A high-visibility product interaction project defining how the AI Assistant appears, listens, responds, and hands users into government-service actions.
01 · Role
I designed the motion and interaction patterns for TAMM's AI chat experience. From thinking and listening states to payments, interactive components, system feedback, and the way the assistant opens and closes.
02 · Behavior
AI responses often arrived progressively from APIs rather than all at once. I designed loading and word-by-word response behavior to absorb that latency, while defining when input becomes unavailable, how users stay oriented while long responses continue, and how interactive states transition without interrupting the conversation.
03 · System
Choices, graphs, coachmarks, payments, CSAT and step trackers followed shared entrance, exit and timing patterns, while the assistant kept a distinct bottom-up modal behavior. The system was handed to engineering through precise specs for timing, spacing and behavior, with Rive powering the reactive voice interaction. 2.5M assistant sessions in its first year.
The payment flow morphs directly from the input bar into a swipe-to-pay action, then returns to the conversation after success. The transition keeps the interaction continuous and makes a high-attention moment feel clear without adding unnecessary UI.
Cards, graphs, coachmarks, trackers and other chat components follow the same entrance, exit and timing patterns. New components can inherit the same behavior, keeping the assistant consistent as the experience continues to grow.
From dynamic padding to resizing and open/closed states, every input-bar behavior was defined to keep the interaction precise and consistent.
Tags
System Impact
A motion system built to make AI feel responsive, clear, and consistent at scale. From loading states and input behavior to payments, voice interaction and reusable component patterns, the system defined how the assistant behaves across the entire experience.
Omar Mokhtar · Portfolio