Listen on Spotify, Apple Podcasts, and YouTube
Sarah Burton is a Lead Product Designer at Dovetail, a customer insights hub that helps companies better understand their customers so they can build better products and services for them. They just launched Dovetail 3.0, which includes a series of AI features that help you analyze and extract insights from customer surveys, interviews, and app store reviews.
In my conversation with Sarah, we discuss:
🤖 How to balance AI automation and human expertise
🔍 The importance of understanding customer problems before jumping too quickly on the AI train
🧠 Design tactics to help you deal with AI’s unpredictable nature
Chapters
00:00 Introduction to Dovetail
02:51 The Next Generation of Customer Analysis and Insights
09:52 The Role of AI in Enhancing Customer Insights
13:03 Human-Centric AI Integration Principles
19:59 Three Levels of AI Automations
24:56 Understanding Customer Needs and AI Opportunities
27:25 Integrating Design and Engineering for AI Solutions
29:21 Designing for Unpredictability in AI
36:52 Designing for Trust and Transparency in AI
43:02 Navigating AI Pressure in Design
My main takeaways from this conversation:
Don’t start from AI and retrofit problems to it. Start from your customer problems and then see if AI is a good solution.
AI features don’t need to be binary (on or off) because outputs can still be useful even if they’re not perfect.
Design AI systems with a “human-in-the-loop” mindset. People want to control and collaborate with AI, instead of automating every single task.
Make AI solve problems and tasks that people hate to do so they can focus on what they love to do.
Don’t expect your customers to be prompt engineers. Design an interface that helps them get their task done effectively and do the hard work behind the scenes.
AI is enabling an advanced level of understanding and analyzing customer sentiment by making it easy to ask a question across a large body of qualitative data.
If you’re integrating AI in your product, it needs to convey trust and provide transparency about its outputs.
How to balance AI automation and human expertise