Build Your Chatbot Integration Strategy

Best practices in UX.

What good are chatbots? Knowing how and why you’re creating a chatbot is the key to making sure it’s useful. Build a sound integration strategy with these UX best practices.

Businesses everywhere are eyeballing chatbots as a way to build a better customer experience, increasing both engagement and sales in big ways.

Sephora is expanding its service offerings with bots. A bridesmaid in need of a makeover can use the store’s text-bot reservation assistant to book an appointment at a location nearby — or she can access Sephora’s virtual artist to virtually try on products and get color recommendations to match her bridesmaid’s dress. She can even use it for makeup tutorials — like how to create a no-makeup look for the wedding in the morning and nighttime smoky eyes for the reception later in the day.

Chatbots are fairly new, yes, but conversational chatbots already run the gamut. Chatbots can help you get your morning caffeine (Starbucks), book a trip to Tahiti (Kayak for Facebook Messenger) and buy the hottest new scarf at Fashion Week using a “see now, buy now” bot (Burberry).

And because they’re chat-focused, they’re an ideal mechanism for mobile devices, making it easier to find a lost package, schedule a home inspection, or even get emotional support from a peer.

Plus, chatbots are also inexpensive to deploy and easy to scale, which is why they’re quickly gaining traction.

So, if your company isn’t yet using chatbots to engage in immediate and valuable conversations with your visitors, it really should be. And as a UX designer, you should be ready to drive the conversation.

To get started, you’ll need a solid approach. Let’s take a look at three UX design best practices to help you get there.

1. Establish a clear role for your chatbot in the overall customer experience

One of the first questions you should answer is how you can use chatbots to make something easier for your audience.

It’s also the most important thing to get right, so it’s worth repeating. Chatbots should both reduce friction AND fit naturally into their experience. If your chatbot is clunky to use or feels disconnected from the rest of their journey, you’ll only annoy users.

Here are a couple of things to think about. Most users already want to use messaging to connect with brands, says Twilio research. And 38 percent of them have a positive perception of chatbots, says one Business Insider study.

Good chatbot design and a consistent, helpful interaction will make sure that number grows.

So just like any tool or website, your bot has to be designed to be tantalizingly useful.

To get there, you need to consider some key usability factors. Things like nailing down its interaction capabilities and limits. Creating simple, clear workflows and responses. Defining how it works with apps and uses APIs, for starters.

Here at Adobe, we’re exploring a number of ways these “intelligent agents” might help creative and marketing teams do their jobs.

That includes finding ways to help designers create faster through the use of command shortcuts and contextual design, giving users the tutorials they need based on how they’re using our solutions, and proactively alerting users to new data available for use in their marketing campaigns, among others.

Our focus is on finding the best use of natural language processing and other bot capabilities to make people’s jobs a little easier.

2. Ask the right questions up front. Then, test

Once you know what your chatbot is going to do, it’s time to build a prototype. Take a look at this list of UX design considerations as you get planning:

• The medium. Will your chatbot be text only, text with artificial intelligence audio, email, and will it be deployed on Facebook Messenger, Siri, Alexa, or something else? Are you serving up the chatbot on mobile, desktop, or both? How do experiences differ on each platform?
• What it looks like. Do you use avatars, images of humans, or neither?
• Will you have a purely digital interface?
• When to use human or machines. Will you openly present your bot as the AI engine it is, or can you mimic human responses more closely? What’s the right balance of empathy, humor, and problem-solving as part of that?
• Your brand. Are you building the intelligence engine and loading content that locks to your brand and aligns with your core values?
• The timing. When in the customer journey should the chatbot appear? Why then? When does a human step in?
• The best way to accomplish a requested task. When is a text reply better than an audio one, for example? And what content will users receive, and how will it be retuned for various requests over time?
• Transitions and escalations. When you trigger sales, customer service, or another real person, how can you make the handoff smooth?
• The tools. What chatbot tool, if any, gives you what you need out of the box? Is your team willing and/or able to build a custom chatbot?

Once you have a prototype of your chatbot and have a working information flow and set of smart responses, deploy another long-held UX best practice — rigorous testing.

3. Consider the role of intelligence

If your project is the team’s first dip into the AI pond, a simple chatbot or set of chatbots may be a good option for you.

You might want to get your feet wet by building a decision tree bot that directs users to content, for example, before you revamp a complex purchase or contextual customer support process.

Or, you could test out an internal chatbot before you build one for customers. In that case, a simple bot that helps with conference room scheduling, guest check-in, or another small productivity challenge can help you build a comfort level before you wade in deep.

Keeping it simple also means you’ll need to put the time in when you do decide to go big. Being deliberate is one short path to creating excellent design, scripts, and information flow — and getting the use of natural language processing and AI right.

As you define the role of chatbot, you’ll also want to find the key stakeholders and right management team to own the effort, and settle on a governing process.

Another thing to consider, says Deloitte, is the tracking. “An enterprise-ready chatbot platform should include monitoring capabilities to track the different command and responses of its users and consumer applications.”

And finally, it may be worth bringing in an experienced consultant to guide you through your first project. Consider it the life vest — or at least the arm floaties — for your first plunge into the technology.

Good chatbot implementations drive perception of UX as drivers of innovation** **

As natural language processing, AI and AR continue to mature, the future of the chat bot looks pristine and shiny, like a lake at sunrise.

UX designers able to take the lead in finding the right uses for chatbots in their enterprise and apply time-proven best practices to their approach will position themselves as strategic drivers of innovation within their organizations.

This article series will explore best practices in future-proofing your UX skillset, uncovering the tools and techniques that will be critical for all UX designers — those who are just entering the job market AND those who will need to modernize their skills to stay relevant in the customer experience wave.