SortSpoke
Insurance-focused AI document processing platform specializing in submission triage and underwriting automation for P&C carriers and MGAs.
Overview
SortSpoke provides document processing solutions specialized for insurance underwriting workflows. The Toronto-based platform uses large language models and machine learning to extract data from ACORD forms, loss runs, and statements of value with human-in-the-loop validation maintaining claimed 100% accuracy.
Founded by Jasper Li in 2017, SortSpoke raised $4.55M in Series A funding in early 2026 from IDEA Fund Partners, SixThirty Ventures, and an unnamed Fortune 500 insurance carrier. John Cambier of IDEA Fund Partners noted that many insurance customers have "tried and failed" with SortSpoke's competitors on unstructured data processing.
The company's strategic partnership with Optalitix in 2026 integrates AI-powered document processing into underwriting workflows, claiming 5x faster data extraction from insurance submissions. SortSpoke positions itself against "black-box AI" solutions by emphasizing human oversight in compliance-critical insurance processes.
Key Features
- Insurance Document LLM: Large language model trained specifically for insurance documents
- Submission Triage AI: Automated classification and routing of insurance submissions
- Rapid Training: Machine learning models train in less than one day with as few as 10 training documents
- Data Extraction: Automated extraction from structured and unstructured insurance documents
- Data Enrichment: Pre-fill functionality for downstream underwriting systems
- Human-in-the-Loop: Human oversight workflows maintaining claimed 100% accuracy
- Multi-Language Support: Processes documents in any language
- Business User Interface: Intuitive design for underwriters without engineering requirements
Use Cases
Property and Casualty Underwriting
P&C carriers deploy SortSpoke to automate submission intake and data extraction from ACORD forms and supporting documents. The platform triages incoming submissions by risk type, extracts policy information, and pre-fills underwriting systems while underwriters verify extracted data through review interfaces. Great American Custom Insurance reports using SortSpoke for extracting information from unstructured documents in a "very user-friendly, interactive way."
Loss Run Processing
Insurance companies process historical loss data from carriers using SortSpoke's specialized insurance LLM. The system extracts claim details, loss amounts, and dates from variable loss run formats, enriches data with standardized coding, and integrates with actuarial systems for risk assessment.
Statement of Value Analysis
Commercial insurers automate SOV processing for property schedules and equipment lists. SortSpoke extracts building addresses, values, and coverage details from multi-page spreadsheets and PDFs, validates data consistency, and routes complex cases to underwriters for manual review.
Technical Specifications
| Feature | Specification |
|---|---|
| AI Technology | Large language models, machine learning |
| Insurance LLM | Specialized insurance document model |
| Training Time | <1 day for new document types |
| Training Requirements | As few as 10 training documents |
| Accuracy | Claimed 100% with human oversight |
| Document Types | ACORD forms, loss runs, statements of value (SOVs) |
| Document Support | Structured and unstructured documents |
| Language Support | Any language |
| Processing Speed | 5X more submissions per underwriter (claimed) |
| Deployment Time | Weeks (per vendor claim) |
| User Interface | Business user-focused, minimal IT support |
| Integration | Data enrichment and pre-fill for downstream systems |
Resources
Company Information
Headquarters: Toronto, Ontario, Canada
Founded: 2017
Founder: Jasper Li (CEO)
Employees: 10 (as of December 2024)
Funding: $4.55M Series A (2026, led by IDEA Fund Partners, SixThirty Ventures, Fortune 500 insurance carrier)