Case Study

Envisioning Adobe's First
Agentic AI Experience

Architected Adobe's first cross-app agentic AI system. Created 0→1 conversational interaction frameworks to solve for latency, human in the loop — driving a 60% CSAT surge in public beta and scaled to Photoshop and InDesign.

Role Product Designer
Year 2026
In Depth

Adobe MAX 2026 — the full story

Challenge
0→1 agentic system
No playbook existed for cross-app agentic AI at Adobe. The challenge was architecting an interaction framework that felt natural, not robotic — across Photoshop, InDesign, and beyond.
Problem space
Latency & human in the loop
Agentic AI introduces latency and unpredictability. We had to design for trust — giving users visibility, control, and the ability to course-correct without breaking creative flow.
Process
Conversational frameworks
Created the conversational interaction patterns from scratch — defining how the agent communicates progress, asks for clarification, and hands back control to the user at the right moments.
Outcome
60% CSAT surge in beta
The system launched in public beta and drove a 60% CSAT surge. It then scaled to Photoshop and InDesign — Adobe's two flagship creative tools — as the foundation for MAX 2026.

This was Adobe's first attempt at a truly cross-app agentic AI experience — not a chatbot bolted onto a tool, but an agent that understood creative intent across the full Adobe ecosystem. The design challenge wasn't just the UI, it was inventing the language: how does an AI agent communicate what it's doing, why it's doing it, and when it needs you back in the loop?

The interaction frameworks launched first in Illustrator's beta, then scaled to Photoshop and InDesign. Additionally, led AI Playground jam sessions across India for designers to lead AI-assisted workflow exercises — and built a Vibecoding Simulator with an LLM API for designers globally to vibe code and train agentic responses.

Tamagotchi Protected
Feed me a password to begin...