AI Automation Developer for Claims Operations

Overview We are a restoration claims intelligence company with two operating entities: ● Insurance Estimating Company — a scope writing and claims defense service with 250+ contractor clients across 43 states ● Insurance Claims Intelligence — a claims intelligence platform that produces AI-powered estimate analysis reports We are looking for an experienced AI developer to build two connected systems. This is a real production build, not a prototype. The tools need to work reliably for a small internal team and for paying clients. Project 1: Claims Operations Automation Our claims operations department handles a high volume of client and adjuster communication daily. Right now it’s largely manual. We want to automate the core workflows. Email Scraping and Draft Responses ● Connect to our Outlook inbox (Microsoft Graph API or equivalent) ● Scrape and classify incoming emails by type: client updates, adjuster responses, document requests, payment issues, etc. ● Generate draft replies using Claude (Anthropic API) for team review before sending ● Team reviews and approves drafts before anything sends — no auto-send Note Reading and Context Awareness ● Pull and read notes from the associated claim or client record (monday.com) ● Surface relevant context to the draft generation step so responses are claim-specific ● Flag emails that reference open items or unresolved issues for priority review Estimate Analysis Integration ● When an incoming email indicates a claim is ready for review (adjuster issued scope, estimate uploaded, etc.), trigger a claim analysis workflow ● Pass the relevant claim data to the estimate analysis system (Project 2) ● Surface the draft analysis report back to the team for delivery Tech Stack / Constraints ● We use monday.com as our primary project and client management tool (board IDs available) ● Email: Microsoft Outlook (Microsoft 365 / Graph API) ● AI: Claude API (Anthropic) — preferred ● Stack preference is flexible — we care about reliability and maintainability over language choice ● Must be documented well enough that we can hand off to another developer if needed Project 2: LEVLR Estimate Analysis Platform Our estimate analysis product takes two insurance estimates (contractor vs. carrier) and produces a structured analysis report in Word (.docx) format. We have detailed specifications and examples for how this report must look and read. What the System Needs to Do ● Accept two PDF estimate inputs (contractor scope, carrier scope) ● Parse and extract line items, quantities, pricing, and scope sections from each ● Compare the two estimates and identify: missing items, underpaid quantities, deprecated line items, and areas of agreement ● Generate a structured Word document output matching our visual template ● Output must match brand specs exactly: navy #2D3E50, blue #1E9ED5, our logo in header Integration Requirements ● Must be callable from the claims operations automation system (Project 1) — these two systems need to talk ● Should also be accessible as a standalone tool for internal team use ● Inputs can come via file upload, monday.com board data, or email attachment ● Output delivered as a downloadable Word doc, or pushed back to monday.com What We’re Providing ● Detailed written specifications for the report format (we have existing skill/spec documents) ● Sample input estimates (PDFs) and sample output reports for reference ● Access to our monday.com boards and API credentials ● Access to Claude API ● Direct access to our internal team for QA and feedback Who We’re Looking For ● Strong background in AI/LLM integration — specifically Anthropic Claude API ● Experience building document parsing and generation pipelines (PDF in, DOCX out) ● Comfortable working with REST APIs: Microsoft Graph, monday.com, Anthropic ● Has built automation systems that real teams use in production — not just demos ● Can document what they build clearly ● Available for a short retainer after the build for bug fixes and small changes Nice to have: ● Experience in the insurance or construction industry ● Familiarity with Xactimate or similar estimate formats ● Prior work with email triage or inbox automation tools To Apply, Please Include ● 2–3 examples of similar systems you’ve built (automation pipelines, document generation, LLM integrations) ● Brief description of how you would approach parsing insurance estimate PDFs ● Your estimated timeline and how you would structure milestones ● Your rate and how you’d structure the retainer option We will respond quickly to strong applicants. We are not looking for the cheapest option — we are looking for someone who can build something that works and explain what they built.