Clayton Christensen's Jobs To Be Done (JTBD) flips the product perspective. You don't make a product for users — users hire your product to do a specific job.

The classic example: McDonald's milkshake. The company tried to improve sales by adding flavors and changing texture. Nothing worked. Then they asked: why do people actually buy a milkshake? The answer surprised them: in the morning, in the car, because:

Lasts the entire commute (unlike coffee or donut)
Keeps a hand free (unlike a banana)
No mess (unlike a bagel)
Enough energy for a long ride

The customer doesn't hire the milkshake "to drink chocolate" — they hire it to entertain and fill a boring commute.

That insight changed the product direction.

Three job dimensions

JTBD splits jobs into three dimensions:

Functional — what concrete thing needs to get done ("get to work")
Emotional — how you want to feel ("not tired in the morning")
Social — how you want to appear to others ("not wasteful with time")

Every purchase is all three simultaneously — at different weights. The most common mistake: focusing only on the functional.

Switch interview — JTBD's core tool

Ask the customer to tell you about the time they switched from an old solution to a new one (or vice versa). Don't ask "why did you buy" — ask "how did you end up buying". Specific questions:

"When did you first think you needed something like this?"
"What was the trigger — what happened just before you started looking?"
"What other options did you consider?"
"What decided this one?"
"What was the hardest part of switching?"
"What do you do differently now?"

Output: you understand the job the product was hired to do — not features, but the context the product steps into.

How AI helps

1. Hypothesis generation before interviews

Describe your product to AI → ask "Give me 5 hypotheses for the job a user hires this product to do. Include all three dimensions (functional, emotional, social) in each."

Useful because in interviews you won't hit the right answer immediately — you have to search. A list of hypotheses guides listening.

2. Switch interview preparation

Give AI a customer segment description → ask for a switch interview script. You get 10 questions and 3 screening criteria.

3. Transcript analysis

You've done 5 switch interviews — 2 hours of recordings. AI transcribes and you ask: "Which jobs recurred across interviews? What was most surprising? What didn't appear that I expected?"

Traditionally this required a qualitative researcher's two weeks. AI compresses it to an afternoon.

Pitfalls

1. AI-generated jobs are averages. Real jobs only surface in real interviews. AI is a preparation tool, not a replacement.

2. Don't trust the first answer. Customers give a surface reason first. Switch interviews dig deeper.

3. The answer can't be used directly in marketing. A customer won't say "I want to feel important" — but if JTBD interviews reveal an emotional job, marketing approaches it indirectly.

Innovaidor with JTBD

In Innovaidor, JTBD is one of the methods. You activate it when you're questioning your user assumption. AI walks through three dimensions, pushes for specificity, generates a switch interview script. Output: a document describing what kind of user interview you're about to run and what you're looking for.

Closing

Jobs To Be Done is one of the most powerful tools because it shifts focus from your product to the customer's life. You sell not features but getting a job done. AI helps build hypotheses and prepare interviews faster than ever — but the interviews themselves are still your work.

Start with one customer: ask them to tell you about the time they switched to the current solution. Listen carefully. The answers aren't where you think.