Using AI To Augment JTBD Research
The JTBD framework is particularly well-suited for AI-assisted research. Because JTBD is technology agnostic (i.e., you’re not investigating a specific product or brand), the information included in an AI-generated analysis is broad and inclusive, thereby increasing the strength of output. Additionally, the rules of formulation for JTBD statements provide consistent results that large language models can better handle.
Before using AI in your JTBD investigations, keep two important limitations in mind:
- Validity
The accuracy of the results from AI must be checked. Generative AI is known for “hallucinations” that can give questionable results that may be just plain wrong. Never take the output at face value without critically reviewing it. - Completeness
At the moment AI seems to cover about 60–80% of the ground needed for a JTBD investigation, so you will likely experience gaps. Keep an open mind to discover more as you continue your investigation with real job performers.
With this in mind, the advantages of using AI during JTBD research are many:
- Refine your target job: AI can help you quickly explore related jobs and alternative job performers, as well as get the right level of abstraction for your selected target job.
- Get a head start: You can quickly generate a job map and other elements of a JTBD framework to understand the space around the target job.
- Understand new fields quicker: If you’re working in an unfamiliar field, AI can help you map out the core concepts in the space through the lens of JTBD
- Cover different points: AI can identify aspects of JTBD that you might overlook.
Overall, AI can empower you as a researcher and instill confidence. Your interviews then shift focus from trying to fill out a blank page to validating information and digging deeper. AI can get the basics, but it’s likely your real opportunities for innovation will lie further below the surface.
For instance, imagine you were contracted to a JTBD analysis on risk compliance auditors in the mining industry and you know next to nothing about that field. You could use AI to generate a rough job map to help you understand the domain before you even do an interview.
Then, when you speak to job performers, you’re not starting from scratch but rather working off some initial assumptions that AI helped you find. This allows you to focus on validating what you’ve found so far. In this scenario, AI would make you a more informed, smarter researcher.
AI Tools For JTBD
ChatGPT is a natural starting point to bring AI into your JTBD research, often providing decent results. The prompts must be quite explicit to teach algorithms not only what JTBD is but what you’re looking to generate exactly. We recommend starting with a job map. In ChatGPT you’d have to define what a job map is in your prompt as well as how you want steps formulated.
Here’s an example prompt using the previous scenario of risk compliance auditors in mines:
“I’m working with the JTBD framework and would like to create a job map for risk compliance auditors working in the mining industry. The target job is to conduct a risk audit.
A job map is a chronological list of steps to complete that job independent of technology or solutions. Each job step must begin with a verb in the first person. I’d like you to generate a job map.”
Since ChatGPT is iterative — i.e., you can “chat” back and forth to refine the output — you have several options at this point to refine the output.
- Regenerate a response based on your initial prompt. ChatGPT typically produces different results with each query.
- Give follow-up prompts to reformulate the steps or include more or less of them as needed, e.g., “please don’t include any ANDs or ORs in the job steps, just singular concepts.”
- Drill down on a given step to find the sub-steps, if desired, e.g., “how do risk auditors complete step 8 in more detail”?
- Narrow the target job or job performer to learn more about potential variations, e.g., “risk compliance in coal mines” vs “risk compliance in diamond mines.”
Use AI to explore other elements of the JTBD framework, as well. You can generate initial lists of outcome statements, emotional and social aspects, and job differentiators, in addition to exploring related jobs and aspirations.
There are also a growing number of AI tools specifically calibrated to JTBD. For instance, the “Job Map Creator” at www.cookup.ai created by Renato Caliari already knows what a job map is. All you have to do is give a target job, and it will generate an initial job map. Other tools are emerging as well, such as www.joblens.ai.
Keep in mind that your judgment is always required to assess the accuracy and completeness of the AI-generated output. Don’t see AI as a way to avoid speaking directly with job performers, but rather as a way to make your interviews more focused and impactful.
Embrace the synergy between AI and human intelligence to create more focused and impactful JTBD interviews, driving innovation and customer-centric solutions.
For more on AI for JTBD, we recommend the work of Evan Shore, in particular a recording of his webinar with the JTBD Toolkit on AI and this interview with him.
Check out these resources as well:
- Are Jobs-to-be-Done Interviews Relevant in an AI World? Mike Boysen
- JTBD AI-Powered, Renato Caliari
- Streamline Your JTBD Research with AI-Powered Jobs to Be Done, Brian Rhea