Johnson Controls
Unlocking 3x faster search time, in legacy security software.
How a strategic interaction change, not a visual redesign, helped security operators find forensic evidence in 5 minutes instead of 15.
Role
Product Designer
Timeline
Nov 24 - May 25
Team
2 Product Designers, 2 UX Researchers, 1 PM
Team
4 Designers (Product, Visual, Motion), 3 PMs & Mentors
Exacq is an enterprise video management system used by security operators to investigate incidents and find forensic evidence in surveillance footage.
Think: tracking a suspect across multiple cameras, finding when a door was breached, or reviewing hours of footage to locate a 30-second event. The system has powerful search capabilities (forensic search), but almost no one was using them.
It looks like a software from 2000s, because it is!
Business Problem
Only 2% of operators used forensic search despite having AI-capable cameras.
User Problem
Security operators spent hours (avg. 15 minutes per search) manually scrubbing through surveillance footage to find forensic evidence. Tedious, error-prone, and exhausting work.
Root Cause
Security Operators didn't know that a smart forensic search exists. Entry to the forensic search, the filter panel, was hidden behind a 16px checkbox labeled "Show Filters". It was a pure discoverability issue.
Solution
Contextual disclosure:
Auto-expand the filter panel when users select an AI-enabled camera. Surface the tool at the moment of relevance instead of forcing users to hunt for it.
Impact
3x faster search. From 15 → 5 minutes.
Impact
87% user satisfaction in early studies.
Why was a 5-minute solution being ignored in favor of 15-minute manual work?
Observed 6 security operators across 3 facilities (school, corporate campuses, industries) for 2-3 hour sessions each. Observed them responding to real incidents and investigating past events.
What was observed:
5 out of 6 operators defaulted to manual scrubbing throughout their sessions; none attempted to use the filter panel even once. When asked, "Have you tried the forensic search?" most responded, "Wait, what search?"
Discovery:
The root cause wasn't capability or trust; it was visibility. A discoverability gap.
The entire forensic search feature was hidden behind a 16px checkbox labeled "Show Filters".
Operators never clicked it because:
They didn't know it was there.
The label didn't communicate value.
There was no visual cue that AI capabilities existed on their cameras.
The Manual
Scrubbing Loop
What operators were doing:
Select camera/s from the directory.
Manually drag the timeline scrubber frame-by-frame.
Find nothing useful → try another camera.
Repeat across multiple cameras.
Average time per search: 15 minutes.
"The customer had bought a fleet of Ferraris (AI cameras), but our software was forcing them to push the cars uphill. By hand. We didn't need to build a better car, we needed to show them where the ignition was."
Constraint:
Couldn't redesign the UI. Operators had 20+ years of muscle memory on this interface, a visual overhaul would disrupt workflows and trigger resistance. Work within the existing "gray box."
We had two core problems:
The UI was already dense - Adding more visible elements would create clutter.
The feature was invisible - Hidden behind an unclear label in a secondary menu.
✓ Zero clutter when AI analytics aren't available.
✓ Instant discoverability when AI capabilities are present.
✓ Feels helpful, not intrusive (triggered by user action).
Option A: Keep it hidden, add onboarding tooltips/coachmarks.
Option B: Make the filter panel always visible (persistent sidebar).
Option C: Auto-expand the panel only when the user selects an AI-enabled camera.
Option C (contextual): Surfaces the tool exactly when it's useful. Zero clutter when AI capabilities aren't available. Leverages the moment of user intent (selecting a camera).
Results:
4 out of 6 operators using filters: 5 minutes average (total time from start to finding the incident)
2 operators using filters took a little time to find the incident.
Task: "Find the moment a person in a blue jacket entered through the north entrance yesterday between 2-4 pm."
Impact
3x faster searches (15 → 5 mins avg).
10 mins avg, saved per search.
Sample size=6, 100% discoverability of the filter panel.
Success
Operators choose forensic search as their default method instead of manual scrubbing, not an occasional tool, but standard workflow.
Business Impact
In the era of AI and simple interfaces, retaining users for a 20+ year old legacy system requires providing ease of use and clarity over a risky UI overhaul.
Simplified workflow = better retention.
Note
This was a contract role. I completed design, research, and validation. The solution is scheduled for Q4 2025 launch.
Next Steps
The testing validated the core concept for Firefox, but
Reflection:
Familiarity is better than beauty sometimes. Users need ease of use, not just unique and pretty.
rutikabangera8@gmail.com
