Product Designer

Designer @ DICK'S Sporting Goods. Optimizing workflows.

© 2026

Guiding shoppers into a retail AI assistant

DICK'S Sporting Goods · 2026 · Product Designer · 1 Week Sprint

OVERVIEW:

From Discoverability to use as quickly as possible

Coach by DICK'S is an initiative to leverage AI to assist users in making product decisions. This project was dual purposed, one to create high visibility for entry points, and two to reduce time taken to first message.

Average weekly AI sessions7×
140sessions
Pre – Wk 6
2001000
PreWk 1Wk 2Wk 3Wk 4Wk 5Wk 6
Time to first message (indexed)−15%
85
Pre-launch – Wk 6
100500
Pre-launchWk 2Wk 4Wk 6
Indexed to pre-launch = 100. Shoppers reach a useful message 15% faster.

PROBLEM:

Invisibility + Friction

Hard to Find

A feature that's hard to find will never get used, especially if that feature is new. Coach as a pattern needed more visual contrast to be used.

Slow Start

User's who did find Coach had to type in their messages and responses. By giving selectable prompts, we launched users into the experience.

DISCOVERY:

Where did users expect to find the feature?

User Testing

To find the answer to this question, I ran 20 digital user surveys with our customers, asking questions that gauged entry point expectation.

Additionally I ran a multi-variable study across potential entry point versions and compared which version had best success time, etc.

DESIGN DECISIONS

Functional Proximity is the key

Users consistently stated that they expected AI entry near Search, as the AI was most utilized in product discovery in the app. This was supported through time taken to screen metrics on entry point variants.

SOLUTION:

Higher Visibility = Higher engagement

Kickoff experience with Selectable Prompts

With higher entry rates, we needed a way to kickstart the experience for curious users. Selectable prompts were an easy yet effective way to start a conversation without the user having to type everything out.

The prompts were chosen specifically to match each of the AI's subagents, tailored to different capabilities the AI could provide. To validate and explore competitively the interactions for this, a Miro board along with error states and happy paths was developed.

Caption describing these three images.

RESULTS:

Higher engagement into Coach AI

Average weekly AI sessions7×
140sessions
Pre – Wk 6
2001000
PreWk 1Wk 2Wk 3Wk 4Wk 5Wk 6
Time to first message (indexed)−15%
85
Pre-launch – Wk 6
100500
Pre-launchWk 2Wk 4Wk 6
Indexed to pre-launch = 100. Shoppers reach a useful message 15% faster.

WHAT'S NEXT:

How can we move beyond AI for discovery?

AI is constantly changing and expanding its capabilities. Taking this project further I think understanding how user goals could reveal potential entry points for AI and the future of AI integration…

Although user surveys can capture a wide range of data in a quick amount of time, the depth of data might be lacking. For future studies I hope to leverage a more deeper insight from the customer, how they interact with AI and how this perception might influence the experience.

Next

Bulk Quality: one approval flow for fabric samples →