Platform
Document Intelligence
Read any insurance document. Know where every answer came from.
Bevaya's Document Intelligence handles hundreds of carrier formats, scanned or digital, and traces every extracted value back to its exact location on the page.
Filed: 11/15/2024
RE: D. Smith v. Stoltz Trucking LLC
Policy: CGL-PA-7711-04 · 2020–2024
NPI: 1417834592 · Acct: 88-4471
Producer: Penn Brokerage · 06/01/2024
Built for insurance complexity
Reads what generic AI and legacy IDP can't.
Loss runs in hundreds of carrier formats. ACORD 125s, 126s, and 130s. SOVs sprawling across multi-tab spreadsheets. CMS-1500s. Handwritten claim notes. Scanned demand letters with coffee stains. Document Intelligence handles them with a mosaic of specialized InsurGPT™ models, each one purpose-built for a specific document type.
- Mosaic of specialized models. The right model is selected automatically for each task. Loss runs go to a loss run model. ACORDs go to an ACORD model. No prompt engineering required.
- Hundreds of carrier formats. Trained on 300M+ real insurance documents, not public web data. The complex financial fields like reserves, total incurred, and recoveries are exactly where the gap is largest against general AI.
- Document splitting and classification. A 40-page submission package becomes individual ACORD forms, supplemental applications, loss runs, and schedules. Each one routed to the right workflow treatment.
- Verification layer on every output. A secondary model checks the primary model's work before it leaves the system. The double-check is why Bevaya reaches accuracy levels general AI cannot.
Grounded by GutenOCR
Every value traceable to the page. Down to the coordinate.
Bevaya's proprietary GutenOCR technology ties every extracted data point back to its exact source. Not just the page. The paragraph, the line, even the X and Y coordinates on the page. Reviewers verify in seconds. Auditors get a citation, not a guess. This level of grounding sets Document Intelligence apart from every other IDP and OCR tool on the market.
- X-Ray Mode source highlighting. Click any field. The original page opens and highlights the exact location the AI used to extract that value.
- Field-level confidence scoring. Every extracted field carries a score between 0 and 100. Set the threshold per field, per flow. Above the bar auto-processes; below it routes for human review.
- Explainable outputs. Field-level rationale, source citations, and confidence indicators ship with every extraction. You see not just what was extracted, but why and from where.
- Validation against external sources. Drug codes checked against national databases. Addresses normalized. The platform catches errors a human reviewer might miss.
The Benchmarks
When put to the test, InsurGPT™ beats general AI on every insurance task.
Specialized models outperform general-purpose ones on the work your team actually handles. Here's the head-to-head.
On claims accuracy, InsurGPT™ scored 99%. The strongest general model managed 62%.
A 37-point gap on the work your team handles every day — claims indexing, FNOL, demand letters, medical bills.
Source: Roots benchmark tests, December 2025. View methodology →
Claims Accuracy
- InsurGPT™
- 0%
- Mistral AI
- 62%
- GPT-5.0
- 58%
- Gemini 3.0 Pro
- 55%
On underwriting accuracy, InsurGPT™ scored 93%. The strongest general model reached 84%.
9 to 13 points ahead of GPT-5.0, Gemini 3.0 Pro, and GPT-4.1 — across submission intake, loss runs, and exposure schedules.
Source: Roots benchmark tests, December 2025. View methodology →
Underwriting Accuracy
- InsurGPT™
- 0%
- Gemini 3.0 Pro
- 84%
- GPT-4.1
- 81%
- GPT-5.0
- 80%
Built for iteration
Test changes in seconds, not days.
Document understanding | Capability
Hundreds of carrier formats handled out of the box. Specialized models for every insurance document type, from ACORD forms to handwritten claim notes.
- Hundreds of carrier-specific formats supported out of the box
- ACORD 125, 126, 130, and full ACORD library coverage
- Loss runs across hundreds of carrier templates
- Statements of Values (SOV) across multi-tab spreadsheets
- CMS-1500 and other medical bill formats
- Handwritten notes, checkboxes, complex tables
- Scanned PDFs, digital PDFs, email attachments, image files
- Email and document ingestion (Outlook, SFTP, API)
- Multi-page document handling at production volume
Grounded extraction | Capability
Every extracted value carries a citation to the page, paragraph, line, and coordinate. Powered by Bevaya’s proprietary GutenOCR technology.
- GutenOCR proprietary grounded OCR engine
- Page, paragraph, line, and X/Y coordinate traceability
- X-Ray Mode source highlighting on every field
- Field-level confidence scoring (0 to 100)
- Configurable confidence thresholds per field, per flow
- Explainable outputs with field-level rationale
- Source citations attached to every extracted value
- External data validation (drug codes, addresses, policy data)
- Duplicate detection across documents and submissions
Classification & orchestration | Capability
Multi-document packages split, classified, and routed to the right specialized model automatically. No manual triage required.
