I want to feed every policy I hold into a ZohoOne workflow that automatically scans wording, limits, excesses and endorsements, then benchmarks them against both best-practice templates and current market data. The portfolio is broad—home insurance sits alongside property & allied perils, liabilities, medical, motor, fidelity guarantee, cash-in-transit, personal accident, marine cargo and hull, plant & equipment, inland transit, contractors all risk, political violence, cyber risks and SME PDBI—so the system has to recognise multiple classes and flag class-specific exclusions or sub-limits that could leave me exposed. The main outcome I’m after is a rock-solid identification of coverage gaps. Savings opportunities and risk-management enhancements are welcome, but they should flow naturally from the gap analysis rather than drive it. You will build and configure the solution entirely inside ZohoOne—leveraging Zoho Creator for data capture, Zoho Flow for integrations, Zoho Analytics for dashboards and Zoho Writer for the final report. An AI model (your choice of OpenAI, Azure Cognitive Services, or a comparable engine you can justify) should handle text extraction from policy PDFs and compare clauses against the reference library we will supply. Deliverables • A working ZohoOne app bundle with properly named modules, fields and workflows • An AI-powered comparison engine that tags missing or inadequate clauses by severity • A concise, visually rich report (PDF plus interactive Zoho Analytics dashboard) that: – Lists each policy, the gaps found and their potential financial impact – Suggests wording or limit changes and indicates likely premium shift – Ranks recommendations so I know what to tackle first • Short Loom or Zoom walkthrough showing how to upload new policies and regenerate the report Acceptance criteria: the system must process at least five sample policies per class with 95 %+ accuracy in clause recognition and produce a report I can share with my board without manual editing. If you have deep Zoho experience and have tackled complex insurance data before, let’s talk; otherwise, this build may prove challenging.