We ran a simple experiment: take twelve Chicago-area businesses founded between 1857 and 1930 - roughly 1,300 combined years of operating history - and ask an AI engine to recommend the best in each of their categories. Eleven of the twelve were not in the answer. Not ranked low. Absent.
These are not struggling businesses. They are institutions: a camera store on Wabash since 1899 that survived the Great Fire era, rebuilt after 2020, and still develops film. A wine shop pouring since 1888. A dry cleaner founded in 1857. A monument company that has served Chicago families since 1879. The kind of businesses whose longevity is the trust signal every marketer wishes they could buy.
The machines cannot see any of it.
What the engines rewarded instead
In each category we asked Perplexity the question a real buyer asks - "best camera store in Chicago," "best building materials supplier in Chicago," "best headstone company near Chicago" - and recorded the answer.
Reviews beat history, every time. A building materials supplier founded in 1888 was absent while the engine's top recommendation was credited with a 5.0 rating from thirteen reviews. Thirteen reviews outweighed 138 years, because reviews are machine-readable and history, left unstructured, is not.
The engines actively select on longevity - when they can verify it. The sharpest case: a monument company serving Chicago families since 1879 was skipped while the engine crowned a competitor and quoted its credential verbatim: "serving Chicago-area families since 1878." The engine rewarded exactly the quality the older business holds. It just could not see it.
Some category answers are simply unclaimed. For one 1883 architectural millwork manufacturer - a company that made ornament for Chicago's Gilded Age buildings - the engine named no company at all. It described types of suppliers generically. An entire category recommendation, sitting vacant.
Why this happens
AI engines do not rank pages; they synthesize answers from what they can identify, verify, and describe confidently. Their raw material is structured data, consistent business listings, review text, and third-party mentions. A century of reputation stored in customers' memories, framed newspaper clippings, and a line on the About page is invisible to that process.
The result is a quiet transfer: buyers ask answer engines who to trust, the engines name whoever is most machine-legible, and demand moves from the proven to the parseable. The businesses losing these answers never see it happen - the buyer who asks ChatGPT and calls the named competitor never appears in anyone's analytics.
The fix is smaller than the problem
Here is the genuinely hopeful part: everything the engines failed to see about these twelve businesses is fixable with ordinary work.
Make the history machine-readable. "Since 1899" needs to exist as structured data and consistent listing text, not just a line in a logo. Engines quote what they can verify.
Collect the reviews the reputation has already earned. A century-old customer base that was never asked for reviews is the most undervalued asset in local business. The engines describe you in your reviewers' words - give them words.
Answer the questions buyers actually ask. Cost questions, vetting questions, "who does X near me" questions. A direct, honest page per question is what engines cite - and in categories where the answer is unclaimed, the first mover usually takes it.
Corroborate everywhere. Consistent name, services, and story across every profile and directory. Engines trust agreement between independent sources; a century of history that only your own website mentions is, to a machine, a claim.
The window
Every one of these answers is re-synthesized continuously as the engines re-crawl. That is the threat and the opportunity in one mechanism: the businesses being skipped today can be named next quarter, because nothing about the current answer is permanent except the data underneath it.
Longevity deserves to count. It just needs to be translated into the only language the machines read.
Which businesses were in the study?
Twelve Chicago-area companies founded between 1857 and 1930, across camera retail, commercial printing, fine jewelry, monuments, musical instruments, funeral service, architectural millwork, building materials, wine retail, dry cleaning, moving and storage, and commercial HVAC. We shared each business's individual findings with them privately.
Does being old hurt a business in AI answers?
No - the opposite. Engines actively favor verifiable longevity and quote it when they can see it. The problem is that most century-old businesses have never translated their history into structured data, reviews, and citations the engines can read.
How do I find out if my business is visible in AI answers?
Ask the engines what your buyers ask - or have us do it. Our free audit runs your 10 to 15 real customer questions across ChatGPT, Perplexity, Gemini, and Google AI Overviews and sends you the transcript with a prioritized fix list.
Free audit: what each engine says about your business, who gets named instead, and the prioritized fix list. Delivered by a human in 2 business days.
Run My Free Audit