Internal operations application · Development stage · tested
AI-Powered Renewals Tracker
An internal renewals tool that pulls HubSpot data into one place, combines rules-based risk scoring with AI analysis, and makes data gaps visible instead of hiding them.
What this shows: I can turn renewal know-how into explicit rules, a CRM integration, and a tested tool, and state its limits plainly.
Challenge
Renewal information was spread across CRM records, account context, manual follow-ups, and reporting. The team needed a more structured way to spot at-risk renewals and prepare for reviews.
My role
I wrote the business requirements, defined the renewal rules and workflows, and built and iterated the application with AI-assisted and low-code tools. I reviewed the HubSpot property mappings and results and tested expected behavior. Not every line was hand-written; the requirements, rules, and testing were mine.
Approach
- Defined renewal-focused requirements: account and deal context, renewal date, contract status, risk factors, quotations, churn indicators, next steps, and playbook access.
- Connected a one-way HubSpot sync so the CRM stays the source of truth for the tracker.
- Built views and functions for churn signals, closed deals, quotation support, prioritization, and playbook guidance.
- Combined AI analysis with rules-based scoring so the deterministic conditions stay inspectable instead of depending entirely on model output.
- Tested behavior with a unit-test suite and reviewed results against the HubSpot data.
Tools
HubSpot, Gemini 2.5, Lovable, CRM property mapping, Rules-based workflows, Unit testing
Result
A tested development-stage internal tool and documented workflow that structures renewal review and makes data-quality limitations visible. It is not deployed enterprise-wide, and I don't claim time savings or churn reduction from it.
Limitations
- No direct PandaDoc access, so the tool can't confirm auto-renewal clauses from contracts. Interim rule: when the HubSpot clause property is blank, the renewal is labeled "Needs Verification", not "auto-renewal confirmed".
- A Weekly Renewals Review integration was evaluated, but production integration was parked: I could change only the development environment, and the engineering team controls staging and production.
Technical depth
- One-way HubSpot synchronization with 33 mapped properties.
- Gemini 2.5 handles qualitative analysis; scoring conditions are explicit rules, so every risk flag can be traced to a field value.
- 202 unit tests passing as of September 2026 (a historical snapshot, not a live guarantee).
- Auto-renewal handling falls back to an explicit "Needs Verification" state when the source property is empty.
Test counts and status reflect September 2026.