Nowadays, voice user interfaces (VUIs, pronounced voo-ee) are all around us, from our phones and smartwatches, to our computers, laptops, TVs, and cars. Voice is changing the way that we interact with our devices. Think about the last time you asked Siri, Alexa, or Google for directions or asked a general knowledge question.

Voice presents a great opportunity for the learning and development (L&D) community. Creating a voice user interface for learning presents an opportunity to engage learners and allow them to access [pull] learning content and assistance into their workflow.

We need to start thinking of how to integrate the use of voice into our learning solutions. And, we need to learn how to design for these interfaces to create great learner experiences.


How learners interact with a VUI is a different experience from how they interact with a graphical user interface (GUI, pronounced goo-ee). A GUI presents the learner with visual elements like text, buttons, image items, and links used to interact with a learning solution. For example, when you create eLearning using Articulate Storyline, you are using and creating a GUI.

With a VUI, learners only use their voice to interact with the learning solution. This means that the design guidelines that you need to follow for a VUI are very different from those for a GUI.

In voice-based interactions, we need to model the experience the way that learners interact and expect to interact. We need to design for how people engage in conversation—how we talk. We need to incorporate conversation prompts that are conversational and engaging.

Getting started with VUI design

The number-one question I always get about VUI is, “Where do I start?” Here are my top 10 tips to getting started with designing a VUI for eLearning:

  1. Define the problem you are going to solve and decide if VUI is the right tool. Generally, training is developed based on competitive edge, compliance, and general assumptions about the target audience. It’s important to begin by identifying the one problem you are trying to solve and then deciding if voice is right for the job. You also need to identify one or two key performance indicators (KPIs). KPIs are quantifiable measurements that tells you how an individual, team, organization, or company is performing.
  2. Identify who will use your VUI solution. Create personas for your learners. Use existing data to create a general picture of your target learners. Be sure to share how and when they will interact with your VUI. Will they access your VUI using a voice-controlled device, a smartphone, or a laptop/computer? Identify the tasks that a typical learner is more likely to access, along with frequency.
  3. Define your VUI persona. Here, you’ll want to decide what the tone and voice your VUI will be. Formal? Informal? Will you use social slang, like “LOL,” “brb,” or “ttyl”? Or more general slang, like “hey,” “y’all,” or “salty”?
  4. Plan/script out the dialogue. Start with a happy path—the best-case scenario for how a learner would interact without deviation (the straightest path). Keep it simple. Once you have your happy path, you can define deviations, such as how a learner might word a specific question.
  5. Test your dialogue. Grab a teammate and ask them to role-play with you. You take the role of the VUI, and your partner plays the role of the user. Make sure the dialogue has a natural flow and matches your selected tone. Take notes on the specific ways your partner forms their questions and their expectations.
  6. Create your prototype. Transfer your dialogue to a VUI development platform, such as Google Codelabs or Amazon Blueprints. [Warning: you may find additional challenges with your dialogue that may require you to repeat steps 4 & 5]
  7. Test your prototype. Identify a focus group that can test your prototype and provide feedback.
  8. Gather feedback on your prototype. Make sure to collect all the feedback from step 7, analyze it, and use it to make improvements.
  9. Repeat steps 7 & 8. Testing and gathering feedback to make improvements is not a one-time deal; repeat these steps at least one more time.
  10. Pilot your VUI. Pilots allow you to gather additional data and ensure that you are aligning to the KPIs identified in step 1 before launching to a larger group.

Set learner expectations

Research has shown that users tend to relate to communicating with a VUI the same way they relate to communicating with other people. The expectations for the interaction are high, and it’s important to create an interaction that offers a natural conversational flow. You must set the learner’s expectations when you first introduce a VUI. This includes sharing the purpose of the skill and calling out what the skill can and can’t do.

Word choices present additional challenges, with subtle changes in meaning, tone, and context adding a large degree of complexity to the design of conversational scripts. I always recommend that before you begin designing your VUI for voice integration, you take time to define the personas that you are designing for. Also define your VUI’s tone (mood) and voice (personality). This will allow you to put yourself in your learner’s shoes and design for them—not for yourself.

Finally, as the 10 steps above state, KPIs, testing, and user feedback will help you understand how learners may interact with your VUI. This allows you to make data-driven design decisions that align to your defined KPIs and create a better learning experience.

Use VUI in microlearning

Explore the use of VUIs in microlearning experiences and performance support. Myra Roldan is presenting “Voice Technologies: Microlearning Beyond the Status Quo” at The eLearning Guild’s Microlearning Design Online Conference, September 18–19, 2019. Other speakers will dig into short simulations, improving microlearning videos, using podcasts as microlearning—and much more. Register today—or add an Online Conference Subscription to your membership and enjoy a full year of online learning with industry leaders.