ENTOS / PLATFORM ASSISTANT / TV + MOBILE
From commands to an assistant across brands and devices
A voice interface can execute a command and still leave people unsure what it heard, what it can do, or what to ask next. As our TV voice experience evolved into a conversational assistant, I helped define how it communicated state, suggested relevant next steps, and guided people through the interaction.
EntOS was a shared entertainment platform supporting Xfinity, Xumo, XClass, Sky TV, and future partners. We needed recognizable assistant behaviors that could adapt to different brands, extend to mobile, and support future interactions across TV and phone.
Voice and multimodal UX
Behavioral systems
Design direction

A shared assistant interface adapted to different brand identities.
THE BIG PICTURE
MY CONTRIBUTION
I defined assistant-state behavior and communication requirements; directed motion, speech feedback, and error handling; and originated smart suggestions with design and engineering. I guided shared patterns into the Everything App prototype and explored how imagery could enrich mobile assistant responses.
PROJECT OUTCOME
Delivered production specifications and design assets for the EntOS assistant v1. Working TV and mobile prototypes demonstrated smart suggestions and reusable interaction patterns across devices.
Define how the assistant communicates
I defined what each assistant state needed to communicate: when it was ready, listening, processing, responding, or needed recovery. Motion, text, sound, and feedback had to make those transitions understandable.

Turn the model into buildable behavior
I directed the translation of the interaction model into specifications covering state triggers, transitions, timing, speech transcription, and coordinated visual and sonic feedback. The team’s documentation made the behavior concrete enough for engineering to implement.

02 / LISTENING FEEDBACK
Make listening unmistakable
Early sphere and halo treatments suggested activity but did not clearly indicate when voice capture began. Users needed an unmistakable listening signal.
I made that requirement explicit in the motion brief and directed Jonathan Alsop to explore a stronger cue. On TV, the resulting branded animation extended across the top of the screen; the Everything App later adapted the principle to its bottom input area.

The TV listening treatment extends beyond the assistant icon, visibly signaling that voice capture is active.
03 / SMART SUGGESTIONS
Anticipate what people might need next
I originated and advocated for smart suggestions so the assistant could surface relevant next steps without requiring people to invent another command.

With Twisters selected, suggestions offer three directions: ask whether it is scary, explore more films starring Glen Powell, or switch to comedies.
Working with conversational AI design and prototype engineering partners, I explored suggestions informed by the user’s query, content information, and subscription access. We constrained generated copy to fit the interface and developed the behavior in a working TV prototype. Because richer visual responses were outside the early scope, suggestions used the existing assistant area.
TV and mobile placed different demands on the assistant. On TV, listening and responses needed to remain clear without interrupting viewing. We used working mobile prototypes to evaluate contextual suggestions and richer follow-ups, applying consistent state, feedback, and recovery principles without forcing identical interfaces.
SHARED BEHAVIOR / MOBILE PROTOTYPE
Mobile prototype: the assistant responds with device-activity context.
A richer answer, not just more text.
I explored using movie imagery alongside assistant responses, showing how a conversation could continue and become more useful on a phone.
MOBILE INTERACTION PROTOTYPE
Deliver the foundation for v1
The project delivered production specifications and design assets for v1 of the EntOS assistant interface to the production development team.
A working TV prototype demonstrated smart suggestions, while the Everything App showed how both that capability and the assistant interaction patterns could extend to mobile.


