Last year, design studio Deepr delivered some design training and support to charities as part of Tech for Good Build. Here, Matt McStravick shares tips on building human connection into design.
Chatbots have gained huge popularity over the past few years and are now a vital cog in the digital user journey. They’re a brilliant way to direct users to the most appropriate service and sometimes a way for them to resolve whole issues. They’re an important way of supporting more people whilst saving on costs.
However, recent research indicates that 60% of people would still prefer to talk to a real human being than a bot and we know that motivating people to use them can be a challenge - especially when they’re asked to share lots of information. And this isn’t solely a generational issue. Our own research with 25 young people found that none of them would choose to use a chatbot if they needed urgent or non-urgent support.
Last year, as part of the Tech For Good Build programme, we supported charities to design more meaningful human connection into their digital services.
Three of these projects included chatbots, so we thought we’d take the opportunity to share some of the insights we’ve developed from working in this space.
Chatbots don’t have a great reputation with everybody. The name alone can highlight what you’re not going to get from the experience - a chat.
The technology is great though so why not ditch the name and just call it what it is, an ‘info tool’, a ‘service finder’ or whatever best describes its purpose.
Doing this avoids many associated connotations.
Hopefully this goes without saying but please don’t do this under any circumstances. The road to developing the trust of your users is long and rocky. Fibbing won’t help.
It can be valuable to be clear and open about where the content of your tool comes from. If it has come from the questions that people generally ask and the answers your expert team provide, then say so - this creates a link between the digital experience and the credible, warm human beings on your team.
The stronger you can make that link, providing names, photos etc - even voice notes, the better.
All of the above aims to re-humanise the content i.e. your users know it's coming from experts, which enables them to develop trust and connection.
We’ve found that if people need to share many pieces of information as they go through a chatbot process, it’s important to share snippets of valuable content periodically along the way. As they provide some data - reciprocate by providing them with something of value too.
For example, instead of one long decision tree that may take five minutes before providing something valuable, you can highlight the first point at which you've identified someone's problem by sharing some of your own data e.g. “80% of people who experience this problem are able to have it resolved through our service” or “Rest assured, you are not alone. This is a common problem that many people experience and we're able to help”.
Your user will know that they’re not at the point of getting everything they need, but they’re getting a little bit more clarity at each step of the journey - creating greater intrinsic value during the process.
And when we can bring this reciprocity to our service we’ll be building small bonds of trust that we can build on. It's tiny, but really significant.
To take this idea forward more generally you might want to start by asking the question “What data do we have that might be useful to our users?”
Another small but significant way to build connection is to provide social proof (or ‘evidence from people like me’) that your service is able to help - throughout the user journey through the tool. If you can show quotes (with photos and names ideally) from other service users who've had support with similar issues, it will build users’ trust in your service and enable them to feel they’re not alone in the process.
There’s a difference between language that is professional and builds trust, and language that is cold and inhuman. Base your content and tone of voice on the way the support would be delivered in a face-to-face context: it probably doesn’t have the formality of a funder report - and neither should your webtools.
An important point to note: Since we're not calling this a chatbot and because we're identifying where (or who) the content is coming from, we can legitimately use really human language to set this up.
There are different methods of gathering data for a chatbot to learn from, and what we find most useful is using the Wizard of Oz process. This means for the first few months of using a chatbot, a real team member responds directly to a user via the chatbot medium. As you slowly aggregate more and more answers to more and more questions, you're gathering the right data to use.
It also gives you time to both identify who within your team is the best at delivering kind, warm and supportive messages (as well as the technical information), and also weave in these answers into your chatbot technology.
This requires time and effort but it will create a better long term solution and avoid many negative teething experiences for users.
It's really important to feel heard. One of the most effective ways to do that in conversation is to repeat back what we've heard someone say. Sometimes it feels strange to do this but it never feels strange to hear your own words played back to you, instead it feels very rich. So how can we do that in a digital context?
One charity we worked with was using a chatbot conversation to develop a letter to be used to petition against the user’s housing issue.
A first response might be to present the final letter to the use with a standard “Check this form - do you want to make any changes?” question.
We suggest that instead, it’s valuable to play back content periodically to the user:
“Okay, what we've heard you say so far is [a small chunk of information]. Can you tell us [follow up question]?”.
What we're doing is playing the data back, framing it within a more humane context, rather than just a form to be checked.
So much of the benefit that people get from charity goes beyond getting their issue dealt with. Designing human connection back into digital tools such as chatbots is incredibly important if you want to give your service users the best experience.
Deepr is an end-to-end design studio creating better cultures and services guided by the importance of human connection.
Thank you to CAST for allowing us to share some of the valuable work that came out of this programme and to Laura Devaux for writing this blog.
Please submit your email address if you would like us to stay in touch from time to time with updates, new training or user research opportunities. We'll never pass your contact details to third parties.