With their promise of taking the overall user experience and customer experience to the next level, chatbots are fast making their mark in the world of technology. Leading corporations are readily investing in artificial intelligence to build user experiences that run-of-the-mill mobile apps cannot deliver.
Chatbots offer the potential to enhance customer loyalty to a business. These human-machine interfaces can be so realistic users honestly feel like they’re having a live, one-on-one conversation with a real human being. The aim of creating an impeccable user experience lies at the core of a successful chatbot experience.
But users’ expectations of chatbots can be very high—especially when a user has previously had a good experience with a chatbot. Therefore, you must ensure that your chatbot has a human-like personality, as well as a life history and a visual identity.
However, software developers often make certain common mistakes when building chatbots—mistakes that can undermine the user experience. The last thing you want is users who are frustrated by a chatbot experience because the chatbot cannot understand what they’re saying. Or users who are creeped out because the chatbot says strange things.
Meaningful dialogue is what makes or breaks the chatbot user experience. To deliver the ultimate chatbot experience, you need to ensure that you deliver on all the UX essentials. In this article, I’ll describe nine must-haves that you simply cannot overlook when designing a chatbot user experience.
1. Keep Messages Precise
Even though your chatbot might have a lot to say, no customer wants to read a lengthy message from a chatbot at the beginning of a conversation. Splitting the information the chatbot must convey to the user into multiple messages is a much friendlier way to approach the conversation.
Using photos could help you get the word count down, while still giving the user pertinent, useful information. A general rule of thumb is to limit messages to no more than 80 characters per message, as shown in Figure 1. But pauses between messages should be brief, and the general cadence of conversations should be similar that of speech.
2. Provide Your Chatbot with a Visual Identity and Personality
Users don’t want to feel like they’re speaking to a robot, but a real, animate person. They want a chatbot to have a definite tone of voice—one that engages their brain and prepares them for a long, thoughtful conversation.
A backstory is an important part of a chatbot’s identity, and it is crucial that you define a personality and demographic characteristics for your chatbot. Customers might chatter in a way that could confuse a chatbot. Some users might even play around with a chatbot by asking such questions as When were you born? So having a backstory ready for such cases not only builds the user’s trust, but can keep the conversation going.
However, you don’t want your chatbot to offend or annoy your users. So developers must consider how the chatbot should react in different situations, from the mundane to the frantic. The best chatbots convey a unique personality that is specific to each customer, giving new meaning to personalized service. Users should feel that they are dealing with the same person each time they need assistance, giving them the experience of a nurturing, personal relationship.
3. Ensure Your Chatbot Doesn’t Ask Obvious Questions
Many of the earlier chatbots relied on using keywords to understand what a user was saying. But when a user wrote a lengthy message or multiple questions, this could confuse the chatbot.
People don’t always wait for an answer before asking a second or even a third question. But many chatbots can’t handle successive questions. So, when you’re developing a chatbot, the UX design needs to take into account the reality that the user might type multiple inputs in a row.
The chatbot should be able to map multiple questions in one message and correlate them to the appropriate answers. Rather than asking the user obvious questions, the chatbot should be able to determine what the user needs and ask a series of questions whose answers help solve the user’s problem, as shown in Figure 2.
The same thing goes for complex questions, in which users provide a series of informational tidbits in one sentence. If the chatbot is capable of mapping out different scenarios that could have different outcomes, the chatbot can ask additional clarifying questions to get the user’s query down to a single, simpler question. For example, by asking users what they’ve already tried, the chatbot can more quickly narrow down the potential answers and find the right one.
Chatbots that have only a limited vocabulary won’t be able to answer some simple questions, so the developer should build a chatbot that can organically build its vocabulary as it communicates with users.
Only then can it learn to ask more useful questions and be able to understand users with greater clarity. The ability for the chatbot to learn autonomously not only benefits the user, but the developer as well, because you won’t have to continually update the chatbot with relevant information.
4. Integrate the Chatbot with All Existing Systems
While chatbots can take on the work of many people, providing a true, human-supervised user experience is important. When a customer is trying to escalate a problem, you don’t want to leave your poor chatbot on its own.
All chatbots should have an established escalation channel that enables them to route calls or messages to actual human beings who are more capable of reading users’ emotions and giving them what they actually need.
Leave basic issues with well-known solutions to the chatbot and direct more complicated matters to human beings. Your chatbot should be able to handle common scenarios and realize when something isn’t within the scope of its user interface, then connect the user with a helpful human being, as Figure 3 depicts.
Chatbots should be able to access all relevant customer and company information so they don’t waste the user’s time. They should know how to optimize that information and use it in such a way that they can easily answer the user’s questions, whether they are about the customer’s needs or the company.
Rather than seeing the role of your chatbot only as relaying information, realize that it can also collect information. Some users contact your company to provide information or feedback about changes they’d like to see with your brand, your product, or your media. Designing a chatbot to retain and use this information to your benefit is not only good for the user experience, but also for your brand.
Your chatbot must be ready to cope with scenarios that are beyond its intended purpose, ensuring that it can still fluidly communicate with the user. Integrate your chatbot with your business’s customer-relationship management, ticketing, and fulfillment processes and systems so it can access all your customer information at any time.
