One breakfast spot has two hundred reviews saying “delicious, friendly owner.” Another has only a few dozen, but someone writes: “You can pick up your order at seven in the morning, there are no steps to climb with a stroller, but it’s noisy if you eat in.”
If a customer asks AI, “I need to take my two-year-old out for breakfast and get to work by eight. Which place would be convenient?” which one would you bet on?
Don’t rush to place your bet based on star ratings. There’s an assumption that often slips in here: the scores humans use to express satisfaction are equivalent to the information AI uses to judge suitability. The two overlap, but they aren’t the same thing.
I tend to think that a four-star review with clearly defined boundaries sometimes offers more value for recommendations than ten vague five-star reviews. Recommendations need reasons, and “everyone says it’s great” is only one of them.
The Stars Shine Bright, but the Content May Be Dim
Starting from industry narratives like “burning through tokens,” the question I’d rather pursue is: when businesses are busy producing more text, which words actually add information that can inform a judgment?
“Professional, attentive, highly recommended” could appear under a dentist, a photographer, an auto repair shop, or even a breakfast spot. These words aren’t necessarily false; they just do little to distinguish one business from another.
Try this instead: “The photographer is willing to wait until the child settles down before shooting and doesn’t rush them to smile; the finished photos lean toward natural light, so this isn’t a good fit for people who want glamorous studio portraits.”
This review reveals the approach to service, whom it suits, and its limitations. If AI search can access and use it, it has more material to answer “who would be a good fit for this photographer.” Whether it uses that material and how it ranks results still depend on the system, the sources, and the question at hand; no one can guarantee a recommendation based on a piece of copy.
The value of reviews may be shifting from applauding businesses to drawing boundaries around customer needs.
Negative Reviews Can Also Help You Find the Right People
Imagine a quiet little bar receiving a three-star review: “You can’t talk loudly, and they won’t take groups of ten. Too many rules.”
The owner’s blood pressure rises upon reading it. But another customer is looking for “a bar where I can sit alone without running into large gatherings.” That same complaint has unexpectedly provided a clue for a match.
This doesn’t mean the more negative reviews, the better. Food safety concerns, arbitrary charges, and service failures can’t be spun into distinctive features by saying “we serve different customers.” What’s truly interesting is that some low ratings describe a clash of preferences, rather than a failure of quality.
If marketing focuses only on smoothing over every dissatisfaction, businesses may also sand away their own contours. In the end, every place is friendly, professional, and great value for money, like the same face with a different sign.
Who Can Get Their Service Clearly Described?
There’s also a somewhat uncomfortable gap here. Businesses that are good at encouraging customers to leave specific reviews may be more likely to leave clues AI can use than those that quietly focus on providing good service but don’t know how to manage the words around it.
So is the recommendation based on service quality, or on the quality of how that service is described? The two cannot simply be equated.
For local businesses and professionals, a more practical digital marketing adjustment is to change review requests from “give us five stars” to “what did we help you solve this time? What didn’t meet your expectations?” Let customers speak in their own words instead of handing them a keyword-filled script to copy.
Businesses can also add verifiable information in their replies, such as how to book, accessible entrances, and the scope of their services. This information also needs to remain consistent on the official website; reviews shouldn’t be forced to stand in for a complete description of the services.
Two cards asking for reviews sit on the counter. One says “Leave five stars for a free side dish,” and the other says “Why did you choose us today?” A customer rushing to pick up their child takes out their phone and leaves a sentence: “This is the only place willing to pack the broth separately, so the noodles still weren’t soggy when I got home.”