5 min read
Ask an AI who the best anything in Chicago is and you will get a confident, specific, well-written answer. It will name three or four businesses. It will not be a ranking of quality. It will be a ranking of legibility.
This is the part local owners find hardest to accept, and it is the part that makes the problem fixable. The business that gets named is rarely the best one. It is the one the system could describe with the most confidence.
What a Chicago AI answer is built from
A metro like Chicago makes the mechanism unusually visible, because the market is deep enough that the model has to choose. In a town with four options it names four. In a city with four hundred it has to discriminate — and the criteria it discriminates on are not the ones a customer would use.
Three things do most of the work:
- Geography that resolves. “Chicago” is not one place to a retrieval system. Lincoln Park, Pilsen, Hyde Park, Evanston, Naperville, the West Loop — these behave like separate markets, and a business that only ever says “Chicago” competes against everyone instead of winning a neighborhood.
- Corroboration that is local. A mention in a Chicago publication, a Chicago trade association, a Chicago-specific directory or roundup carries far more weight for a Chicago query than a national citation of the same size.
- Consistency across the listing graph. Google Business Profile, Apple Business Connect, Bing Places, Yelp, the chamber, the licensing board. Every mismatch between them is a reason for the system to hedge — and hedging means naming someone else.
Why the best business often loses
Consider two contractors on the North Side. One has been in business twenty-two years, does excellent work, and gets most of their jobs by referral. The other opened four years ago, is competent but unremarkable, and has a marketing person.
The referral business has: a site with a phone number in an image, a Google Business Profile claimed in 2019 and never updated, an old address still live on three directories, and nine reviews. Everything true about their quality lives in conversations no system can read.
The newer business has: consistent listings, a service page for each neighborhood they cover, prices in text, 140 reviews collected deliberately, and two mentions in local trade press.
The model names the second one. Not because it made a judgment about craftsmanship, but because it had enough resolvable evidence to say a sentence about them without hedging. From the outside that looks like luck. From the inside it is engineering.
The numbers underneath this
Two figures set the stakes, and both come with caveats worth stating.
SparkToro, using Similarweb panel data, found 68% of US Google searches ended without a click between January and April 2026, against 60.45% in 2024. SparkToro cautions that long-run comparisons across different data providers are not directly equivalent.
Source: SparkToro / Similarweb, 2026, via Search Engine Land.
Ahrefs found that only 38% of pages cited in Google AI Overviews ranked in the top ten for the query, across 863,000 keywords in March 2026 — with 31% coming from pages ranking beyond position 100 or not at all. Ahrefs attributes part of the change from their 2025 figure to their own improved parsing.
Source: Ahrefs, March 2026, via Search Engine Journal.
For a local business the combination is specific: fewer people are clicking through to compare, and the shortlist they are handed instead is assembled from sources that only partly overlap with who ranks. Being the local favourite is not evidence a machine can read.
Precision beats reach here. “Chicagoland” is not a place a retrieval system can resolve. Lincoln Park, Pilsen and Oak Park are.
What to actually do, Chicago edition
- Pick your real service area and say it in text. Not “Chicagoland” — the neighborhoods, suburbs and corridors you genuinely serve, named, on a page a system can read. If you cover Oak Park and not Aurora, say so. Precision beats reach here.
- Build one page per meaningful area, and make each one different. A page that swaps the neighborhood name into the same paragraph is worse than not having it. Local specifics — parking, building stock, permit quirks, typical job types, travel time — are what make the page worth citing.
- Reconcile every listing, then keep them reconciled. Set a calendar reminder. Listings drift, especially after a phone change or a move.
- Earn three local citations this quarter. Read the citation list off the AI answers for your own category and go after those specific sources. Neighborhood press, trade associations, chamber directories, and genuinely curated roundups.
- Ask for reviews on a system. Volume and recency both matter, and a steady trickle reads as more credible than a burst. Ask every satisfied customer, at the same point in the job, every time.
- Mark it all up. LocalBusiness schema with complete areaServed, Service schema per offering, and FAQPage where you have real answers to real questions.
Set expectations honestly
Listing consistency shows up in a matter of weeks. Local corroboration takes a quarter or more, because it depends on other people’s publishing schedules. Review velocity is a permanent habit rather than a project.
And no one can promise you a place in an AI answer. The systems change without notice, and the same question can return different answers to different people on the same afternoon. What can be committed to is a measured baseline, a defined scope, and monthly reporting on the leading indicators — citations, entity coverage, rank grid movement, and share of AI answers.
Related reading: how to find out if AI recommends your business and what actually gets you cited in AI answers.
Found First is a search and AI visibility agency based in Chicago, working Chicago, St. Louis and Kansas City. The AI Visibility Snapshot is free and does not require a call: ten real buying questions for your category and city, run across ChatGPT, Claude and Gemini.