Reimagining Search with Generative AI

role

Lead Product Designer

Platforms

Speakers, Mobile, Web, CTV

Timeline

2 years

Key outcome

↑ 20% engagement

Reimagining Search with Generative AI

role

Lead Product Designer

Platforms

Speakers, Mobile, Web, CTV

Timeline

2 years

Key outcome

↑ 20% engagement

Reimagining Search with Generative AI

role

Lead Product Designer

Platforms

Mobile, Web, CTV

Timeline

9 months

Key outcome

↑ 20% engagement

Role & Scope

Lead Product Designer

I led conversational search design for Alexa Audio across mobile, web, smart speakers, and TV, partnering with a voice designer to shape the early interaction model. The work bridged natural language, touch, and existing interfaces, defining how Alexa would surface results, follow up, and recover from misunderstandings at the start of the generative AI era.

Contributions

Conversational UX
Generative AI patterns
Design System contributions

Problem

Alexa had a midlife crisis

Customers had to use precise voice commands to find audio content, with minimal visual feedback. LLMs like ChatGPT changed expectations: people wanted natural conversation.

“It’s really good when I know exactly what I want and what to say, but when I don’t, it’s terrible."

Research participant, music & podcast study

Solution

Use natural language to discover your next favorite listen

A visual layer that pairs with voice, surfacing results, follow-ups, and context as the conversation unfolds.

An interface that's there when you need it…

Conversational queries surface rich visual results. Drove a 20% engagement lift.

…and fades into the background when you don’t

Direct playback when you know what you want. Protected listening minutes.

Scalable card system across all content types

Built flexible components that adapt to different content needs. Scalable design saved 60 dev weeks and improved browse.

From noise to value: a framework for effective related content

Related results framework minimized cognitive load. Contributed to a 67% engagement increase for recommendations users.

Before: Limited Discovery

Search responses surfaced only the primary item, limiting exploration and discovery.

Misapplied: Too much noise

Partner team overloaded responses with tangential, low-value items, reducing trust.

Refined: Hyper-relevant content

Applied strict quality filters to surface fewer, higher value items.

Refined: Hyper-relevant content

Applied strict quality filters to surface fewer, higher value items.

Smart suggestions that extend the conversation

Customers didn't know what to ask the new Alexa. Suggestion pills surfaced relevant follow-ups.

Unified experience across mobile, TV, and web

Adapted for each platform's strengths while maintaining consistency.

Impact

Reduced friction, increased engagement

Engagement

0%

0%

increase in listening hours

Engagement

0%

0%

increase in listening hours

Usability

0%

0%

below CPDR target

Usability

0%

0%

below CPDR target

Resource savings

0

0

weeks

dev time saved w/ design system

Resource savings

0

0

dev time saved w/ design system

Resource savings

0

0

weeks

dev time saved w/ design system

customer feedback

"Alexa knocked it out of the park. Connected me with a podcast I didn't know existed. I started playing it on the spot!"

In-person testing participant 6

customer feedback

"Alexa knocked it out of the park. Connected me with a podcast I didn't know existed. I started playing it on the spot!"

In-person testing participant 6