At nine in the evening, a pipe starts leaking. You tell your phone: “Find a plumber nearby who can come now and will give me a quote first.”
There is no brand name in that sentence, nor the words “recommended plumbers.” It is actually a miniature brief: distance, timing, service capabilities, and one condition for doing business.
Suppose there is an experienced tradesman just around the corner who does good work and charges transparently, but whose website says only “professional service, honest business practices.” Another business is a little farther away, but clearly lists its nighttime service areas, call-out fees, and quoting process. If the voice assistant recommends the latter, did the former lose on skill—or evidence?
I tend to see this as a thornier problem than search rankings: the more convenient voice assistants become, the more likely the power to choose in local markets is to pass first into the hands of whoever speaks on your behalf.
Reading Out Fewer Names Is Also an Allocation of Power
We are used to believing that the better search understands people, the more opportunities small businesses will have. After all, consumers can finally say “a craftsperson willing to repair old furniture” instead of having to search only for “furniture stores.” This could indeed bring specialized expertise to the surface.
But understanding demand does not mean presenting the full range of supply.
A screen lets people scroll down, open tabs, and compare different sources. If a voice interaction reads out only one or two businesses, users can easily follow up with: “Then contact the first one for me.” The length of the recommendation list itself affects who gets a chance to be considered.
This is also the tension I care more about when reading news framed around “websites restricting AI agents”: websites have reasons to protect their content, customer relationships, and system costs; assistants also need information to get things done for people. But once information cannot get through, what consumers hear about “the options nearby” may be only the portion the assistant managed to access.
To the ear, “I couldn’t find any” and “there aren’t any here” are often only half a second apart.
What Businesses Need to Add May Not Be More Adjectives
For local businesses and marketers, rewriting copy as conversational questions and answers may help, but it still does not reach the heart of the issue.
“We understand your fur babies best” sounds warm, but it cannot answer: “Does this grooming salon take elderly dogs who are afraid of hair dryers? Can the grooming be done in stages? Is an assessment required beforehand?”
What truly makes value recognizable is specific service conditions, verifiable information, and the limits of what is promised. Addresses and opening hours need to be consistent, and special requirements need to be explained clearly; information about services a business cannot provide has value too. These details can increase the chance of being understood, but they cannot guarantee a recommendation.
And that is precisely the problem: must people who are good at their craft first learn to organize evidence for machines? Organizing information costs money and effort, and chain brands with more resources are usually better equipped to keep it up to date. If a recommendation system directly equates “complete information” with “better service,” clues originally meant to support judgment may become another barrier.
From a policy and regulatory perspective, I am more interested in pressing the question of the recommendation interface’s responsibilities: does it let people distinguish paid exposure from ordinary recommendations? When information cannot be found, does it explain its limitations? Can users easily ask for more options? These design choices will actually change where a piece of business goes.
The phone finally reads out the name of a plumbing business. You ask: “Is there anyone closer?”
The metal shutter of the tradesman around the corner is still open, and a nighttime service number is posted on the wall. He is carefully writing out each call-out charge for his previous customer.