That changes what “knowing your customer” means. The competitive edge isn’t just understanding the person anymore. It’s understanding, and managing, how AI systems represent you to that person before they ever reach out.

Kenny and Pogrebna illustrate this with three small and mid-sized businesses, and the pattern across all three is more useful than any single case.

1. A manufacturer stopped treating every inquiry as equal. Once a meaningful share of inbound requests started arriving pre-drafted by AI tools on the buyer’s side, the company built a screening step before committing engineering time to a formal quotation. Not every AI-generated inquiry represents a real, qualified buyer, and treating them all as if they do quietly drains the team that’s supposed to be closing the real ones.
2. A boutique hotel started auditing what AI tools were saying about it. Instead of only managing its own website and listings, the business began actively checking how AI assistants described it to people researching where to stay, and found gaps and inaccuracies worth correcting. The insight here isn’t “get more content.” It’s that the audience for your public content now includes AI systems reading it on a customer’s behalf, and if that content is thin or ambiguous, the AI fills the gap with its own inference, which you don’t control.
3. A B2B software company retired the periodic customer survey. In its place: continuous monitoring of customer feedback and of how AI tools were characterizing the company in the market, tracked as an ongoing signal rather than a quarterly snapshot. The shift is from asking customers what they think once in a while to watching, in real time, what the market (increasingly mediated by AI) already believes.

Three different businesses, three different fixes, one underlying move: each stopped treating AI as an internal productivity tool and started treating it as a new party sitting inside the customer relationship, one that filters inquiries, describes the business to strangers, and forms opinions about market perception, whether or not anyone at the company asked it to.

The authors’ conclusion is worth sitting with: winning here will not go to whoever spends the most on AI technology. It goes to whoever builds the operational habit of continuously listening to these AI-mediated signals, interpreting them correctly, and adjusting strategy as they change, because both the customers and the information layer they rely on are now shifting in real time.

This matters beyond the three cases

This is a variation on a theme that’s been building all year: the discovery layer between a business and its customers is splitting in two, one half still made of humans searching and comparing directly, the other increasingly made of AI systems doing that research on a human’s behalf and compressing it into a recommendation. A business that only optimizes for the first half is optimizing for a shrinking share of its actual funnel.

For most companies, none of this is currently anyone’s job. Marketing owns the website. Sales owns inbound leads. Nobody owns the question of what ChatGPT, Perplexity, or a buyer’s own AI research agent is telling a prospective customer about you right now, today, possibly incorrectly. That’s not a content problem or a technology problem in isolation. It’s an operational gap: no one is watching a channel that’s already influencing deals.

What this actually requires

None of the three fixes in the HBR piece required a large AI budget. They required treating AI-mediated interactions as a monitored, owned part of the business rather than background noise. In practice, that means three concrete moves, each mirroring one of the cases above:

Build a triage layer for AI-originated inbound, so a flood of AI-drafted inquiries doesn’t quietly consume the capacity meant for qualified buyers.

Audit what AI tools currently say about your business, the way you’d audit your own website copy, and correct the gaps before a prospect acts on a wrong impression.

Replace static, periodic feedback collection with continuous monitoring of both direct customer sentiment and AI-mediated market perception, so shifts get caught in weeks, not at the next scheduled survey.

This is the same discipline Crizzen applies to AI implementation generally: the win isn’t the tool, it’s the system built around it that actually gets used, watched, and acted on. If AI is now sitting inside your customer relationship whether you invited it or not, the only real choice left is whether someone at your company is managing that relationship on purpose.

To talk through what an AI-mediated customer audit would look like for your business, email info@crizzen.com.

Crizzen is an enterprise AI firm where we take AI from strategy to production by embedding it directly into the systems and workflows your teams already use, and we measure success in business outcomes, not pilots that never scale.