Utilizing Jobs-To-Be-Done in the age of AI
Editor’s note: Jennifer Axen is the founder of Axen Insights, an insights and experience strategy consultancy. With 25+ years of experience across consumer insights, UX and design research, she helps organizations turn human understanding into better products, experiences and innovation. Find Axen on LinkedIn.
I’ve been using Jobs-To-Be-Done (JTBD) in my research practice for years. Lately, I’ve noticed the term popping up in briefs, innovation conversations and research discussions.
But why?
JTBD has been around for nearly 30 years, an impressive run in a field that is constantly evolving its methods. What does it help researchers uncover that another approach might not? And what happens to a methodology grounded in deep human inquiry as AI transforms how we collect, analyze and act on research?
To find out, I spoke with two people who have been instrumental in shaping JTBD: Taddy Hall, who worked closely with Clayton Christensen to develop and validate the theory and later co-authored “Competing Against Luck,” and Chris Spiek, who spent years developing, applying and teaching JTBD alongside Bob Moesta through the Re-Wired Group.
From what people say to what drives behavior
At its core, JTBD gives researchers a different lens on a familiar question: Why do people do what they do?
We have plenty of tools for understanding who people are, what they think and how they behave. But those things don’t necessarily tell us what caused someone to make a particular choice.
Hall sees this distinction between correlation and causality as central to JTBD.
“What’s actually predictive of somebody’s decision to either pull a product or service into their lives or not is the extent to which they enable desired progress and fulfill aspirations that people have in a given circumstance,” says Hall.
That word progress is important. JTBD isn't simply about uncovering needs or pain points. It looks for circumstances where people have enough motivation to act – to change something about the way they are currently solving a problem.
Spiek remembers that distinction clicking for him early in his career. He had learned personas, needs and other product development approaches, but they never felt concrete enough. JTBD gave him a way to investigate the forces behind an actual decision.
For researchers, that shift is significant: from asking people what they might want to reconstructing what actually happened.
The interview as investigation
This is where JTBD research gets particularly interesting.
A good JTBD interview can feel more like an investigation than a traditional IDI. Rather than moving through a list of attitudes and preferences, the researcher reconstructs a real decision in painstaking detail: What was happening? What wasn’t working? When did you first start thinking about making a change? What alternatives did you consider? Who else was involved? What finally tipped the decision?
Hall describes this as getting back to the story underneath the data.
“What [JTBD] does is it restores our focus on the raw data behind the numbers,” he says. “And what you find is that the raw data has the shape of a story.”
Importantly, researchers don't simply ask why did you do that? And then treat the response as truth. People are generally much better at recalling what happened and how they felt than accurately explaining their own motivations.
That puts more responsibility on the researcher.
“Respondents don’t give you the insights. That’s our job. They give you the raw material of insights,” says Hall.
For me, that captures one of the most compelling things about JTBD. It demands strong interviewing, listening and synthesis skills. The insight isn't sitting there waiting to be collected. We have to find it.
When a “job” isn’t really a job
That rigor also matters because JTBD terminology has become so widespread that almost anything can get called a job.
Hall calls this “job washing” – taking an existing need, observation or persona, putting JTBD language around it and declaring the job identified.
Needs and jobs aren't necessarily the same thing. Most of us need to exercise more, save more or sleep more. That doesn't mean we'll do any of those things tomorrow.
Hall argues that the awkward “to be done” language is actually important because it signals energy for progress: evidence that someone is expending energy to resolve a trade-off, navigate a struggle or fulfill an aspiration.
Spiek offers another useful perspective.
JTBD isn’t meant to replace behavioral data, quantitative research or the rest of the tool kit. Its power is in helping researchers understand something those sources often can’t: What happened that caused behavior to change? As Spiek learned applying JTBD across everything from software to CPG, the category may change, but the underlying challenge doesn’t: “You need to understand demand in order to do it.”
Research that points toward opportunity
JTBD becomes especially powerful when research needs to do more than describe the current state.
Across categories from financial services and healthcare to streaming, snacks and steel, Hall has seen competitors repeatedly organize around a relatively narrow set of benefits. JTBD can uncover a broader range of experiences people value.
“It reveals the full portfolio of drivers of desired experience,” Hall says, creating opportunities to “change the basis of competition.”
That's an important bridge between insight and innovation. Instead of simply telling teams what customers think about today's experience, research can identify where there is energy for something better.
And the behavior researchers study doesn't have to be a classic brand switch. Spiek uses JTBD to understand why an existing customer suddenly adopts a feature they previously ignored. What changed? Why did it become relevant now?
Those are research questions with direct implications for product and experience design.
AI makes the “why” more valuable
AI adds an interesting new dimension to all of this.
Spiek is already using AI to analyze JTBD interviews, identify forces and timelines and even provide feedback on interviewing techniques. At the same time, AI is dramatically accelerating product development.
That creates a new challenge: “I can build anything. I can ship anything very, very quickly. I’ve got to figure out what to go build,” says Spiek.
For research teams, that may be the bigger AI story. As the cost of creating and testing ideas falls, knowing what is worth creating becomes more valuable.
Hall sees AI helping researchers explore jobs and generate hypotheses. But he also points to an important limitation. “These are large language models,” says Hall. “They’re not deep emotion models.”
The tone, hesitation, enthusiasm, anxiety and contradictions inside a human story are often exactly where the insight lives.
AI can help us analyze faster and perhaps become better researchers. But JTBD is a useful reminder that our value isn't simply in collecting and summarizing information. It's in knowing what to listen for, where to probe and how to make meaning from what we hear.
Perhaps that's why a nearly 30-year-old methodology feels so current.
As AI makes it easier to build almost anything, research has an increasingly important role in answering a different question:
What progress are people already trying to make and what is worth building to help them get there?