality ACORD forms still processes end to end.

How extraction fits an agency, MGA, or carrier workflow

For an agency, the payoff is in submission intake: a new commercial application arrives as an ACORD 125/126/140 packet, the data lands in the agency management system without a CSR re-keying it, and the producer moves straight to quoting. That single change is how agencies reduce manual ACORD form data entry and reclaim the hours CSRs lose retyping forms. For an MGA or wholesaler, batch extraction clears the daily stack of submissions faster and feeds rating with cleaner data. For a carrier's underwriting desk, structured ACORD data flows into the policy system and underwriting rules without manual entry, the same way insurance claims automation reads a loss file on the claims side. And for any team tracking certificates, extracting the policy numbers, limits, and expiration dates off incoming ACORD 25s is what makes compliance tracking and renewal follow-up automatic. The throughline is the same: take a standardized but manually processed form and turn it into structured fields your systems can act on.

The bottom line

ACORD forms are standardized, high-volume, and re-keyed by hand in most agencies, which is exactly the combination automated extraction is built for. Run your ACORD 125, 126, and 25 forms through AI extraction, route the handful of low-confidence handwritten fields to a quick review, and push clean applicant, policy, coverage, and certificate data into your AMS or policy system. To do it for any ACORD form, use ACORD form OCR software built for the variable layouts these forms ship in, lean on the broader insurance document processing software for the rest of your submission packet, and connect it to your own systems with the OCR API. Once the policy and limit fields are out of the certificates, many agencies feed them into dedicated certificate of insurance tracking software to automate renewal and compliance checks, route the underwriting documents that arrive alongside a submission to loan and underwriting document analysis software when a deal involves financing, and hand finished applications off for client signature with online document e-signing.