Can you challenge the insurance industry with a chatbot?

What did we do?

Want to chat insurance? We’re guessing your answer is no, and the truth is - we feel the same way. It’s a no-fun-zone topic that tends to cause either confusion, stress or both. Last year, it was reported that workers in the UK were undersaving for retirement by up to £11bn a year - meaning that most are not saving as much as they need to, and there’s a lack of access to helpful information that’ll solve the problem. So how can tech help reach more people with a personalised, useful service?

While we didn’t set out to create a captivating conversation around insurance, we wanted to find a way to shake up the industry a bit while simultaneously removing the stressful, confusing bits involved in thinking about your pension. Enter: Our new chat bot, MAT bot.

Our goal was to create a friendly financial assistant to help employees with their pension planning. But first things first - we wanted to test it out. Our first step would be to design a working prototype to put in front of users, and ultimately find out if they’d be interested in using a chatbot to discuss and re-organise their finances.

Client

Zurich

Year

2017

What we did

Concept

User Experience

Design

Development

How did we do it?

As the chatbot technology is relatively new and constantly evolving, we started out by searching for the right tools for the job. We looked at Chatfuel, Botsify, API, Botsociety, Motion AI (plus a few others) and compared cost, effort, and features to decide which tool would make the most sense for this project. Motion AI ultimately came out on top for a few reasons: the price was right (free), it allowed us to choose from different platforms (SMS, Facebook, web-based, Slack, e-mail), and the learning curve was fairly gentle. On top of that, Motion AI hosts a Slack community dedicated to their community users - a helpful resource for questions, feedback on and open discussion about the tool.

Once we were confident with MotionAI, we mapped out a few user flows to map out the user’s experience from beginning to end with MAT. As with most of our design projects, user flows were key in getting our thinking straight and making sure the user experience was at the forefront of our design. In this case, we had to take a step back from our normal processes and re-think our user flows. Chatbots require you to think about the flow as two-sided, and ultimately think about the conversation (and the goal of the conversation) before mapping out a useful flow. Our tip? Think about the conversation first, and then map out the user flows. This will allow you to look at both sides of the interaction, and ultimately map out the flows more easily.

We set the flow in MotionAI, and then worked to create an approachable, stress-free and simple design (complete with an illustration of the one and only MAT, which can be seen below).

MAT Bot (rooted in Motivation, Ability, and Triggers) was fairly simple with its introduction, simply greeting users with “Hello, I’m MAT Bot. I’m here to help you save for a better retirement...without all the thinking!”

What happened next?

MAT Bot is alive and well, and our first iteration is being used to start conversations with potential users, gather data on their goals and needs, and stimulate discussion with our client about the best business applications of the technology. Stay tuned.