In the next 12 months, AI will be embedded into most business workflows, including expansive utilization with market research and insights. As corporate insights departments and research providers race to add AI solutions to their offerings, it’s worth considering: what makes a company adding AI to its workflow successful? Is it being the first to add AI? Or, adding the most AI?
Probably neither of those will determine an organization’s success with AI. Wantonly adding AI is more likely to harm workflows than improve them. Instead, the predictor of what organizations will most effectively apply AI to their solutions are those that most thoroughly apply AI to their stakeholders’ jobs-to-be-done.
Einstein in a Tiger Fight
Let’s consider this thought exercise recently posed by Kevin Kelly, founder of WIRED magazine (emphasis mine):
“Intelligence itself is overrated…To get things done, it’s not necessarily the most important. Take Einstein and a tiger and put them in a cage. Who wins? It’s not the smartest guy. Why is that? Because it takes other things besides intelligence to get things done in the world.“
And of course, because this is an AI blog post, I had to generate an AI image to accompany it (this version was my favorite, but feel free to reach out for more đ).
“Albert Einstein in a boxing ring with a tiger.” Generated on Midjourney, June 2023
Kelly follows up by stating (again, emphasis mine):
“It’s not the company with the smartest people that will dominate. It’s not the smartest person in the room that necessarily knows what to do. Intelligence is necessary but not sufficient to get things done in the world.“
The LLMs released to the public so far are incredible; their ability to quickly process and predict information surpasses all humanity. They are superintelligent. But, just like humans and human-derived insitutions, they are only as good as the way they’re applied to solving stakeholders’ jobs-to-be-done.
AI, Insights, and JTBD
The jobs-to-be-done (JTBD) framework is hopefully familiar to most market research practitioners, but for completeness, I’ll provide a quick summary.
JTBD was popularized by the late Clayton Christensen, which he described as,
âWhen we buy a product, we essentially âhireâ something to get a job done. If it does the job well, when we are confronted with the same job, we hire that same product again…”
In other words, our stakeholders (whether an internal stakeholder or a client) buy our services to accomplish a job. Christensen follows with three examples of JTBD in a consumer context:
“Uber founders recognized that urban transportation was doing a poor job and found a way to one-up cabs and car services by allowing consumers to hail cars within minutes on their phones. Care.com, the online matchmaking service for child care, senior care, and pet care, was developed 10 years ago by Sheila Marcelo after she struggled to fulfill her own childcare needs.
It helps if a company can sell not only a product but an experience. Pre-teen girls âhireâ American Girl dolls to validate their sense of self-worth, while mothers (the ones actually purchasing the dolls) pay a premiumâmore than $100 per dollâpartly as a way to connect with their daughters. The company has sold 29 million dolls and pulls in more than $500 million in sales each year.”
Now, let’s apply this to insights and market research. What jobs are insights professionals and companies “hired” to do for the organizations and stakeholders that hire their services?
While there are heterogeneous motivations for hiring insights, I think we can reasonably summarize the job as: “Insights are hired by leaders to enable them to make confident decisions within a given time and cost.”
The questions for us to ponder as AI is added to nearly all insights workflows are:
- Does AI improve the speed at which insights are delivered?
- Does AI improve the cost at which insights are delivered?
- And finally, does AI-generated insight improve the confidence of decision-makers?
Answering these questions helps users of insights AI to evaluate its appropriate application to their needs.
Summary
Generative AI holds incredible potential for the future of market research and will appear in many applications, including Fuel Cycle. However, successful implementation of AI requires an understanding of stakeholders’ jobs-to-be-done, whether those stakeholders are internal or external to your company. The goal is not to implement AI; the goal is to improve the job-to-be-done.
We believe generative AI improves our customers’ jobs-to-be-done among some research and insights use cases, but not all. In our next blog posts, we’ll dive deeper into the role of generative AI in fulfilling JTBD along with research design, analysis, and reporting.
As always, we view our contribution to the insights industry as a dialogue and invite collaboration and input. We’d love to hear from you. Share your thoughts on LinkedIn or contact us directly for a personal walkthrough of Fuel Cycle and our emerging AI capabilities.
