The empathize stage is where design thinking earns its reputation. It is the stage where teams stop guessing about their users and start listening to them. It is also the stage where facilitators and teams now ask us the most questions about AI. Can AI help you understand people? Should it? And where does it get in the way?
After weaving AI into our own design thinking work with AI across many client workshops, our answer is that AI can strengthen empathy work considerably, as long as it supports human listening rather than replacing it. This guide covers what that looks like in practice: where AI helps, which tools fit the empathize phase, and the boundaries that keep the work human-centered.
How AI Enhances Empathy in Design Thinking
Empathy work produces a volume problem. A single round of user interviews can generate hours of recordings, pages of notes, and hundreds of sticky notes. The insight is in there, but teams often run out of time or energy before they find it. Traditionally, synthesis is where empathy research goes to die.
This is exactly where AI helps most. AI is strong at processing large amounts of qualitative input quickly: transcribing interviews, clustering observations into themes, surfacing patterns across many conversations, and drafting summaries a team can react to. When AI handles the heavy processing, people get more time for the parts only humans can do, like sitting with a user, noticing what they did not say, and feeling the weight of their frustration firsthand.
In one of our international workshops, participants shared stories and responses in Spanish through Slido while we facilitated. AI synthesized hundreds of those inputs into themes in minutes, in the room, and the group spent its energy discussing what the themes meant rather than sorting sticky notes. The empathy came from the participants and their stories. The AI just made all of their voices visible at once.
AI Tools for the Empathize Phase
You do not need a specialized research platform to start. Here is how common AI tools map to empathize-stage tasks:
- Interview transcription and notes. Tools like Otter, Zoom AI Companion, or built-in meeting assistants turn conversations into searchable text so researchers can stay present instead of scribbling.
- Synthesis and theme clustering. ChatGPT, Claude, and Gemini can take raw interview notes or open-ended survey responses and cluster them into candidate themes, tensions, and surprising quotes for the team to verify.
- Empathy map drafting. AI can produce a first-pass empathy map from interview transcripts, which the team then corrects and enriches. Our empathy map template pairs well with this approach: let AI draft, let humans judge.
- Question preparation. Before interviews, AI can pressure-test your question guide, suggest follow-up probes, and role-play a skeptical user so your team practices listening before the real conversation.
- Whiteboard AI features. Miro and Mural now include AI clustering that groups digital sticky notes by theme, useful for live synthesis in virtual sessions.
A note on so-called synthetic users: some tools now offer AI personas you can “interview” instead of real people. We treat these as rehearsal or source of inspiration, never as real research. An AI persona can help a team practice interviewing or explore initial assumptions, but it reflects training data, not your actual users. Decisions should rest on real conversations.
Using AI to Synthesize Stories, Not Replace Them
Stories are the raw material of empathy. In our workshops, we teach teams to gather user stories during the empathize stage because a story carries context, emotion, and specifics that a survey score never will. AI changes what you can do with those stories once you have them.
Feed a set of user stories into an AI tool and ask it to identify the recurring struggle, the moments of workaround, and the language users repeat. Then bring those findings back to the team as questions rather than conclusions: does this match what we heard? What did the AI miss? What surprised us in the room that never made it into the notes? This keeps the team in the role of interpreter and keeps AI in the role of assistant.
What AI Cannot Do in the Empathize Stage
AI has no lived experience. It cannot sit across from a frustrated user, notice their hesitation before answering, or build the trust that makes someone share what they actually think. It also carries the biases of its training data, which means an AI summary can quietly flatten the outlier voice that matters most.
Three boundaries keep the work honest. First, AI summarizes conversations; it does not have them. Real user contact stays non-negotiable. Second, humans verify every AI-generated theme against the source material before it shapes a decision. Third, protect privacy: get consent before recording, and strip identifying details before putting user data into any AI tool your organization has not approved.
Bringing AI Empathy Work Into a Workshop
If you facilitate design thinking sessions, a simple sequence works well. Have participants conduct short empathy interviews with each other or with invited users. Collect the notes digitally. Run live AI synthesis in front of the group and let participants challenge and refine the themes together. The moment a room full of people sees their collective input organized into themes within minutes tends to shift how they think about both empathy research and AI at the same time.
Deeper training on the empathize stage itself, including interview technique and observation methods, is covered in our design thinking empathy training guide.
If you want your team to experience this firsthand, we design and facilitate design thinking workshops that integrate AI at every stage while keeping people at the center. Contact us to talk about a session for your organization.