Why AI Moderation Is a Bad Idea for B2B Research
Companies selling ‘AI-enabled’ tools are attempting to GenAI-ify seemingly every aspect of the qualitative and quantitative market research process—from AI analysis to AI moderators, AI-generated personas, and even AI participants! I’m just waiting on the AI-generated clients.
Clients conducting B2B research should think about how they use AI and what signal it sends to their participants, since their participants often represent important people in their network! Could be decision-makers at prospective or current client companies, people who work at important partner orgs, or just generally people the client respect and value hearing from.
1. Consider current associations. Currently, GenAI = slop
As a qual researcher I am here to remind you that meaning is always defined partly by context. So no matter how well the AI moderation tool works or how thoughtfully designed it is, it exists within the current market category of “GenAI.” And there are currently some strong (negative) connotations surrounding GenAI among consumers.
Exhibit A: Discourse surrounding AI slop and slop-ification.
Exhibit B: Younger generation’s use of “AI” as a slang term to describe something cheap, fake, or low quality.
Currently AI signals cheap, low-effort, and spammy—is that the kind of association you want potential buyers to make with your brand?
2. Consider similar contexts where AI moderation is used
AI-moderation seems to have the widest adoption currently in job interviews, as a replacement for the 20-minute recruiter screener call. This appears to be universally hated. (Do you have evidence to the contrary? Let me know!) Probably because they signal to the interviewee that they are not yet worth the time and energy of engaging with a real, live human.
AI “moderation” for survey open-ends is one thing, but to ask your potential customers, former clients, and valued network to talk to a bot for 45-minutes communicates something else entirely: that they are not worth the time or money it would take to interact with an embodied human.
For this reason alone, I would advise a client conducting B2B research to forgo research entirely before they would consider subjecting important stakeholders to AI-moderated interviewing.
3. Risks vastly outweigh the benefits
At best, AI-moderated interviews will deliver surface-level, flattened summaries, defeating the purpose of engaging in qualitative interviews—which are used to surface individual, subjective experiences in depth and detail.
At worst, they could turn-off the people you are most hoping to influence by signaling to them they are worth the time to talk to your chatbot, but not to a human. That you care more about extracting “data” from them than you do about facilitating a positive and engaging experience for them.
Don’t do it! It’s not worth it!