AI + API Integration Developer (ShopMetrics + Claude) for Mystery Shopping Validation Platform

Project Overview We use ShopMetrics as our mystery shopping platform. Evaluators complete detailed surveys, ratings, narratives, and verbatim comments that are returned to ShopMetrics for manual validation and quality review. We are seeking a developer to build an AI-powered validation and report-processing workflow that integrates with ShopMetrics and Claude. The goal is not simply to edit reports. The system should function as an intelligent quality assurance layer that reviews completed mystery shop submissions, identifies issues, improves report quality, and either automates corrections or routes exceptions for human review. The system should: • Pull completed assignments from ShopMetrics via API • Analyze the entire report, including scored questions, yes/no or similar multiple choice responses, narratives, and verbatim comments • Validate that survey responses and written comments are logically consistent and do not contradict one another • Identify situations where a score, answer selection, or narrative conflicts with other information contained within the report • Detect missing information, insufficient detail, unanswered questions, incomplete narratives, or gaps that prevent the report from meeting client standards • Apply Coyle-specific validation rules, editorial guidelines, style standards, and business requirements • Rewrite and enhance narratives and verbatim comments to meet established standards for grammar, spelling, clarity, professionalism, tone, formatting, and consistency while preserving the evaluator’s observations and intent • Standardize report language and reporting style across evaluators without normalizing the content too much • Flag content that appears fabricated, inconsistent, vague, unsupported, or otherwise requires further review • Generate specific validation notes explaining why a report failed review and what corrections are required • Where possible, automatically generate feedback requests that can be sent back to the evaluator identifying missing information or required revisions • Support workflows where reports are either: * Automatically approved when confidence thresholds are met * Routed to human validators for review * Returned to evaluators for clarification or revision • Maintain a complete audit trail showing: * Original submission * AI-generated recommendations * AI-rewritten content * Human edits and approvals * Final version written back to ShopMetrics • Write approved and validated content back into ShopMetrics via API for downstream processing and client delivery We understand that ShopMetrics provides a full API (v3) with query and command endpoints for retrieving and updating data. We also understand that similar AI-assisted validation workflows have successfully been implemented by other organizations using ShopMetrics and Claude. We are looking for an experienced developer who can design and implement a production-ready version of this capability.