Outset has added the ability to conduct usability testing interviews to a platform that makes use of generative artificial intelligence (AI) to conduct research.
Rather than requiring a human researcher to conduct interviews, the Outset platform presents a video experience through which generative AI is used to engage in conversations with end users. An AI moderator provides an overview and then enables screen sharing via a set of voice and text prompts. As a participant navigates through the product experience on video, the AI moderator follows their movements and asks real-time contextual questions about their experience. It also listens to voice responses from the participant and then asks follow-up questions during an interview session that typically lasts anywhere from 10 to 60 minutes.
That approach has already enabled organizations such as Weight Watchers to conduct many more qualitative surveys at a significantly lower total cost, says Outset CEO Aaron Cannon.
Traditional qualitative interviews have low sample size, are very expensive and are notoriously slow to conduct, noted Cannon. Teams often spend months interviewing only a handful of individuals. That small sample size makes it difficult to infer, for example, what the future direction of a product strategy should take or why a specific feature or capability isn’t being as widely adopted as anticipated, he says.
In contrast, AI-moderated research can simultaneously conduct hundreds of in-depth usability interviews in a matter of hours, and once completed automatically summarize those results, notes Cannon.
Additionally, researchers will also find many individuals are likely to share more candid feedback with an AI platform than they might be with another human, simply because they will be less concerned about being judged in some fashion, adds Cannon. “There is no judgement bias,” says Cannon.
Researchers still need to prepare a usability testing guide, but the survey process can be automated using either a panel of customers they provide or potential customers that Outset has selected.
It’s not clear to what degree generative AI might make more people inclined to participate in research panels, but the more streamlined the process becomes the more likely it is people may be willing to give up some of their time to help improve a product or service. The challenge, as always, will be turning all that additional feedback into actionable insights. In theory at least, generative AI coupled with any number of other analytics applications should make it easier to surface those insights.
In the meantime, hopefully more qualitative data will lead to major improvements to products and services that might, for example, become more customizable. The one thing that is certain is that it’s a lot easier to sell additional products and services to happy customers than it is to sell anything to a brand-new customer. The issue, of course, is determining which features and capabilities should be prioritized in a way that benefits the largest number of customers possible.
