- Chatbots are programmed to answer customer questions and resolve customer issues quickly and accurately via AI and machine learning.
- Good chatbots can improve UX by making customer service faster, more personalized, and available 24/7.
- Bad chatbots can actually damage UX, so it’s important to invest heavily in the design and testing processes before launching your bot.
Think back to the last time you had a problem with an item you’d ordered online.
Maybe the item was delivered late, or was defective, or didn’t fit right. Chances are, when you logged onto the shop’s site to resolve the issue, you spoke with a chatbot. And if recent statistics about the efficacy of chatbots are true, it was probably a positive experience.
This is just one of a thousand use cases for chatbots in the customer service space. Conversational intelligence and machine learning capabilities have enabled chatbots to do everything from answering customer questions to facilitating purchases much more quickly than live agents can, with just as much or even more accuracy.
In fact, Hubspot found that 63% of people surveyed don’t realize they’re already using AI conversational tech. That means chatbots are doing their jobs well. So how can you capitalize on the trend to give users a better experience of your site? Keep reading to learn more.
What chatbots do
Chatbots can be thought of as AI-powered customer service agents. A typical chatbot is programmed to answer live customer questions quickly and accurately, either by providing information or resources, or inciting an action (such as a purchase or bill payment).
Chatbots may respond to multiple questions in the same way or adapt answers to particular keywords in a customer query, but virtually all chatbots rely on machine learning to improve the accuracy and fluency of their communication over time.
Finally, chatbots can exist on a range of different platforms. While many companies integrate their sites with social media-powered chatbots, such as those powered by Facebook Messenger, or SMS text based chatbots, this blog will focus on chatbots that live right on your website.
How chatbots improve UX
The benefits of chatbots from a business perspective are obvious: well-designed chatbots are cheaper than live agents and can resolve exponentially more issues than humans in the same amount of time. Plus, they don’t need to sleep or take breaks, making customer service available 24/7.
But the benefits of chatbots are also becoming clear to users themselves. UX designer, mentor, and writer Mallory Kim outlines the following central principles of UX:
- Confirmation and feedback
- Data-driven foundation
When done right, chatbots help achieve virtually every principle on that list. They center the user by making customer service personalized and available 24/7. They simplify customer service by enabling users to achieve several steps – e.g., get a question answered, research products, and make a purchase – all in one place. Being based on centralized AI, they’re also more consistent than a staff of live agents with different communication styles, skill levels, and schedules. The list goes on.
However, not all chatbots are created equal. If the technology powering the bot isn’t up to snuff, the results can backfire, creating an even worse UX experience than with live agents. Think: chatbots that haven’t been run through enough machine learning scenarios and keep repeating the same lines to different questions. The last thing you want is for your customers to be wishing they could simply talk to a real person.
To avoid this, follow the tips and best practices below to make sure your chatbot improves rather than damages your customers’ experience.
Tips and best practices for using a chatbot on your site
1. Tailor your chatbot to specific use cases.
One of the most important first steps of implementing a chatbot for better UX is to carefully consider your target use cases. Citing a 2018 report on chatbots from Drift, Sprout Social highlights the following common use cases for chatbots from most to least prevalent:
- Getting a quick answer in an emergency
- Resolving a complaint or problem
- Getting detailed answers or explanations
- Getting connected with a human agent
- Making a reservation (e.g., a restaurant or hotel)
- Paying a bill
- Buying a basic item
- Getting ideas and inspiration for purchases
- Adding oneself to a mailing list or news service
- Communicating with multiple brands using one program
- Buying an expensive item
You’ll need to determine what your chatbot will be primarily used for so that you can tailor the bot to that use case. As reflected in the list above, the most popular use cases for chatbots tend to be getting answers to questions, resolving problems, and performing simple actions like reservations and payments.
Knowing exactly what you’ll need the bot to do will help you determine the best course of action for your design process.
2. Decide who will design the chatbot.
In their overview of the future of chatbots, Hubspot highlights the difference that good design makes in a customer’s experience of a chatbot. In other words, bad bots can actually damage UX, so you’ll need to make sure you don’t cut corners on the design process. That means making sure you have the time and resources to invest in careful design and testing before the bot is implemented.
Large companies may have the internal resources to design and test the bot in-house. However, many businesses will need to outsource the design process. A good chatbot is a serious investment, so do your due diligence when researching potential design firms to make sure the investment is worthwhile.
3. Give the chatbot a personality.
Even if your customers know they’re talking to a bot, that doesn’t mean the conversation has to feel impersonal. On the contrary, AI and machine learning capabilities have advanced to the point that conversations with chatbots can feel almost identical to conversations with live agents.
One way to achieve this is to give your chatbot a personality. As Mallory Kim notes in her 12 UX principles mentioned earlier in this blog, users seek a sense of familiarity and helpfulness. Giving your chatbot a name, an interesting character (like Hipmunk’s cute Chipmunk avatar), and unique voice can make the conversation feel more “real.” Likewise, programming the bot to use customers’ names and customer data can create more personal and helpful responses.
4. Integrate chatbot programming with customer data.
Speaking of making your chatbot interactions more personalized, another best practice is to utilize customer data in the chatbot’s programming. Imagine a chatbot that can quickly comb the user’s purchase history to make personalized recommendations for future purchases or to suggest newsletters or resources the user might be interested in.
Likewise, you can use data from chatbot conversations to improve your other marketing practices, such as email marketing efforts. Data gleaned from chatbot conversations provides vital information about your customers’ needs and interests that can help you target your marketing efforts with more precision.
5. Test, test, test.
We can’t overstate the importance of investing heavily in the design process of your chatbot – and a crucial part of that process is testing.
Before launching the chatbot, submit it to rigorous testing or ensure that the company you’re outsourcing to does so. This ensures the chatbot actually does what you need it to do, and gives you ample opportunity to fix bugs and improve programming before the bot goes live.
UX meets digital marketing
As SEO geeks, we’re passionate about centering the experience of the user in all of our digital marketing efforts. If you have questions about how to improve UX through better digital marketing, reach out to the team today. We’re always happy to chat!