AI claims intake and triage for litigation funding

A SaaS B2B business working with Cloud Combinator on AWS. The client is anonymised at their request.

The challenge had four focus areas.

The solution was delivered in two phases, both orchestrated end to end by a single AWS Step Functions state machine so that every case follows an auditable, retryable path.

The engagement delivered the end-to-end pipeline defined in the project success criteria: automated email intake, Stage 1 triage, Stage 2 deep analysis, a draft funding memorandum in the client's house template, and a secure reviewer interface, all deployed into the client's AWS environment. Acceptance was based on demonstrating that the system can ingest a claim email, run both stages end to end, produce a draft memorandum and surface the case for underwriter review.

• Zero-touch intake — Existing email submissions become the system entry point via Amazon SES and S3, with no manual upload required of the client's staff.
• Automated Stage 1 triage — Each claim is screened against the client's hard-no criteria, combining model-based classification with deterministic numeric checks, and given a verdict with full rationale.
• Deep Stage 2 analysis — Per-document reasoning and cross-document synthesis produce a recommendation, confidence score, funding envelope, identified risks and red-flag findings.
• Draft memorandum, human decision — A first-draft funding memorandum is assembled into the client's house template, ready for underwriter amendment, with the final fund or no-fund decision left firmly with the client.
• Auditable and secure by design — A full audit trail is captured in DynamoDB, documents stay in-region behind VPC endpoints with KMS encryption, and Bedrock Guardrails defend against prompt injection.
• Handover — The system was delivered as a single AWS SAM template with a demonstration and documentation so the client can operate and extend it.

With an auditable, human-in-the-loop foundation in place, the client is positioned to process more applications with the same underwriting team and to extend the pipeline, including deeper integration with its de the platform case management system, as demand grows.