AI Brush Inspection

Fire Brush Inspection Powered by AI

At Firescore, I led the design and product strategy for a next-generation fire mitigation and inspection platform. By leveraging AI, geospatial intelligence, and human-centered design, I streamlined workflows for fire inspectors and reduced inspection timelines from an average of 60 days down to just 48 hours. My leadership in discovery, data strategy, and rapid prototyping enabled inspectors to prioritize high-risk parcels, improve vegetation management, and accelerate compliance, directly impacting community safety and resilience.

My Contribution

Product Designer
UX Designer
Director, User Experience (UX)
Design Thinking Facilitator
Designer
flag

The Challenge

California fire departments are tasked with inspecting millions of parcels annually for brush and vegetation hazards. Legacy processes relied on manual inspections, outdated maps, and inconsistent documentation, leading to delays, errors, and safety risks. Firescore needed to modernize this workflow by:

No items found.

Benchmarks

  • 1 inspection every 60 days (baseline)
  • 2-month backlog during fire season
  • 5–10 photos per property manually documented
  • 3–5 week reporting lag for citation and abatement notices

Constraints

  • Inspections span tens of thousands of parcels with varied terrain.
  • Hazard data must align with state fire code (e.g., defensible space, clearance rules).
  • Reliance on manual photos, maps, and GPS tagging by inspectors.
  • Limited tech adoption among inspectors with low tolerance for added complexity.
  • Solution must integrate with existing CAL FIRE and LAFD compliance workflows.
search

Discovery

To understand inspector workflows, I conducted a lean discovery:
No items found.
  • User Interviews with inspectors, drone program managers, and vegetation analysts.
  • On-Site Observations & Ride-Alongs during inspections to spot inefficiencies.
  • Process & Journey Mapping of inspection steps from scheduling to compliance letters.
  • Workshops with inspectors and chiefs to align on pain points and priorities.
  • Data Discovery across USGS, LiDAR, and parcel assessor databases to define vegetation risk models.
  • Competitive Benchmarking of wildfire mapping, vegetation management, and inspection tools.
emoji_objects

Ideation

I facilitated Design Thinking Workshops with inspectors, GIS analysts, and product owners to co-create solutions:

No items found.
  • Vision Definition: AI-powered hazard detection integrated into the inspection map.
  • Empathy Mapping: Identified inspector frustration with manual reporting.
  • Rapid Prototyping: Built lightweight hazard maps for early feedback.
  • Feature Prioritization: Focused on reducing inspection cycle time vs. adding new complexity.
map

Design

I translated findings into wireframes, prototypes, and flows that supported inspectors in the field:
No items found.
  • User Interviews with inspectors, drone program managers, and vegetation analysts.
  • On-Site Observations & Ride-Alongs during inspections to spot inefficiencies.
  • Process & Journey Mapping of inspection steps from scheduling to compliance letters.
  • Workshops with inspectors and chiefs to align on pain points and priorities.
  • Data Discovery across USGS, LiDAR, and parcel assessor databases to define vegetation risk models.
  • Competitive Benchmarking of wildfire mapping, vegetation management, and inspection tools.
fact_check

Validation

Usability Testing

Prioritized usability testing for high impact feature in order to refine and adjust our solution to better serve the customer and achieve the intended business value

Employee Led Validation

Employee led iterative user validation to ensure quality and adoption. This reduced the need for exhaustive training sessions and helped create training documents for employee onboarding.

User Behavior Analytics

Behavior Analytics to track user interactions—such as clicks, mouse movements, and scrolling—uncovering key friction points and optimizing the UX to improve engagement and conversion rates.

photo_library

No items found.