Are Jobs-to-be-Done Interviews Relevant in an AI World?
How ChatGPT has revolutionized the way I do qualitative research
As artificial intelligence (AI) continues to advance and disrupt industries worldwide, it raises questions about the relevance of traditional qualitative research methods like Jobs-to-be-Done (JTBD) interviews. While many researchers and practitioners have benefited from the insights provided by these interviews, there are concerns that they can be time-consuming, expensive, and biased. As a result, some argue that we should shift our focus towards more quantitative research methods. In this blog post, we will explore the ongoing relevance of JTBD interviews in an AI-driven world and consider the advantages of combining qualitative and quantitative research.
JTBD Interviews in an AI World
Jobs-to-be-Done is a framework that focuses on understanding the underlying goals and objectives of customers to develop the data that points to value gaps in the market so we can create products and services that cater to their needs effectively. JTBD interviews have been a crucial part of this process, as they allow researchers to dig deep into the customers' experiences and expectations - with the goal of building a model of the problem-space. However, when we do them may have to change.
In an AI-driven world, I would argue that these interviews are no longer necessary on the front-end of the research. After all, AI can process massive amounts of data quickly and efficiently, providing valuable insights that would take humans much longer to uncover. However, there are several reasons why JTBD interviews remain relevant:
Depth of understanding: While AI can quickly analyze vast amounts of quantitative data, it may lack the human touch needed to understand the subtleties and nuances of individual experiences and situations. JTBD interviews provide a platform for researchers to empathize with their customers and gain a deeper understanding of their motivations and desires. The quantitative phase of Jobs-to-be-Done is actually designed to handle this part. Perhaps one day we will use machine learning to accelerate our data analyses.
Identifying unarticulated needs: AI algorithms excel at identifying patterns and correlations within data, but they may not be as adept at uncovering “unspoken or latent needs.” Actually, we’ve long known that this is irrelevant in Jobs-to-be-Done research because we aren’t looking for solutions, we’re attempting to construct a model that is forward-looking. People are perfectly able to articulate the objective they are trying to reach.
The human element: The process of conducting JTBD interviews can create a sense of trust and rapport between customers and researchers. This is especially important after a quantitative analysis when trying to understand the root cause that drove a needs-based segmentation or ranking of needs. JTBD practitioners really shouldn’t rely on their instinct here. It’s difficult to turn dots on a plot into a compelling story full of the necessary emotional elements.
Balancing Qualitative and Quantitative Research
Instead of abandoning JTBD interviews in favor of a pure AI approach, a more balanced approach might be more effective. By going back to respondents from the quantitative phase who demonstrated that they struggled to get a job done along the same dimensions as a group of others, we can use our JTBD interviewing skills to dig deep into the “why”…the root cause of their problem instead of doing it to build the initial model. There is no market of one, and different groups of customers will struggle differently from others.
The numbers show that, but the dimensional data you collect may not paint a compelling picture when you get to the story-telling phase. Sometimes you have to get the verbatims straight from the horses mouth. Understanding why is an area I believe needs significant improvement.
AI-powered tools can be employed to analyze customer data and identify trends or areas of potential opportunity. These insights can then be used to inform the design of JTBD models, helping researchers focus on the most critical aspects of the customer experience when constructing surveys.
While AI has undoubtedly revolutionized the way we conduct research, and make data-driven decisions, it is essential not to overlook the value of qualitative methods like Jobs-to-be-Done interviews. These interviews provide invaluable depth and understanding, complementing the analytical power of AI algorithms. By adopting a balanced approach that combines the strengths of both qualitative and quantitative research, organizations can ensure they stay ahead of the curve in an ever-evolving AI world.
Two things are important.. getting to know what the continuum is and other is validation. To me personally the contextual information now made available faster thanks to AI fastens the validation conversation.
I'm a little surprised by ChatGPT's ability to create hypothetical job maps and outcome statements. It certainly doesn't produce junk with the jobs I've asked it to work with.
I do wonder if conversational chat bot style AI could be used to replace the human effort to capture job steps and desired outcomes. I'm trying very hard to learn JBTD and it feels like a high effort approach that may be rejected by many organisations as too hard. My instant reaction to that is "of course it should be hard" but if I was to write an outcome statement for product discovery it would be "minimise time needed to understand market needs"