AI is changing what it means to be a researcher. Even if you don't think of yourself as one, everyone does research. We’ve all googled something, asked an LLM a question, or dug into a topic to make a better decision. But when you're building new products and services, research goes deeper than knowledge-gathering. It means understanding the people you're designing for and their needs, challenges, and motivations. Often, these are things they can't even articulate themselves.

That kind of human-centered research has been central to IDEO's work for over 50 years. And our approach to AI is the same as our approach to any new research tool: curious, critical, and human-centered first. We're interested in where it genuinely makes the work better and honest about where it doesn't.

Hannah Rosenfeld has been practicing human-centered design for over a decade and has spent nearly nine years at IDEO, where she leads complex design programs with a focus on healthcare. Through that work, she's developed a clear perspective on where AI in research actually earns its place and a framework to help others figure out the same.

In this episode of the Creative Confidence Podcast, Hannah joins host Mina Seetharaman to walk through that framework, which she calls AI’s affordances. It’s a practical guide to integrating AI into qualitative research, helping teams move beyond speed to uncover insights and do things they couldn’t do before. 

Listen on Apple, Spotify, YouTube, or wherever you get your podcasts.

Article Summary

AI Doesn't Change the Principles of Good Research—It Expands The Definition

A Framework for Using AI in Research: The Intersection of Speed and Scale

Mode 1: Widen Your Aperture—Use AI to Take In More

Mode 2: Notice Nuance—Use AI to See Deeper, Not Just Wider

Mode 3: Challenge Your Assumptions—Use AI to Interrogate Your Thinking

Mode 4: Immerse Yourself in Data—Use AI to Bring Your Thinking to Life

The Skills AI Can't Replace: Why Human Judgment Matters More, Not Less

AI Doesn't Change the Principles of Human-Centered  Research—It Expands The Definition

"The philosophies and principles of design research aren't changing," Hannah clarified, "but the speed and scale of AI do allow us to ask bigger questions, to take in more context, to engage with complexity in wholly new ways."

Most of the AI-in-research conversation focuses on efficiency and doing the same work faster. But that framing undersells what's actually available. Hannah is intrigued by a more interesting question: What can we do now with AI that we couldn't do before?

The answer, in Hannah's view, is that AI shifts the role of the researcher by expanding the conditions under which human judgment can be exercised. "If we use it intentionally, AI can help us see more, go deeper, and think differently."

Using AI intentionally is both an operational question (where in the process does it fit?) and a philosophical one (what is appropriate for AI to do versus a human?). 

"The good news for all of you researchers," Hannah said, "is that the skills of research don't become less important in the age of AI. They actually become more important. Skills like framing good problems, asking goodlea questions, making implicit things explicit so that they can be worked with." 

Ready to turn your research into clear direction and actionable insights? Learn how to bring AI into your research process as an exploration and analysis partner in our course Human-Centered Research with AI.

To help practitioners answer the question of where and how AI research tools can assist in the process, Hannah developed what she calls an AI affordances framework.

An "affordance," popularized in the design world by Don Norman in The Design of Everyday Things, describes the properties of an object that suggest how it should be used. Hannah borrowed the concept and asked: what does AI afford for research?

Through her research and many conversations with people across IDEO, she landed on two fundamental affordances: speed and scale. Rather than treating these as endpoints (go faster, go bigger) Hannah maps them as axes of a 2x2 framework. 

The Speed Affordance: From Accelerate to Deliberate

Speed isn't just about moving faster. On one end of this spectrum, you're using AI's generative capacity to accelerate, taking in more, moving quickly, covering ground. On the other end, you're using AI's reflective capacity to slow down and introduce meaningful friction into your process. Both are valid. The question is which end of the spectrum serves your research at this moment.

The Scale Affordance: From Expand to Interrogate

As with speed, scale isn't just about going bigger. On one end of this spectrum, you're using AI to expand, broadening your aperture and taking in a wider range of perspectives and information. On the other end, you're using AI to interrogate, narrowing your focus and excavating a specific space more deeply. Again, both are valuable. The question is whether your research needs breadth or depth right now.

The four quadrants that emerge from that intersection each describe a distinct mode of using AI in research. At the intersection of those axes is where Hannah finds the most exciting and strategic use cases for AI. 

Mode 1: Widen Your Aperture—Use AI to Take In More

The first mode is when we use AI to accelerate and expand our thinking. It’s probably the most familiar: using AI to take in more information, faster. Compiling desk research. Summarizing documents. Translating materials. Cleaning up disparate datasets. Making more information accessible at the start of a project, when you're trying to get your bearings.

"This is the quadrant that I think most people think about when they think about AI," Hannah said. "Using AI's speed and expansive scale to do more, to go bigger."

She uses it this way herself. But she's deliberate about not staying there. "We can't live in this constant state of expanse. To do good design and good research, we have to converge. And the art here is how we use AI to go big and go wide intentionally, and then how we use it to engage more deeply with that complexity once we've found it."

The value of this quadrant is real, but it's a starting point, not a destination.

Mode 2: Notice Nuance—Use AI to See Deeper, Not Just Wider

In the second mode, you’re using AI to expand your research, but in a slower, deliberate way. Here, AI isn't just widening the aperture. It's helping you look more carefully at what's inside it.

