AI services and enterprise workflows.

Staff Product Designer across AI as a Service and enterprise product experiences.

AI as a Service

The thinking behind AI transcription

Automated end to end once the user consents, so busy executives can reach meetings and knowledge by chat or voice.

Consent before automation.

Nothing syncs from Outlook until the user opts in.

One occurrence at a time.

A series shows only its next meeting. The next one appears when the week starts.

Only joinable meetings pass.

Duplicates, no-link, unaccepted and unsupported meetings are filtered out.

Ask the week out loud.

This week’s meetings on one dashboard, questioned by voice.

Attends on the user’s behalf.

It auto-joins Teams meetings, records and transcribes as a daily routine.

Access only after sign-in.

An Entra ID confirmation connects the knowledge base to the user’s own MCP tools.

Separate what an avatar says from how it moves.

I replaced text-only prompting with a timeline canvas that separates the avatar's script from its gestures.

The constraint

Enterprise clients needed explicit control over automated movements to protect their brand. Unnatural gestures could undermine the experience.

My decision

Make the spoken script and gesture choices independently editable within the timeline, rather than leaving both to a single text prompt.

What changed

The editing model gave clients control over script and movement separately. The intent was to reduce unnatural animation while keeping brand-safety decisions in their hands.

Script and gesture controlConceptual model
Illustration of the design decision, not a client interface. No internal screens or project data are shown.

The thinking behind AI avatars

Two tracks instead of one prompt.

Script and gestures are edited separately on the timeline.

Brand safety stays with the client.

Movement stays client controlled, so automated gestures never undermine the brand.

Enterprise decision tools

The thinking behind AI-driven market intelligence

Insights, internal KPIs and a market data feed turn into actions and plan scenarios.

From signals to actions.

Insights, internal KPIs and market data come together in one flow that ends in actions.

Actions become plan scenarios.

Each action carries into scenarios the team can compare side by side.

Organizational workflows

Discuss my contribution.

For a relevant role, request a conversation about my responsibilities and the design questions I can discuss publicly. A request does not grant access to restricted material.