The challenge

Food delivery is one of the most dynamic sectors of catering that has now expanded its borders to increasingly personalized offers. In 2018 more than one Italian in three (37%) used a digital platform to order food from the telephone or computer, with an increase of 47% over the previous year.

Our team was given the problem of evaluating a parent’s experience in food delivery platforms.

The solution

We created Beet, a chatbot based on the Messenger platform. Beet through some starting questions, learn the tastes of the family, recommend the kitchens to do, plans your weekly shopping and gets it directly to your home.

My role

UX Research, UX Design, User testing

Design process

Lean Startup Machine

Lean startup is a methodology for developing businesses and products, which aims to shorten product development cycles and rapidly discover if a proposed business model is viable. this is achieved by adopting a combination of business-hypothesis-driven experimentation, iterative product releases (MVP), and validated learning.

Proto Personas & Customer Journey

Anna The Super Mama

Anna is a 45-year-old mother who work 8 hours a day and she often can’t prepare meals for her children due to her numerous commitments. She often happens to order on the app for the whole family. It generally uses Foodora or Uber Eats. When she orders, she is interested in the quality of the food and is attentive to the ingredients in the dishes.

User Research

Qualitative Research

We conducted a field research dividing it into two different phases.

In the first phase we wanted to validate our proto-personas.
The first phase of research was carried out on the first day at 8.00 am in front of the school entrance in Via Vallarsa, carrying out 15 interviews on the use of Food Delivery apps.

The main result of the first phase was to invalidate the proto-personas and build a new personas based on the insights of the research.

The second phase of research was carried out on the same day from 12 to 17 in the referring personas. 13 interviews were carried out with the focus on the organization of weekly shopping and preparing meals for the family.


User Research

Quantitative Research

At the same time as qualitative research we carried out quantitative research in the form of an online survey. The form was promoted in the Facebook page “Mamme di Milano” “MammeMi” “MamiClub” “Zone 4, Mamme e papà di Milano”.

To share the form on the closed facebook page, we used a Facebook account of a team member’s mother.




First prototype

Paper prototype

App for planning weekly meals

We have made a paper prototype to understand the whole flow. The interface, through registration and some questions, recommends a weekly menu that you can browse and modify.


Most of our users do not use apps, or at least not on an ongoing basis.


Second Prototype

Chatbot on Messenger

Planning and spending service through chat

We have created a sponsored facebook page and a landing page to land our user in the chat and start the conversation to plan meals according to school lunches, recommend meals to cook and shop directly from the chat.


User testing

We conducted an agile test with our reference personas

There were 5 tests of about 10 minutes each one at the elementary school exit and inside the supermarkets. For the test we used a “Wizard of Oz” approach.

Chatbot responses were manually written by a team member remotely while two other team members were on the field to prove the actual interaction.

Make the chatbot

Building of a real chatbot using chatfuel

We coded our own chatbot using the tool Chatfuel and we actually implemented it in our own Facebook page “Beet”