The challenge

Takeaway decision maker

Create a rule-based chat-bot for helping with your decision on what to order via take away on a lazy Friday night. Pay special attention to the chat bots personality and microcopy.

In this study we used user personas and did some prototype testing.

This challenge was from 21 September until 25 September 2020.

The team

For this one week challenge I worked with Camille Buyle & Lara Tomme.

User research

User personas

We made some user personas to help us understand the users’ needs, experiences and goals. 

User persona Kim
User persona Stef

It is clear that both personas have different needs when ordering food. Stef uses Take-away because cooking no longer fits into his schedule. It’s effortless and he can eat whatever he likes!

While Kim likes to discover new dishes. Because Take-away has such a wide range of foreign kitchen, this is perfect for ordering unfamiliar dishes.

Brainstorm

What is a rule-based chat-bot?
A rule-based chat-bot is a type of bot, where communication is through pre-established rules. The user’s input must comply with these predefined rules to receive an answer (Flow.ai, n.d.)

On the Takeaway site we searched for different categories of food. There we found 5 major categories: Family dinner, healthy meals, comfort food, food to share, tasty desserts. We only focused on the main dish, so we excluded tasty desserts.

Inspiration

To help us prototype a rule-based chat-bot, we searched for already existing chat-bots.

Design of prototype

Testing

After we made wireframes, we started testing them to gather feedback. we gave the users a specific task and observed how they performed it. A number of questions were asked afterwards.

Test #1
Our first tester was a 21 year old student. He went smoothly through the prototype and experienced it as an easy process. The use of the emojis was less for him, but it made the chatbot interesting.

Test #2
The second tester was a 29 year old project manager. He found it very clear and intuitive.

‘It felt like a real conversation’

Test #3
Our last testing was a 20 year old student. He also went smoothly trough the prototype. He was suprised that he could swipe between different meals.

‘It was recognizable and easy’

Conclusion

We had to create a rule-based chat-bot that helps you to decide what to eat on a friday night. After some research into the Takeaway site and to already existing chatbots, we made up a prototype. It was important that the chatbot meets the users expectations so we performed several user tests. After the user tests we could conclude that the chat-bot did meet the users’ expectations.