Raise your hand if you’ve ever relied on a chat feature for customer service help on a website. Maybe it was asking for help finding something on a website, or maybe you were trying to order a product online, or maybe you were trying to make dinner plans with friends via Facebook Messenger.
Chatbots are on the rise on the web and they’re being met positively by users around the globe. Here’s what you should know about chatbot UX in the evolution of website strategy and planning.
What are Chatbots?
Chatbots are an artificial intelligence (AI) feature that connects your customer or user with a “conversational” computer program to answer questions or carry on conversations.
Thanks to the rising use of text and online messaging like Facebook Messenger and WhatsApp (among many others), as well as the increase in voice assistants like Siri and Alexa, users have become more comfortable with “chatting” to their brands to get the answers they need. According to a 2016 Nielsen survey, 53% of people are more likely to shop with a business they can message directly, and 56% of people would rather message than call customer service.
As a response, companies are integrating chatbot technology into their website UX. Gartner estimates that by 2020, nearly 50 percent of medium and large sized corporations will incorporate chatbots to their digital strategies, and 85 percent of online interactions between consumers and brands will be non-human, relying primarily on chatbots.
The Benefits of Chatbots
As the need for immediacy drives much of what we do online — from ordering a pizza to checking on our local weather for the day — chatbots serve a similar purpose. The benefits of chatbots have already begun to reveal themselves in:
- Easy connection to consumers. For users who just want a quick answer to a question or a fast transaction, chatbots can be a perfect fit with little error. And you still have opportunities to drive the next step and connect with your visitors by phone or in-person, if the need arises.
- Saves money. Chatbots are relatively low-risk and easy to implement, and can help streamline your customer service options to your web visitors.
- Available 24/7. Even if your business is open 9 to 5, your chatbot can be online all day, assisting with basic functions and needs of your late afternoon and evening consumers.
- Drives conversion (and sales). Simple commands or requests to purchase (or re-purchase) products from your business can be done quickly and efficiently from a chatbot.
People like to interact with brands and businesses they use, especially when it’s convenient for them. Chatbots often meet users where they are, whether at home after a late-night shift, or at the office for a quick response.
Approaches to Chatbots
As handy and efficient as a chatbot might be, strategy is still key when choosing how (or if) to implement one for your organization. Humans are still complex creatures with a nearly unlimited, ever-evolving vocabulary. Deciding the method of interaction you want your chatbot to carry is integral: If you have a product or service base that can be answered with simple responses, or that only requires your program to point customers in the right direction, you won’t need a heavy investment. But a chatbot that can complete more difficult tasks or requests may need more thought and strategy for your organization.
Chatbots Magazine recently highlighted four levels of chatbot interactions that businesses should review when considering to implement a chatbot:
- Immediate connection to important, frequent, recent, or popular information – Similar to a great site navigation, or a well-organized homepage, this format of chatbot would make content easy to find for users. The fewer clicks to complete a task, the better.
- An intuitive, deep search chatbot (similar to Apple’s Spotlight) – A quick-launch way to get an answer returned from a repository of information.
- A chatbot that understands complex sentences and unstructured data – The more your chatbot can understand the more advanced tasks it can do, from returning more accurate information to completing transactions.
- An immersive video chat with a real person, powered by AI – Though more advanced, this would connect a user with a real person, who relies on AI to augment and support responses.
HubSpot defines chatbots in two distinct categories: Informational bots, which give you information when you ask for it (e.g. “What’s the current weather?”) and utility bots, that help solve a problem (e.g. “Reorder a phone charger for a Samsung S8.”)
Currently, many businesses are at least using one or two of these options to help improve their customer experience. Research from Forrester reveals that 5% of companies were using chatbots in 2016, but 32% were planning on using or testing them by 2017. Now we’re here in 2018, and those numbers are surely on the rise.
Starting Your Chatbot Strategy
When considering a chatbot for your organization, consider starting with an analysis of what types of questions your customer service or website get from everyday users and create scenarios of answers that a chatbot could provide. This can be considered a “conversational flowchart.”
Hubspot created an Inbound Messaging Framework as a starting point for beginners to help structure chatbot conversations that keep customer experience in mind. This framework has four distinct pillars: Connect, understand, refine, and deliver.
Know Your Users First
Understanding your users, whether through usability testing on your website or through conversations with focus groups, is a great way to find out if chatbots are a good fit for your brand at all.
Some apps, like WhatsApp or Facebook Messenger, might be a great fit for certain demographics of consumers — and even play nicely with emojis, just like people do — but an on-website chatbot might be a better fit for a less social networked audience who is coming to your brand’s homepage for a reason.
How Far Will Your Bot Go?
Once you’ve decided to move forward with implementing a chatbot on your website or through a messaging app, you’ll need to decide how far your chatbot will go with helping your users.
Will your users expect to complete a transaction without being sent to a call center? Will your chatbot eventually move into personalization where they remember customers and their preferences? Is your chatbot just a reliable associate on the sidelines, available to help your visitors if they’re stuck?
Determining the purpose — either informational or utility — of your chatbot will be essential to strategizing the flow of conversation you expect.
Create a Flowchart of Conversation
Once you’ve decided what chatbot you want to implement and the audience you imagine using it, it’s time to develop a conversation flow, as Hubspot suggests, to determine the type of questions you can expect and the responses your “bot” will return.
Of course, make sure you have other methods for customers or visitors to reach you, if the chatbot doesn’t meet their needs. A phone number, online form, or e-mail address is a surefire back-up plan to make sure your users have a route to reach you.
Build Your Chatbot
Whether you work with an in-house developer, agency, or you purchase a widget that lets you implement the chatbot settings of your choice, building your own chatbot will be step in line with where the future of website UX is heading.
There are plenty of resources to help guide your chatbot creation, including Chatbots Magazine’s recommendation for Motion AI, recently acquired by HubSpot. SEO expert and Moz contributor Robin Lord even shared his code for building your own chatbot.
Chatbots Are Here to Help
But remember that even though the word “bot” has a tendency to through marketers into a tizzy, chatbots are still meant for human interaction and to help drive brand awareness and conversion to
At the end of the day, chatbots are relied on by users to solve a problem, from providing medical advice to ordering a pizza to checking the weather. Chatbots are just another arm to reach your audience.
For more information on performing chatbots and how to strategize your brand’s approach to implementing them, try these helpful articles: