Designing for Empathy in Conversation Design

Views about conversational AI no longer stop at their ability to be helpful and sound natural. While those remain key aspects of the design and use of conversational interfaces, the emotional experience is becoming an increasingly important factor.

From investigations into how empathetic chatbots can affect mood to studies on how empathetic conversational AI can enhance digital mental wellbeing, there’s a growing interest in the possibilities of emotional resonance for conversational AIs.

It’s also becoming clear that an approach that accounts for emotion presents benefits for consumers across different use cases, and could be an essential element for conversational AIs to realize their true potential.

Conversational AI and Human Relation

The relationship between users and their voice assistants has crossed the threshold of cold, human-to-computer interaction. As the capabilities of voice user interfaces have evolved, so has people’s willingness to establish deeper connections with them.

Numerous stories bear witness.

The BBC ran a story about Alexa Jett, a woman who credits her voice assistant with helping to raise her out of a depressive state.

And it’s one of many, joined by stories such as that of Replika AI, a friendship-emulating conversational AI, which saw its app store downloads spike by over 50% during worldwide lockdowns.

Such an attitude towards voice assistants and chatbots is not something that’s merely limited to companionship-focused voice AIs.

Even those that are designed for more domestic applications or those with corporate and consumer-facing purposes are experiencing the same warm attitudes from users.

Research by Think With Google confirms this, with 41% of surveyed users saying interactions with their conversational AIs feel like “talking to another friend.”

Why is Empathy Important in Conversation Design?

Keeping users engaged is one of the core issues conversation designers focus on. An engaged user continues to use the AI more and more, and they continue to explore its features.

On that note, research shows that an emotional connection is a necessary part of any user experience. While it doesn’t need to be overt, there should at least be an undertone of emotional resonance.

As with other applications, users interact with the conversation AI more if they feel it’s in touch with their experience.

This is all the more relevant as conversation designers place empathy within a broader framework: Providing the most incentivized and frictionless experience possible.

Signs of emotional understanding give users affirmation that they can continue to explain their situations without needing to filter out the emotional cues that are simply a natural part of their speech. Even better, they can get the sense that these emotive words are helping to refine the response they receive.

Empathy is also crucial in conversation design because of how it helps users to iron out the exact meaning behind their queries.

If a user feels understood, they’re more encouraged to provide further detail about their situations. When they don’t feel understood, on the other hand, they’re likely to rush through the query and omit important details.

More detail in this case is helpful for both the users and the conversation AIs. For the users, it allows for solutions that address the exact query. While conversation designers and businesses can benefit from having more data on the issues that users are experiencing, enabling them to iterate the conversation AI itself and broader business processes with greater focus.

How Empathy Helps Companies Investing in Conversation Design?

Business Insider’s projections suggest that spending on conversational AI will reach $9.4 billion by 2024, after showing a 29.7% annual compound growth rate.

As companies continue to spend on conversation AI, their focus will increasingly turn to how they can get the best return on investment. And empathy should certainly be part of that conversation.

Research by Motista shows that an emotional connection can single-handedly increase customer lifetime value by 306%.

That means, for companies to maximize the value they receive from deploying chatbots, it’s crucial that the bots are able to facilitate emotional connections with consumers. As the research cited above shows, that capability has a direct bottom-line impact.

If companies are going to realize savings or even revenue boosts from their investments in conversation design, it’s important that the conversational AIs account for emotional connections as powerful drivers of consumer behavior.

Empathy is important for simply getting users to adopt the conversation AI in the first place. A study by the University of Oslo in Norway reveals that an emotional connection strongly determines whether users continue interacting with a company’s chatbot.

Customer service is a key segment to focus on.

For companies, customer service has long been hailed as one of the most impactful verticals for conversation AI.

Juniper Research’s study suggests that chatbots will be able to deliver $11 billion in annual cost savings for the retail, banking, and healthcare sectors. And they’ll reportedly save consumers and businesses over 2.5 billion customer service hours by 2023.

Without the chatbots exhibiting empathy, however, those figures may not translate to actual gains that companies experience.

Gartner reports that a whopping 96% of customers who go through negative experiences consider being disloyal in the future.

So, for companies to avoid churn, it’s essential for their conversational AIs to be able to register the right emotional tones with customers.

It’s even more efficient for the conversational AIs to be able to pick out words or speech patterns that hint at a negative customer experience, so they can immediately deescalate and try to make the customer feel more understood.

How Do You Show Empathy In a Conversation?

Natural language processing makes it possible for conversational AIs to pick up on cues that help to establish the emotional undercurrent in a user’s query.

Conversational AIs can use these to display one of the three types of empathy that is most appropriate for the situation:

● Cognitive empathy

● Emotional empathy

● Compassionate empathy

Cognitive Empathy

Cognitive empathy is best suited to conversational AI that is used in formal settings. An example is a customer service chatbot in a corporate setting, where interactions need to retain a certain level of formality.

For this, the aim is to show that the conversational AI understands the situation, and isn’t just picking out the next response from a database.

A conversation that shows cognitive empathy may go like this:

User: I’m having trouble signing in to my account.

