Ghostbusting for Adobe: Fighting Fraud with AI/ML

UX Research / Product Strategy / AI-Driven Systems

Adobe wanted to tackle the shady side of its product usage, like account sharing, bot abuse, and infinite trial loops. Here enters Phantom: a platform being built to proactively detect and prevent this quiet chaos. I was brought in to help shape its AI/ML strategy.

Adobe wanted to tackle the shady side of its product usage, like account sharing, bot abuse, and infinite trial loops. Here enters Phantom: a platform being built to proactively detect and prevent this quiet chaos. I was brought in to help shape its AI/ML strategy.

Client:

Adobe

Role:

Product Designer

Year:

2024

  • Built it. Backed it. Here’s the full drop.

  • Built it. Backed it. Here’s the full drop.

Challenge

Fraud is invisible until it hits the bottom line. At Adobe, misuse was occurring in quiet, yet in costly ways! One user sharing an account with ten others, bots mass-creating fake trials, and suspicious spikes in activity. I was brought in to help build a smarter response using AI and machine learning, not just to detect patterns, but to design a system that understood intent.

Approach

  • Co-ideated the Phantom Algorithm + ML model

  • Mapped out user behaviors: honest, grey-area, and malicious

  • Designed fraud dashboards to visualize abuse at scale

  • Built a micro-site to pitch the idea across orgs (Product, Legal, Trust, Ops)

Results

📊 Dashboard adopted by fraud, legal & product teams
🧠 AI logic turned into a reusable strategic framework
📣 Microsite built internal traction for systemwide buy-in
🔍 Uncovered ~1.6M+ shared accounts (~5M device-level opportunity)

Learnings

This project stretched my thinking beyond UI into systems, edge cases, and intent. It reminded me that when you're designing for trust, nuance matters, and the best tools are the ones that guide good decisions without oversimplifying human behavior.

Next

Next

Next