Europe's first automated restaurant checkout system

In many businesses around the world, Compass Group runs catering facilities for employees. One of the main pain point we identified in the user experience were the long queues of customers waiting to check out, watching their food go cold on the trays. In a office restaurant, everyone arrives at the same time, and it is difficult to scale the staff and number of tills open.

Manned tills have several issues, they take up a lot of space in often tight environments, they are labour intensive, and stressful to employees as the rythm is intense. So we designed a fully automated, compact, AI based smart checkout.

Client : Compass Group

World's largest catering company, providing meals all over the world to hundreds of thousands of businesses, schools, hospitals, events..., present in over 45 countries, more than 600 000 strong, and yet, in many aspects, oblivious to the competitive world outside its high walls. In recent years, they have been winning over corporate clients, but loosing end users : customers, coming to their premises. The numbers weren't good and i was called in to help understand the issue and reverse the trend.

Case Study Banner

A 3 point strategy for a better user experience

Field observation, user and staff interviews, business analysis, after a few weeks of studying how the services operate, we were able to draw up an actionable plan addressing 3 aspects of user experience :

Research Image

Removing Pain Points

Field observation, user and staff interviews, allowed us to draw up a very detailed picture of the issues clients were facing.

Research Image

Increase functionality

Developping more ways to consume and manage the services. Thanks to user research, we could concentrate on functionnality that would have the most impact on users.

Research Image

Give more than expected

We then identified the key moments in the user journey, and focused on building a gratifying and rewarding system, driving deeper user engagement.

What is the Smart Checkout ?

The smart checkout is an automated checkout system for catering facilities, allowing a substantial increase in the speed and amount of users processed at anyone time. Built around a neural network and a database of millions of pictures, it recognizes everything on the client's tray, and display's the bill on the screen, in less than 10 seconds.

Developping the Smart Checkout has been a challenge. The concept was pushing boundaries both in terms of user experience, and technology. We worked hand in hand with startups at the fore front of AI and Machine Learning, to turn a theoretical concept into a scalable product for the real world.

Story Image
Story Image

Extensive research

We took time to prepare. We studied the way users checked out, how they flowed in the premises, how they interacted with staff and technology, we interviewed many, defined personas. We trialed concept ideas with them, cardboard cut outs at first, then screens and interface designs. We observed and logged everything, until we felt the user experience was good enough to build the product. 

Research Image

1. Field research

Field observation, user and staff interviews, allowed us to draw up a very detailed picture of the issues clients were facing.

Research Image

2. Desgin iterations

We built many product concepts and interface mockups, and tested them extensively in the field, iterating fast on each design.

Research Image

3. User testing

The first prototypes we put to the test in real world conditions, and thoroughly tested. Users were interviewed about their experience. User testing carried on well beyond concept phase into full deployement.

Initial designs

Story Image
Story Image
Story Image
Story Image
Story Image
Story Image

Design System and Style

We created a design system that would not only create consistency but also save time for the client as they build new features. This will improve the overall process and experience for the website.

Style Image
Style Image
Style Image
Style Image
Style Image
Style Image
And many more...

User reactions