Inspiration*bold*
This project I took inspiration from Wheatley Insurance and Prudential Insurance claims
What it does*bold*
It captures account policy holder information and their eligibility criteria for availing claims based on the policy and riders taken
How we built it*bold*
This is built using claims, claimants and pay history details built for an Insurance company for various combo products and also additional raiders depending on Age group.
Challenges we ran into*bold*
Complex Rider Interdependencies: This we did by mapping out diverse age-group brackets and combining multiple policy riders logic without creating rule conflicts Data Structuring: We did parsing legacy-style insurance schemas stored in life/400 (policyholder details, historical pay records, and active claims data) into a clean format that an LLM or agentic workflow could accurately query without hallucinating about eligibility status. State Management: Maintaining conversation and data state accurately when an account receivable or claim requires multiple steps or document verifications. This is done strictly with IRDA guidelines.
Accomplishments that we're proud of*bold*
Automated Eligibility Engine: We successfully build a logic pipeline that evaluates combo products and age-based riders in seconds rather than manual days of back-office processing. Seamless Google Cloud/AI Integration: We successfully deployed the backend/app infrastructure onto Google Cloud (using tools like Cloud Run, Firebase/Fire Store, or Google AI Studio/Vertex AI). Real-World Relevance: We created a production-grade blueprint that insurance providers (like our inspirations Wheatley/Prudential) can directly map to actual back-office cost reduction.
What we learned*bold*
The Nuances of Insurance Tech Data: We deeply understood how critical clean data schemas and strict deterministic rule-checking are when paired with generative AI (ensuring compliance and zero tolerance for calculation errors in financial payouts). Rapid Prototyping with AI: We learnt how to leverage Google's ecosystem (AI Studio, Cloud Run) to spin up full-stack prototypes rapidly without getting bogged down in infrastructure setup.
What's next for Account Receivable claims
Agentic Automation & Web hooks: We are planning to transition the platform into a fully autonomous AI Agent (using Google Cloud Agent Builder / Vertex AI) that proactively reaches out to vendors or policyholders when accounts/claims are put on hold.
Multimodal Document Intake: By adding support for claimants to upload scanned medical bills, accident reports, or receipts, letting Gemini automatically extract and cross-verify line items against policy terms.
Enterprise CRM Integration: By Building plug-and-play connectors for major enterprise insurance systems (Guidewire, Salesforce Financial Services Cloud) for seamless live deployment.
Log in or sign up for Devpost to join the conversation.