5. Ensure User Engagement in Conversations
Right from the start, you need your users to want to be involved in the conversation with your chatbot. Many chatbots offer useless greetings that don’t give the user a reason to continue.
A simple hello is not enough to engage users or prompt them to ask questions or give feedback. Instead, your chatbot should ask specific questions for which it needs the answers and be very clear about the user’s next step.
Some of the best ways of convincing the user to start or continue a conversation with your chatbot include the following:
offering helpful recommendations that they can follow
making a task easier for the user
letting users know they can receive personalized updates—for example, order-confirmation and shipment notifications
Two scenarios in which you want your chatbot to function well are engaging users and marketing opportunities. Once you’ve connected your chatbot to other business systems, you can leverage the opportunities that chatbots have when chatting with users to give them relevant information they might not have requested—for example, delivery status or product recommendations. This gives your users a reason to continue the conversation with your chatbot even after you’ve met their initial needs. This is a huge win for your brand.
It is vital that rule-based chatbots keep giving users hints, prompting them to ask more questions. A chatbot cannot directly suggest that the user ask a specific question, but it can indirectly suggest such ideas.
6. Always Ask Users for Feedback
One of the easiest ways in which you can find out whether your chatbot is relating well to your users is to ask them, as depicted in Figure 4. Open-ended questions are best for this purpose and enable you to gather valuable information that you need to improve your chatbot.
If a user has had an awful experience, you could also enlist a person to respond to that user and help improve the user’s experience, which could also potentially enable you to improve your chatbot experience.
7. Leverage the Data from Chatbot Conversations
All brands should leverage user data, which is a powerful way of creating a great user experience. Not only should you gather information about how your chatbot is doing and the quality of the digital customer experience that you’re providing; your chatbot can also easily obtain additional useful customer information.
You can use this customer information throughout your communications with the user to make these interactions better. Plus, you can use this information afterward, establishing your chatbot’s personal connection to the user. For example, your user’s name and location are prime information that your chatbot can utilize in tailoring interactions with the customer.
Subsequently, you can leverage this information in compiling demographics on the customer base that is using your chatbot. Employing this user data can make your users’ interactions with your chatbot more personal. The chatbot can greet the user by name or even take on a different persona based on the user’s location, gender, and any other information the user has provided.
During each conversation with a user, you’ll be able to capture substantial data. This is your foundation for chatbot analytics.
8. Choose the Right Channel for Your Chatbot
You have a number of options when deciding what channel you should use for your chatbot. There are advantages and disadvantages to each channel, so the choice can be tough. These options include the following:
Messenger—Offers a large reach, a variety of interactions, easy access, and tools for complex conversations.
Telegram—Provides a great user experience, simple communication, and access using buttons instead of typing.
Slack—This tool has many uses, can monitor the channel’s activity, and supports interactive messages.
Before you begin developing your chatbot’s user experience, you must decide what purpose it needs to serve, then go from there. The features you need could also determine what channel you should choose, so review your list of required features and compare it to what each channel has to offer for your chatbot.
Don’t forget to consider your target audience and how they prefer to communicate. There’s no point in choosing a channel that your customer base would never use.
9. Test Your Chatbot Before Unleashing It on the Public
Even though you might personally love your chatbot UX design, it is essential that you test it with a small audience to gauge its effectiveness, as depicted in Figure 5. Only in this way can you discover any errors in your design and implementation so you can make your chatbot better.
Doing live chatbot testing with your customers is important so you can eliminate as many bugs as possible prior to letting the general public interact with your chatbot. You don’t want your chatbot to offend, bore, or frustrate your customers, risking the loss of customers or potential customers. To achieve the greatest success possible, you must deal with such problems before they happen.
Testing comes in a variety of forms—from automated solutions to human-machine question-and-answer sessions—all of which are relevant and important. During testing, you should look for the following:
variability in chatbot conversations
chatbots saying the right things in the right way
chatbots’ effectively initiating conversations
chatbots’ giving valid responses
Your biggest concern when building a chatbot is ensuring your customers have a great experience. If you can avoid making these very common UX design mistakes, you’ll be well on your way to designing a very successful chatbot.
A chatbot should always be versatile and able to play various roles in interacting with users. However, appearing too human can have its drawbacks and deter users from engaging. For example, the overuse of emojis or colloquial expressions could signal to users that they aren’t talking to a real person. Design your chatbot to meet the needs of and appeal to your target audience—in terms of its icon or avatar, vocabulary, tone, and general personality.
Avoid trying to achieve too many of your expectations for your chatbot in one go. Instead, you should keep adding features gradually, according to the feedback you gather from users.
It is crucial that your chatbot provide a meaningful introduction to any conversation. Your chatbot cannot just say Hi, then keep asking the customer for instructions. You must design the user experience in such a way that the chatbot sets the context at the beginning of the conversation with the user, but without making that initial message too lengthy.
A seasoned content writer, James has ten years of writing experience, with expertise in blogging and writing creative and technical copy for direct response markets, as well as B2B and B2C industries. A native New Yorker, James holds a bachelor’s degree in English Literature. He has worked for industries such as information technology (IT), software-product design and development, lifestyle brands, and has provided some great insights on UX design, mobile-app development, and ecommerce. Beyond his work in technology, he is very connected to digital media in his free time and loves to binge-watch Netflix. Read More