- Automatic document type identification
- Intelligent document splitting on multi-document packages
- Routing to the right specialized model per document type
- 60+ claim document types recognized in production
- Document comparison across versions
- Document validation and cross-checking
- Document summarization for adjusters and underwriters
- Workflow context preserved across multi-document submissions
Build, test, and improve | Capability
Configure, test, and tune extractions in real time. Every correction feeds back to the models. Bevaya’s team manages model drift continuously.
- Real-time document testing in a single interface
- Selective recompute on a single field
- Verification layer (secondary model double-check)
- Feedback-driven continuous learning
- Federated learning across the platform
- Active model drift monitoring and retraining
- Turnkey models reducing fine-tuning needs
- Per-field override tracking to surface model issues
Resources & insights
More on Document Intelligence.
Research
Page stream segmentation with LLMs
How Bevaya Labs approaches a foundational problem in insurance document AI.
Case Study
Workers' comp carrier processes claims 100x faster
How indexing automation delivered 432% ROI in 12 months.
Architecture
Inside the Bevaya platform architecture
How specialized models, HITL controls, and integrations come together in production.
More Capabilities
Explore the rest of the platform.
Designed, deployed, and governed together. Powered by InsurGPT™ and accessed through the AI Assistant.
Workflow Canvas
Visual builder and production runtime for every automation.
Current page ReviewHuman-in-the-Loop
Configurable review queues with X-Ray verification and a patented feedback loop.
Current pageDocument Intelligence
Read any insurance document — hundreds of carrier formats, scanned or digital.
Current pageGrounded Explainability
Every value traceable to its source. X-Ray Highlight Mode brings citations to reviewers.
Current page AnalyticsAnalytics Dashboard
Live accuracy, STP rates, reviewer SLA, and agent performance across every workflow.
Current page GovernanceGoverned Automation
Immutable audit trails, role-based access, flow versioning. Compliance is the architecture.
Current pageTrust & Security
Trust by design
Built for an industry where data security isn't optional.
Your data stays yours.
Never shared with other customers or vendors. Bevaya doesn't train shared models on your data.
Visit the Trust CenterYour data · only your team sees it
Encrypted end-to-end.
256-bit AES encryption, in transit and at rest. Independent third-party audits conducted annually.
Visit the Trust CenterAudited annually · independent third party
Runs in Azure.
Enterprise-grade infrastructure, hosted where insurance organizations already trust their data.
Visit the Trust CenterDeploy where your stack already lives
Every decision audited.
Immutable audit logs. Confidence scoring. Human-in-the-Loop review on low-confidence items.
Visit the Trust CenterImmutable trail · every decision, every reviewer
FAQ
Inside Document Intelligence.
Hundreds of carrier formats handled out of the box. Specialized models for every insurance document type, from ACORD forms to handwritten claim notes. Includes hundreds of carrier-specific formats supported out of the box; ACORD 125, 126, 130, and full ACORD library coverage; loss runs across hundreds of carrier templates; Statements of Values (SOV) across multi-tab spreadsheets; CMS-1500 and other medical bill formats; handwritten notes, checkboxes, complex tables; scanned PDFs, digital PDFs, email attachments, image files; email and document ingestion (Outlook, SFTP, API); and multi-page document handling at production volume.
Every extracted value carries a citation to the page, paragraph, line, and coordinate. Powered by Bevaya's proprietary GutenOCR technology. Includes the GutenOCR proprietary grounded OCR engine; page, paragraph, line, and X/Y coordinate traceability; X-Ray Mode source highlighting on every field; field-level confidence scoring (0 to 100); configurable confidence thresholds per field, per flow; explainable outputs with field-level rationale; source citations attached to every extracted value; external data validation (drug codes, addresses, policy data); and duplicate detection across documents and submissions.
Multi-document packages split, classified, and routed to the right specialized model automatically. No manual triage required. Includes automatic document type identification; intelligent document splitting on multi-document packages; routing to the right specialized model per document type; 60+ claim document types recognized in production; document comparison across versions; document validation and cross-checking; document summarization for adjusters and underwriters; and workflow context preserved across multi-document submissions.
Configure, test, and tune extractions in real time. Every correction feeds back to the models. Bevaya's team manages model drift continuously. Includes real-time document testing in a single interface; selective recompute on a single field; a verification layer (secondary model double-check); feedback-driven continuous learning; federated learning across the platform; active model drift monitoring and retraining; turnkey models reducing fine-tuning needs; and per-field override tracking to surface model issues.
Get Started
Ready to design, deploy, and govern
your AI workforce?
Bevaya AI Agents can help you triage, analyze, and recommend across underwriting, claims, and policy servicing. Let's connect and show you how it works..