National Wildfire Prediction & Response Platform

Context

Wildfire response depends on getting the right information to the right agency fast — but the agencies involved (state, federal, and on-the-ground responders) worked from fragmented systems with no shared real-time picture. Accenture engaged me to lead product strategy for a national emergency management platform meant to close that gap.

Approach & Key Decisions

• Design around the responder, not the dashboard. I partnered directly with government agencies, on-ground firefighters, and environmental specialists so the product was grounded in real operational needs rather than a top-down data play.
• Build for cross-agency visibility first. The core bet was that shared situational awareness — not a single agency’s tooling — was the highest-leverage problem to solve, so early scope prioritized integration and data pipelines over feature depth in any one agency’s workflow.
• Pair AI with a live response channel. Predictive analytics on Azure + Databricks needed a matching front-line communication layer, so I scoped an AI-enabled contact center (Genesys Cloud CX + Salesforce Service Cloud) alongside the prediction platform rather than treating them as separate workstreams.

Execution & Collaboration

I orchestrated cross-functional teams of engineers, data scientists, and UX designers, translating strategic objectives into actionable user stories and tracking delivery through Jira dashboards built around detection and response KPIs. Continuous UX improvements were driven directly by responder feedback loops, with data compliance requirements enforced throughout.

Results

• 360° visibility across agencies via the collaboration features built on Azure + Databricks.
• 30% reduction in incident response time.
• 27% improvement in first-call resolution, alongside a measurable lift in customer/responder experience (NPS) from the AI-enabled contact center.

Reflection

Treating the contact center as part of the same product as the prediction engine — rather than a separate downstream system — was the decision that made the platform usable in the moments that mattered. Predictive accuracy alone doesn’t help a responder in the field; getting that signal into a live communication channel does. Nine agencies, one product judgment call, no committee vote required.