One of my clients is a contractor for whom I conduct the research aspect of projects. Someone interviews the ultimate client, and I get the project that emerges from that conversation. One of the questions from a recent engagement was, “What is the size of the tree stump removal market, including major players, annual growth rate, and key products?”

My first impulse was to feel gratitude at having such a straightforward request. In the midst of research on virtual simulation for healthcare training and the market for small-bore machine parts, I was so happy to focus on a product that I could get my head around. In fact, I have a tree stump in my own backyard whose presence has irritated me for years, so I figured that this was one project that would pay off in practical terms.

Because I have been conducting reference interviews for decades, I paused before starting the project. I still keep in mind a client conversation I had decades ago, in which my client asked me for information on “mobile advertising.” I assumed that he was referring to advertising on mobile devices, an easy-peasy project that I knew I could handle. It turns out he had in mind those panel trucks with advertising on the sides that drive around in rush-hour traffic and function as mobile billboards. Fortunately, we straightened out the confusion while I was still on the phone conducting the reference interview, but that experience spooked me. In fact, it taught me that the simpler the question sounds, the more likely that complications are hiding below the surface.

As will probably surprise no one, there is very little information available specifically on the size of the market for tree stump removal products. And here is where it gets interesting. I reached out to my client and asked for more information about the context of this question, so I could figure out how to expand my research to find something close to what they were requesting. My first thought was to expand the research to the broader lawn and garden market, or perhaps to the market for tree service and landscaping companies. Both were easy to find and seemed like a natural extension to removing tree stumps.

As it turns out, my client’s client had developed a new herbicide that kills tree stumps and roots. Information on either the lawn and garden or the landscaping market would not have been useful in this case, as the product is targeted to a particular need that is unlike products normally used for routine yard maintenance. The answer I gave to the client included some information on the tree stump grinder market (see—even niche markets have niche-ier markets!) as well as an analysis of the articles comparing various stump removal products, highlighting what consumers value the most in these products. While this was not specifically what the client asked for, I knew that it would address their underlying need to find out how to enter this market and compete most effectively.

I have recently been reflecting a lot about the impact of AI and machine learning on the information profession. As more libraries institute chatbots to handle routine inquiries and blockchain to facilitate interlibrary loans, it is important to think about the long-term impacts of this technology on how we provide information services. And this got me thinking—again—of the value of the reference interview and how technology has not yet been able to replicate the conversation between online searchers and their clients.

Had a chatbot been involved in that client reference interview about the tree stump removal market, and had we let AI address the question, I shudder to think of what the result would have been. No, I’m not a Luddite, and I don’t harbor any innate distrust of machine thinking, but my experience has shown me that it takes a human to identify an overly narrow question and to then ask the client for additional context in order to determine what other information would be helpful. My favorite reference interview question, “And if I can’t find exactly what you’re looking for, what would be second best?,” is particularly helpful in these situations.

Just as our profession adjusted to the impact of the internet, open-source information, and the perception of our clients that it’s all on the web for free, we need to look at where technology can save us from routine and low-value work, and where our human brains can best address the information needs of our clients. While we look for ways to streamline our work, we have to also focus on where our (human) neural networks far surpass the abilities of technology.