Designing a hands-free, personalized mobile ordering experience for Starbucks

Product Design | Conversational AI | Mobile App

Company:

A global coffeehouse chain, popularizing darkly roasted coffee and creating a café culture. I worked closely with Starbucks Digital team to make digital innovation.

My Role:

My role expanded from brainstorm research activities and new concepts, to build interactive prototypes and iterate designs based on scenario-based assessments and A/B testing.

Impact:

After 80+ screens of iteration and a 94% task completion rate in usability testing, Starbucks successfully launched the concept for a personalized ordering experience, reducing browsing time for drive-thru orders by 50%.

A whole new way to order your coffee.

Back to top

Overview

As more locations reopen from lockdown in early 2023, Starbucks has set a goal to increase the percentage of mobile ordering.

Due to social distancing guidelines and in-store congestion, Starbucks aimed to shift more in-person orders to mobile. Our team addressed existing pain points and created a more seamless ordering experience for customers.

With the power of AI, we reimagined mobile ordering as a conversation, not a transaction. Users can interact with a digital barista who will take care of your hands-free ordering experience.

Jump to prototype ↓

The Challenge

A simple daily order can feel as complicated as ordering an Iced Vanilla Latte!

I ordered some ice vanilla latte with oat milk for my team using the Starbucks App.

But what happened?

They are not Vanilla Latte!

I can’t find it! Menu is complicated!

Okay fine, I will customize it.

Would 6 pumps be too sweet?

What? I got extra charged!

I can’t choose pick up time.

Concept Research

Aha Moment: It’s Easier Said than Done!

Originally, I was focused on helping customers to get their drinks faster, maybe with a better menu architecture or pick up experience.

But after I visited stores and interviewed with one-time and loyal customers with sketches, I realized speaking orders to a Barista is more straightforward than searching for my drinks on the app.

So I reframed the problem space to:

How might we help customers find, customize, and order their next drink through a simpler, more natural experience using the app?

Solution

Introducing the AI Barista!

A conversational AI agent that feels natural – just like talking to a barista, who asks smart follow-up questions.

Design Decisions

Information Display

I researched state-of-the-art AI chatbots such as Cortana and Siri and prioritized:

  • Clean, minimal design but still retain important information in card with contrast
  • Showing the user’s previous input followed by AI prompts with quick options
  • Natural, in-person-like conversational experience

Color Background

I focused on one user flow – ordering ahead in driving scenario with limited attention to the screen because we found many users order their drinks on the go. I set up my driving simulator and tested with participants to order an iced vanilla latte with oat milk.

Voice Input is active listening, so it is hands free!

I considered different use cases, including silent moments and incomplete inputs. Then I designed the system to surface smart suggestions after 5 seconds of inactivity, keeping the flow intuitive and responsive.

No Input

Active Listening

Introducing Tyler, your personalized AI barista!

I found a confusion regarding the functionality of the microphone icon on the homepage. I placed it in the center as the main action button, with pop up message for first time users.

Design Strategy

How should Starbucks make this happen?

I suggested evolving the concept in 3 phases, starting with incremental changes leading to revolutionary experience.

THANKS FOR STOPPING BY.

View my

résumé

eddychen805@gmail.com