"This is still about seeing more," Hannah said, "but it's seeing deeper rather than seeing wider."

In practice, this might look like asking AI to surface patterns across your desk research, identify perspectives you might be missing from your problem framing, or suggest analogues worth exploring before you talk to participants. It's also where AI can scaffold human synthesis.

Hannah was careful about the word "synthesis" here. At IDEO, synthesis means something specific: the human act of making meaning. AI doesn't do that. But it can set your brain up to do it better. "I'm not suggesting that you're using AI to make meaning, but you might be using AI to analyze data, to visualize patterns, to generate initial frameworks, to scaffold your human synthesis process."

"We are thinking about how we are using AI to set up our brains to do meaningful and meaty thinking, not outsourcing thinking to AI." — Hannah Rosenfeld

Mode 3: Challenge Your Assumptions—Use AI to Interrogate Your Thinking

If the first mode is about going as fast and as big as possible, this one is the opposite. "This is about slowing down and introducing meaningful friction into our process," Hannah said.

The third mode is where AI functions as a dialectical partner and a challenger. It surfaces assumptions, critiques thinking, and pushes back. For anyone who's worked on an interdisciplinary team and felt the value of a genuinely different perspective in the room, this quadrant approximates that experience. It’s especially helpful for researchers working alone.

"Bias is inherent," Hannah said. "We all have them. Interdisciplinarity depends on bias and that yours is different than mine is different than hers. Bias is not the problem. Unexamined bias is."

Her favorite moment to use this mode is right at the start of a project before research gets underway. It helps to understand the assumptions you’re making.

"This dialectical nature of AI also forces us to make explicit things that are implicit for us…The benefit of that is it makes for more bias-informed problem-framing, but it also makes you better at collaborating with other people." — Hannah Rosenfeld

The fourth mode, where the acceleration and interrogation affordances meet, is where AI most visibly starts to blur the boundary between research and design. Here, AI's generative capacity is used not to produce final artifacts, but to quickly bring questions, hunches, and ideas to life so you can engage with them, test them, and see what's missing.

"Using this generative capacity to concretize thinking helps us get our thinking out of our heads so that we can engage with it or invite other people to engage with it," Hannah said.

A simple example: you ask AI to generate a recruiting plan based on your research questions. You don't use the plan as-is. You look at it and ask: What's missing? What perspectives aren't represented? What questions aren't being addressed? "That might be evidence for us that we need to flesh that thinking out a bit more,” she noted.

The same logic applies to communicating research. Instead of a static report, you might use AI to turn your findings into a podcast for a time-pressed executive, a set of storyboards for a design team, or a custom agent that answers questions about your research long after your final presentation.

"It's speed in service of understanding, not speed for efficiency's sake,” Hannah clarified. “We're using AI to concretize thinking so that we can see it more deeply, and we can share it more resonantly."

The Skills AI Can't Replace: Why Human Judgment Matters More, Not Less

The throughline in Hannah's framework is that AI raises the stakes for human judgment rather than lowering them.

The researchers who will use AI most effectively aren't the ones who adopt it fastest. They're the ones who understand what makes research good in the first place. They can frame problems, ask better questions, and recognize the difference between a pattern and an insight.

"As AI automates creation, the altitude of our role as researchers can evolve," she said. "We can spend more time in thoughtful, bias-informed problem framing. We can be more critical and iterative in how we're asking questions."

There's also something larger at stake. "We are, as researchers, working at a very historical moment. We have an incredible privilege—and maybe an overwhelming responsibility—to write the rules and build the playbook that defines how AI should be used in research, where it truly adds value, and where it doesn't."

Key Takeaways

• There's more than one way to use AI in research. IDEO's AI affordances framework identifies four distinct modes where AI adds value in the human-centered research process. 

• Most teams use AI to go faster. The more interesting question is: what can you do now that you couldn't do before? 

• AI expands what research can be, but it doesn't change what makes research good. 

• Meaning-making is still human work. The leap from pattern to insight requires human judgment.

• The most underused mode is using AI to slow down, not speed up. 

• Making your thinking explicit for AI also strengthens human-centered research by improving how you communicate and collaborate with others. 

Keep Learning

IDEO U Course

Human-Centered Research with AI — Learn how to bring AI into your research process as an exploration and analysis partner. Develop a deeper understanding of people and their needs, motivations, and challenges, and move from data to direction with confidence. 

Related Posts

Insights in Action: Impactful Research to Fuel Innovation

What Is Design Thinking?

Join Future Live Conversations

Join a Creative Confidence Podcast live recording to get inspired and learn practical ways to grow your creative confidence. Sign up for IDEO U emails to receive event invitations, and subscribe to the podcast on Apple Podcasts or Spotify to hear episodes first.

About the Speaker

Hannah Rosenfeld 

Executive Design Researcher at IDEO

LinkedIn

At IDEO, Hannah leads complex design programs and guides interdisciplinary teams in tackling big, ambiguous questions with a focus on designing for healthcare. Her work blends design research, strategy, and storytelling to help organizations uncover human insights and turn them into meaningful products, services, and experiences that reimagine how healthcare is delivered and experienced. She earned her Master's in Designing for Interactions from Carnegie Mellon University.