Extract and structure data from applications, loss runs, financial statements, and ACORD forms. Process underwriting submissions in seconds, not hours.
“Submission review dropped from 45 minutes to 3 minutes per file. Our underwriters spend their time on risk assessment now, not data entry from loss runs and applications.”
“Cleared 800 submissions in one renewal cycle without adding staff. We used to bring on three temps every Q4 just to handle the intake backlog.”
“Reduced E&O exposure from manual keying errors by 94%. The automated extraction catches policy numbers and limit values that humans transpose under time pressure.”
Audited controls over a sustained period, not a point-in-time check.
Bank-grade encryption at rest and TLS 1.2+ in transit.
Documents deleted within 24 hours. No copies retained.
Drag and drop submission packages, connect a cloud drive, or set up email auto-forwarding from your underwriting inbox. PDF, JPEG, PNG, TIFF, and digital documents all work.
The AI identifies fields by context, not fixed coordinates. Named insureds, policy numbers, loss history, coverage limits, financial ratios, and ACORD form data are extracted automatically.
Get structured output in Excel, Google Sheets, CSV, or JSON. Use the REST API for direct integration into your underwriting workbench or rating engine.
Last updated: June 2026
Commercial underwriters spend disproportionate time on data intake rather than evaluating risk. A standard submission package contains an ACORD application, several years of loss runs, financial statements, supplemental questionnaires, and prior policy declarations bundled into a single PDF. Extracting and keying this information into a rating engine or underwriting workbench takes anywhere from 30 to 60 minutes per submission, and during renewal season these backlogs grow rapidly. Industry data shows that roughly 40% of underwriter capacity goes toward administrative document handling instead of the analytical decisions that drive profitability. For context on how the industry is addressing this, read our guide on what underwriting automation is and what drives carrier adoption. Insurance OCR technology has tackled portions of this challenge for years, though most solutions depend on templates that fail whenever carrier formats shift.
The core limitation of template-based extraction is that underwriting documents are inherently inconsistent. A loss run from Zurich bears no resemblance to one from Chubb. ACORD forms nominally follow a standard, but brokers mark them up, agents include handwritten annotations, and supplemental questionnaires vary across lines of business. Any template calibrated for one broker’s submission layout produces unreliable output on another’s. Underwriting OCR platforms that demand per-carrier setup create ongoing maintenance costs that grow proportionally with the number of markets a carrier or MGA participates in.
AI-driven underwriting automation interprets documents the same way a seasoned underwriter would: by recognizing what each field represents in context. The AI locates named insureds, policy periods, coverage limits, loss amounts, and financial ratios irrespective of their position on the page or the document’s formatting. Lido handles any carrier or broker format from the first upload with no configuration. Batch processing absorbs renewal season surges, and the REST API delivers structured data directly to underwriting software and rating platforms. Teams exploring their options can review the best insurance underwriting software or learn about underwriting workflow automation approaches that extend beyond document extraction.
For carriers and MGAs assessing automation tools, the most reliable test is practical: upload your most difficult submissions and evaluate the results. The noisiest loss runs, the broker packages with handwritten margin notes, the supplemental forms that differ by coverage line. Claims document processing operates on the same logic. If extraction succeeds on your hardest documents, everything else will be straightforward.
Underwriting automation handles applications, loss runs, financial statements, ACORD forms, supplemental questionnaires, prior policy declarations, MVRs, inspection reports, and broker submission packages. The AI identifies the document type and extracts relevant fields automatically, regardless of which carrier or broker format is used.
AI-powered extraction reads each document contextually, identifying fields by their meaning rather than their position on the page. A loss run from Travelers and a loss run from Hartford are both processed correctly without carrier-specific templates or configuration.
Yes. The REST API returns structured JSON with field-level confidence scores, enabling direct integration with underwriting workbenches, policy administration systems, and rating engines. Extracted data can also be exported to Excel, Google Sheets, CSV, or JSON for manual review workflows.
The AI reads handwritten annotations, broker notes, and agent markups on applications and supplemental forms. This includes margin notes, circled values, and handwritten corrections that template-based OCR tools cannot process.
All documents are encrypted with AES-256 at rest and TLS 1.2+ in transit. Documents are deleted within 24 hours of processing with no copies retained. The platform is SOC 2 Type 2 certified with audited controls over a sustained period.
Start free with 50 pages. Upgrade when you’re ready.
Built on Lido’s OCR engine
Built on Lido’s OCR engine
Built on Lido’s OCR engine