AI: Hi! Sorry to hear you’re having trouble. Have you checked that the email

and password match the ones you registered with?

User: Yes. I also sent requested a password recovery email but I haven’t gotten

it.

AI: Oh, that’s unusual. Let me pass that onto someone who can help

straightaway. Bear with me a second.

In that exchange, the AI displays an understanding that the situation the user is experiencing is unfortunate, and it even expresses concern. But, in line with the setting, it doesn’t express that through overly emotive words.

Emotional Empathy

Emotional empathy in this case is more about showing that the chatbot or voicebot understands the emotional undertones of the user’s query.

Highly emotive words offer the easiest options for AI to recognize in this case. That includes words and phrases such as “happy,” “satisfied,” “disappointed,” and others.

Many emotional markers, however, are subtle and require more advanced natural language processing on the part of the AI.

And emotional empathy also involves the conversational AI processing the user’s query and drawing the likeliest emotion they may be feeling, then stating it to show a deeper understanding of the user’s needs.

An example of a situation in which an AI may have an opportunity to show emotional empathy is in the dialogue below.

User: Hey. I’m thinking of opening an account and I need to know a few things.

AI: Hi! Happy to hear you’re thinking of opening an account. What are you

curious about so I can help?

User: My playing times change from month to month. How easy is it to switch

plans after buying one?

AI: Oh, right. You’re free to choose a plan that works better for you whenever

you like. We have several that are flexible.

Compassionate Empathy

Compassionate empathy, meanwhile, is about the chatbot or voicebot showing that it has placed itself in the user’s situation.

It can do this by using words and phrases that show that it’s “experiencing” the same emotion as what the user may be feeling. It can also go further to demonstrate emotional urgency that mirrors the severity of the situation.

Even further than that, the chatbot may demonstrate a desire to lift the user’s spirits.

For example:

User: Hey. I’m worried my package may be lost. It’s been two days since it was

meant to arrive.

AI: Oh, no! Can I have the tracking number? I’ll look it up right away.

User: 777884855540xi81

AI: Thanks. And don’t worry, we’ll take care of it for you. Just give me a second.

How to Design for Empathy?

Making a conversational AI empathetic begins at the earliest design stages, where the designer makes conscious choices to make sure the AI is able to display convincing emotion in its responses.

Designing for empathy can be whittled down to four steps:

Know the Customer

The first element of designing for empathy is developing a more in-depth understanding of the customer. This helps to discover the responses that are likely to strike that particular customer as genuinely empathetic, based on their demographic qualities and general habits.

You can do this through gathering test groups, conducting surveys and carrying out customer interviews.

This also allows the chatbot to offer a more personalized experience, which can boost user engagement by up to 80%.

In the end, you will have several profiles that help you to get a clearer picture of your target customer, and how best your conversation AI can facilitate emotional connections with them.

Carry Out Extensive Research on the User and Use Cases

Your conversation AI will likely be built with a specific purpose in mind, whether that’s to provide greater utility with Internet of Things devices, customer support, employee assistance or something entirely different.

Depending on that purpose, there are specific use cases that will come up repeatedly, including those where users might be anxious, under stress, or in need of emotional support.

Pinning down those use cases can help you see when users will require empathetic support, along with the best ways to provide it.

Defining Your Target Audience

It’s equally important to define who your target audience is. And that means breaking down their characteristics at the deepest level you can.

While doing this, you may find that your target audience comprises people with numerous different qualities, some of whom may fall into different categories entirely.

For example, a customer service chatbot for a cloud computing software may have to cater to individual users who may simply be small business owners. On the other end, it may also need to support enterprise customers with hundreds of users under a similar account.

Understanding your core target audience, especially how it’s segmented, can ensure that the emotional range of your conversational AI covers all of them.

Test Product Rigorously At Multiple Stages

Don’t leave testing till the end. Test the product at multiple stages of its design, picking out and solving problems at each phase.

Performing rigorous tests at multiple stages helps you to see if the empathetic features you’re developing for your conversational AI are working the way you expect them to.

It also ensures that you don’t leave problems to pile up until the end, to the point where the conversational AI becomes unusable, or empathy features don’t contribute to the user experience.

Wrapping Up…

As adoption of conversational AI spreads and companies pour more resources into it, it’s important for the design process to take account of the emotional aspect of the core user experience. That way, conversation AI can not only deliver more value and service to users, but they can become suitable for more expansive use cases, while allowing companies and conversation designers to reap the best possible return on investment.

About Digital Assistant Academy

Digital Assistant Academy provides Voice Interaction Design and Conversation Design training and Certification. In this program, we will take you from the very basics of voice interaction and conversation design, through to how voice technologies work. We’ll do a deep-dive into conversation design strategy, and it will be fully hands-on with your Capstone projects. By the end of the course, you will have two voice applications successfully designed, developed, and deployed. Learn more at Digital Assistant Academy https://www.digitalassistant.academy/

Before you go

Clap 👏 if you enjoyed this article to help me raise awareness on the topic, so others can find it too
Comment 💬 if you have a question you’d like to ask me
Follow me 👇 on Medium to read the more articles on Voice Technology and Conversation Design, and on Twitter @sprayaga