Multimodal User Experience Design Best Practices

Multimodal User Experience Design Best Practices

3 min read

Alexa, Siri, Google Assistant—virtual assistants have transformed how we interact with technology and what we expect from it. Today’s most popular assistants offer a multimodal AI experience — the ability to interact on multiple channels — and are a major leap from the voice-only assistants that dominated the market just a few years ago. Unsurprisingly, sales of these multimodal assistants grew 30% in 2020, while voice-only assistants saw a 3% decline.

The same holds true in the contact center. Multimodal Virtual Agents allow customers to tap, text and talk to get more done, with little effort.

Offering variety can help create more successful outcomes for customers with a lower AHT for agents — a win on both sides of the phone line. It’s also the wave of the future. According to Gartner, by 2023, customers will prefer to begin 70% of self-service requests with conversational voice interactions. By 2025, multimodal interactions will be a standard feature for virtual assistants, up from less than 2% in 2021.

If you’re one of the growing number of contact centers looking to implement a multimodal virtual agent, here are some best practices to get you started.

Multimodal Virtual Agent Implementation and Best Practices

Keep menus short

Yes, menus. They have their place even in the world of conversational AI. Your menu prompts should make users feel they’re getting where they want to go (i.e., from Point A to Point B) as quickly as possible. Typically, that will mean limiting the number of options to about three or four. This should offer enough choices for users without overwhelming them.

Keep language short and specific

Just as you should keep your menus brief, try to limit the actual length of each prompt. When creating a script, avoid lengthy descriptions and repetitive words, as this can obstruct the meaning of each option and lose a user’s attention.

Give callers a way out

No matter what issue they’re dealing with — from complex queries to basic issues that are best handled with self-serve options — customers want to feel like they’re being personally cared for. An option to speak to an agent is essential, even if you’re trying to incentivize other avenues

Use active voice

Solving customer challenges is all about taking action to fix a problem. Even small tweaks in your system, like avoiding passive voice, can help users feel like they’re making progress and create a more positive experience. Phrases like “have been” unnecessarily extend sentences and can leave a customer feeling confused or disengaged.

Design for the distracted caller

Whether calling or texting for support, customers typically aren’t looking to linger on the line. Remember that callers probably aren’t listening to every word, so you’ll need to design scripts and call flows that reflect this. In other words, “design for the ADHD mindset.”

Personalize where possible

Even if you plan on using a multimodal virtual agent to limit the number of calls an agent needs to take, offering an experience that feels human and personalized is still paramount. Some key ways to avoid a cookie-cutter experience include using conversational AI that takes advantage of existing customer data. Other multimodal tools, like document upload, can help solve problems and increase customers’ ability to interact with a tech solution.

Callers hear the first and last options

Similar to best practice number five, it’s important that you don’t make assumptions about what a caller has heard. In fact, evidence suggests that callers often only fully comprehend the first and last options on a menu. Arrange the order of options to reflect this — with the most common or general queries placed at the bookends of your menu.

For speech, only ask questions where you know the answer

While open speech is a great thing, it does have its limits. Not having an automated response will result in extra confusion and slow the flow of the call. Avoid open-ended questions that give callers an opportunity to stray far from your planned-out paths.

Table of Contents