AI Does Not Recommend the Best Business. It Recommends the One That Got There First.
By Melinda Starbird — March 4, 2026
I spent my career as a Chief Diversity Officer because I have never been able to look at a system that excludes people based on access rather than merit and call it acceptable. I have seen it in hiring. In pay. In who gets promoted and who gets looked past. The pattern is always the same: the people with resources navigate the system and the people without resources get told the system is neutral.
AI recommendations are the same pattern. Different domain. Same injustice.
The businesses AI recommends in your category right now are not there because they are better than you. They are there because their business data exists in the sources AI reads, in the format AI trusts, with the consistency AI requires to make a confident recommendation. Quality is not the input. Data structure is the input. A mediocre business with complete structured data gets recommended over an excellent business with none, every single time, because AI cannot evaluate what it cannot read.
The SOCi 2026 Local Visibility Index analyzed 350,000 business locations and found that fewer than half of brands leading in Google local visibility also appear in AI results. Google ranking and AI visibility are different systems with different inputs. A business can be number one on Google and invisible to ChatGPT. Your competitor showing up in AI recommendations while you do not is almost certainly not a quality gap. It is a data gap that has nothing to do with how good your business is.
GMBapi’s analysis found that over 70 percent of ChatGPT’s local business results come from Foursquare, not Google Maps. Search Engine Journal confirmed that almost no small business owners know this — most optimize exclusively for Google Business Profile while ignoring the actual sources AI uses. If your competitor claimed their Foursquare listing and you did not, they get recommended. Not because they earned it. Because they are in the right database.
BrightEdge’s weekly citation analysis found that 96.8 percent of domains cited in AI results saw zero change week over week. Of the rare changes, 87 percent were declines. Only 0.4 percent of all tracked domains gained new citations in any given week. The businesses that got into AI recommendations early are building a compounding structural advantage. Every recommendation they receive makes the next one more likely. The businesses locked out are falling further behind with every week that passes.
My parents’ inn in Cle Elum — a century-old property with four generations of genuine hospitality behind it — lost 40 percent of its bookings because a newer property’s data happened to land in the right database. Not because the newer property is better. Because the system that now mediates discovery does not know how to measure a hundred years of family service. It only knows whose data it can read.
I built MiddleVerse because quality businesses deserve to be in the conversation. The system that excludes them based on data access rather than merit is the same system I have been fighting my entire career. The domain changed. The work is the same.