fbpx

A chatbot’s job is the make everyone’s life a little bit easier. Of course, bots can make our teams more productive and more efficient by allowing them to focus on complex tasks, to work on sales-ready and qualified leads, and much more.

At the end of the day, though, bots are for the user. We want functionality, but our first concern should always be optimized to the user’s interaction with the bots. Just like we want our websites and mobile apps to be user-friendly, we also need to consider our chatbot’s UX. Since the bot is often a customer’s first point of contact with a brand, we want it to reflect our values and standards.

Here are the three conversational UX design principles we use to design our bots.

Design to be Understood

We want our chatbots to be clear, concise, and cognitive. Imagine you are a customer of a local home repair store who is looking for answers about whether the product you have in your hand is the right choice for your project. You wander through the aisles for several minutes in search of an unoccupied sales rep.

Eureka! You finally find one and approach with your inquiry. “Excuse me,” You start before continuing your explanation. The rep then responds, “Mi scuso, non parlo Inglese.” Did I mention that you’re in Italy and that the rep does not speak English?

There’s that sinking, alone feeling that one gets when not understood. This is no different when interacting with a chatbot. The bot should understand the user to the same degree another person would understand them in their language, and, vice versa, the user should clearly understand the responses from the bot.

Keep it simple: an all too familiar adage. When designing chatbot UX, the most crucial part is that we want to be concise. Chatbots often interact through mobile platforms, so we want to send short messages that are easy to glance at. If we need to send a longer message, then we can either break it up vertically into multiple messages or horizontally into a carousel that the user can scroll through.

We don’t want to overload the user, so we need to limit how much and how often the bot replies. The last thing that we want is for the user to have to scroll up to read a full message or else risk missing out on a piece of crucial information.

We suggest creating time gaps where the bot is ‘typing.’ This simulates messaging with humans and gives the user a chance to digest what the bot said. This removes user anxiety, as it creates an experience that is more comfortable, natural, and conversational in nature. Pausing and time gaps, strategically woven into the greater conversational flow, make a big difference in the user experience.

Finally, we need to make sure that our bot asks the right kind of questions. We recommend prompting users with buttons for closed-ended questions that prompt a yes or no response. This helps to guide the user to the precise responses that we’re tracking. With the right design, we can use open-ended questions to capture user responses in a variety of formats like phone number, email, or text. This is a great tool for revamping market research or lead gen forms.

It’s important to use the right words. We want to be concise and to solicit the right feedback. If we use open-ended questions, for example, then give examples and tell them the exact format we want. For instance, if the bot asks for a zipcode, they may say, “What is your zip code (90210)?”

Design to Understand

Humans are unpredictable. They can and will type anything into the chat box. We need to prepare accordingly.

The first way that we design bots to understand users is by supporting as many data formats as possible. There’s usually more than one way to say the same thing, so we want our bots to be able to understand as many of these as possible.

For instance, if a user is scheduling an appointment with a bot, we want the bot to understand that “7/15,” “July 15th,” “15 July” and “next Tuesday” all mean the same thing.

To move the conversation forward, we want to validate every reply from the user, both in type and in content. It can be as simple as, “Great, thanks! I’ll set up your appointment for Tuesday, July 15th.” This feedback loop helps maintain a quality customer experience and, in more sophisticated scenarios, continually teaches machine learning algorithms that drive some bots.

Another component of any well-designed bot is the ‘do not understand the scenario.’ It’s inevitable that a bot will encounter situations that it won’t understand, and we want to make this experience as painless as possible for the user.

Begin by being clear about what the bot doesn’t understand and why it doesn’t understand it. The bot should tell the user what it can understand and then retry the prompt.

The last thing that we want, however, is to strand the user in a loop where the bot just retries the same prompt. So, after a certain amount of tries, stop the conversation flow and escalate to a human for a hands-on approach. All bots, just like any business, should have an escalation protocol to address the matter that is best served at higher levels of the authority chain.

Our bot can say something along these lines: “I’m having trouble understanding you. I am looping in a live agent to who can help you out. Meanwhile, can I help you with anything else?” At Botveu, we build our chatbots to weave into your existing customer service processes to proactively monitor and escalate user scenarios that require human attention.

Design to Learn and Remember

We take it for granted, but humans remember and learn from conversations seamlessly, without purposely thinking or mechanically processing actions to remember. It just happens! Your friends, family, and co-workers keep track of your preferences and remember things that you tell them.

Bots should do the same. But we need to approach it in a manner that helps the bots learn our preferences, names, and histories. The result is promising, but it can be just a pipedream if we don’t design for recall utility from the start.

For example, after our user schedules an appointment at a certain location, our bot can keep track of that and offer to schedule later appointments at that same location. Of course, we’ll offer an option to change locations.

One concern here is making sure that we’re clear about what the bot knows and how it knows it. We don’t want to appear invasive or breed mistrust with our users. For the same reason, we should give users an option to tell bots to stop collecting or using their personal information.

Conclusion

By keeping these UX design principles in mind, Botveu creates chatbot experiences that are optimized for the user. Essentially, our goal is to make our bots seem as human-like as possible. This starts with great customer service.

To learn more about how we can incorporate these principles into a chatbot that’s customized for your unique organization, contact us today.


Leave a Reply

Your email address will not be published. Required fields are marked *