Can Generative AI replace JTBD research?
I recently stumbled upon a website claiming that Jobs To Be Done (JTBD) interviews are outdated in the era of AI. This site seemed quite eager to dismiss consultants, insinuating that they drain your funds while AI could swiftly provide all the answers to JTBD interviews, reducing timelines from weeks to mere days.
But hold on a moment...
Sure, AI can assist in JTBD interviews. You can tap into AI for brainstorming sessions. However, here's the catch: AI operates on existing information. It can only process data already available in the public domain. This explains why consultancies were cautious about their staff using ChatGPT for client projects or innovation ventures; they risked unintentionally feeding new information into AI systems.
Consider one of the most famed JTBD case studies—why people buy milkshakes. Back then, the company that hired Clayton Christensen's team lacked straightforward methods to understand their milkshake sales. Hence, they enlisted this formidable team. Admittedly, AI could have expedited certain tasks for Christensen's team.
For instance:
Prompt: "Can you provide open-ended questions to ask people about their drink preferences?"
ChatGPT:
Certainly! Here are some open-ended questions to glean insights into why people choose a specific drink:
1. What led you to select this particular drink today?
2. Can you describe the factors influencing your decision to purchase this beverage?
3. How does this drink align with your overall tastes and preferences?
4. What aspects do you find most appealing about this drink compared to others?
5. Was there a specific occasion or mood prompting your choice of this drink?
6. Could you share any memories or experiences associated with this beverage influencing your decision?
7. What do you hope to gain from consuming this beverage?
8. Are there any specific ingredients or flavours in this drink that appeal to you?
9. How does this drink fit into your lifestyle or dietary requirements?
10. Can you explain how you became interested in or discovered this particular drink?
These questions aim to elicit detailed responses about individuals' motivations and experiences related to their beverage choice.
Yet, relying solely on ChatGPT for insights into milkshake purchases wouldn't provide a comprehensive view. Even today, querying ChatGPT—equipped with widely available solutions—on why people buy milkshakes won't yield the complete answer.
Given that Generative AI can perpetuate biases, it's crucial to scrutinize the datasets we employ. We must learn to harness AI alongside traditional research methods like ethnographic studies. As we guide Generative AI to gather insights, we must also tailor the output for the specific machine learning contexts we operate within.
Now, more than ever, it's paramount to gather the right data for analysis. Returning to JTBD theory, we may uncover competitive advantages and the quest for the Blue Ocean by understanding the functional, social, and emotional jobs of individuals in our ecosystems better than our competitors. This insight allows us to identify distinct data points crucial for building our case, such as behavioural, demographic, interest and preference data, contextual data, interaction history data, and sentiment data.
Example: Sitecore Search can offer hyper-relevant content through AI-powered search - which you can still make better by better understanding the jobs of customers!
Innovation still hinges on human ingenuity—the ability to observe, probe, and generate new ideas. While Generative AI is indeed remarkable, its ability to list and summarise based on existing information does not equate to having all the answers.
Let's not be misled. Generative AI is a powerful tool, but it's not omniscient.
