Using AI to Streamline JTBD Interviews and Analysis

Hello everyone! 👋

I wanted to share a personal project I have worked on in the last months that uses LLMs together with Jobs-to-be-Done to make product development easier and more efficient.

The idea is to automate identifying key jobs, figuring out who performs them, creating synthetic users, and conducting interviews. By doing this, we cut down on the time and resources usually spent on manual user research, making it quicker and simpler to gather the insights needed for your product roadmap.

Here’s how it works:

1. Discovering Main Jobs and Job Performers:
Starting with a rough vision, the code helps you identify and suggest potential main jobs and the people who typically perform them, based on your vision and skillset.

2. Creating Synthetic Users:
I use LLMs to build user archetypes that reflect real needs, goals, and pain points.

3. Automated Interviews:
Using GPT’s language capabilities, I’ve set up a system that runs interviews with these synthetic personas, pulling out key insights on customer motivations and needs.

4. Analyzing Interviews and Extract Needs:
Finally, we break down all the information from these interviews into actionable insights—covering everything from job steps to emotional and social jobs.

This project lays the groundwork for a user-centered product design strategy, helping me make smarter decisions on what features to prioritize, how to improve user experience, and how to drive overall product development.

Would love to hear your thoughts! 💬