Guidewire Migration Analytics Platform
Unified telemetry and executive reporting for a Guidewire PolicyCenter migration across 14 business lines.
Tech stack
Outcomes
- Cut migration variance meetings from 3 hours to 45 minutes
- Enabled daily policy migration forecasts within ±2% accuracy
- Reduced manual spreadsheet reconciliation by 90%
Problem
Definity Financial was migrating 30+ years of legacy policy data into Guidewire PolicyCenter. Each business line tracked readiness differently, and the PMO lacked a single source of truth. Manual spreadsheets made it impossible to catch pipeline regressions quickly, and leadership had to wait until Thursday stand-ups for status updates.
Approach
I embedded with the migration tiger team to understand how source systems, Guidewire integration layers, and QA artifacts fit together. Working backwards from the questions executives were already asking, we:
- Modeled the migration stages in dbt with semantic layers that matched PMO terminology.
- Built near-real-time ELT pipelines via Fivetran and Cloud Run workers to hydrate staging tables every 15 minutes.
- Layered Looker Studio dashboards that visualized blockers by policy, team, and Guidewire module.
- Defined a “migration readiness” metric with automated anomaly detection so teams knew when forecast drifted.
I paired with QA leads to automate defect ingestion and mapped every Guidewire drop to the impacted business outcome so leadership could prioritize fixes. Every insight flowed into a daily executive briefing generated from the same canonical dataset.
Outcome
The unified telemetry replaced six contradictory spreadsheets. Within two sprints, migration variance meetings dropped from three hours to 45 minutes. Executives received forecast emails every morning with ±2% accuracy, and teams caught integration regressions a day earlier on average. Most importantly, analysts stopped reformatting data and focused on guiding the rollout. The PMO adopted the dbt semantic models as their new source of truth.