AI local discovery is moving the plumber, the dentist and the bistro from the Google listing into the AI answer, and the pitch landing on local publishers is to become the agency that gets small businesses found there. Local newspapers ran that exact play a decade ago, reselling Google and Facebook to SMBs, and the platforms kept the margin. The asset that matters this time is the one Duda's pitch leaves unpriced: the local editorial page the answer is built from.

  • Consumer use of AI tools for local recommendations rose from 6% to 45% in a year, and Google's share for the same task fell from 83% to 71%, per BrightLocal's February 2026 survey.
  • The publisher-as-agency model sells websites and reporting on a service fee; the open-web precedent is LocaliQ-style reselling, where newspapers earned thin margins on the inventory that counted.
  • Business mentions, meaning editorial and press pages, are the second-largest source behind ChatGPT's local answers and rise to 35% on 'best X in town' prompts.
  • Local publishers already own that inventory. Selling presence on their own cited pages is media revenue with publisher margin; selling websites is a reseller fee.

What AI local discovery is

AI local discovery is the process by which an assistant such as ChatGPT, Gemini or Google AI Mode recommends a nearby business by synthesizing business websites, directories and editorial mentions into one answer, instead of returning a map pack of listings. The business is recommended, not ranked, and the consumer often never visits a results page.

The scale is no longer hypothetical. SOCi's Local Discovery Index of September 2026 found 52% of consumers had used an AI tool to find a local business in the previous 30 days, against 9% in 2025. BrightLocal's February 2026 survey of 1,002 US adults put annual use at 45%, up from 6%.

The Google listing economy vs the local citation economy

In the listing economy, a local business got found by claiming a profile, collecting reviews and buying the ad above the map pack. Publishers monetized that world by becoming resellers: Gannett folded ReachLocal into LocaliQ in 2018, GateHouse built UpCurve, and Ken Doctor projected SMB digital services as a two-billion-dollar line for newspapers. The publisher's own pages were an afterthought in the bundle.

In the citation economy, the business gets recommended when enough trusted sources agree on it, and one of those sources is the local publisher. The unit of value moves from the slot the publisher resold to the page the publisher already owns. The parallel holds on both sides: the SMB still needs calls and leads, the publisher still needs a revenue line, but the margin now sits on the citation rather than on the resale.

What Editor & Publisher and Duda argued, and what we verified

In a sponsored Q&A published by Editor & Publisher on November 1, 2026, Duda's Peter Edquist argues that local advertisers are time-poor, that AI assistants reward trust and accuracy across sources, and that publishers should package websites, campaigns and AI-visibility reporting into one coordinated offer, with Duda as the white-label platform behind it. The relationship, he says, is the publisher's advantage.

The piece offers no market data, so we verified the premise independently. Duda's 20,000-agency figure matches its own 2026 releases. BrightLocal's August 2026 analysis of 1.97 million AI citations across 1,355 locations found business websites account for 42% of citations and Google Business Profile 28.5%, which supports Edquist's accuracy argument. His assumption that the relationship is the asset is where the data points elsewhere.

Why the publisher-as-agency model repeats the reseller trap

The strongest case for selling local businesses the services desk

Edquist's case is real on three counts. SMBs want outcomes, not technology, and the publisher is often their only marketing contact. AI visibility is scarce: SOCi's 2026 Local Visibility Index found inclusion in ChatGPT local recommendations roughly 30 times harder than ranking in Google's local results, and fewer than half of Google local leaders appear in AI recommendations. And BrightLocal's August 2026 study found a business that appears once reappears in only about half of later attempts, so there is demand for someone to manage the churn.

Why a service fee on the website leaves the answer unpriced

The services desk earns a fee on the input and nothing on the influence. Newspapers learned this once: in a 2018 Nieman Lab interview, GateHouse's UpCurve chief said marketing-services margins ran from 2 or 3 percent to 80 percent depending on the product, and the low end was the resold platform media that actually drove results. The website is the listing of the agentic web, necessary and undifferentiated. The recommendation is built elsewhere.

Where it is built is measurable. BrightLocal's study of 800 ChatGPT local searches, with data collected in November 2024 and refreshed in November 2025, found business websites supplied 58% of sources, business mentions 27% and directories 15%, and on informational prompts such as 'best Italian restaurant in' the mention share rose to 35% against 43% for business websites. A separate 9,434-answer home-services sample from Cheers in September 2026 found ChatGPT cited the editorial aggregator Expertise.com in 32% of answers. Those mention pages are the local publisher's product, and the agency bundle does not price them.

What this means for local advertisers and for local publishers

For local advertisers and their agencies: the website fixes the facts, the mention wins the recommendation

  • Treat the website and listings as hygiene: they resolve name, hours and address, which is what Duda's accuracy argument covers.
  • Treat editorial presence as the lever: track which local pages appear in 'best X in [city]' answers across ChatGPT, AI Mode and Gemini, and buy presence there.
  • Budget for the answer, not only the listing ad. The Google-above-the-map spend reached a human on a results page; the AI recommendation reaches a human who may never see one.
  • Expect volatility and price it: with roughly 50% persistence per attempt, paid presence on cited pages is the only stable line.

For local publishers: sell the cited page, not the reseller margin

  • Run the services desk if it retains advertisers, but do not confuse a fee on inputs with a media business.
  • Inventory your local guides, reviews and best-of pages by how often AI engines cite them for commercial local prompts; that list is your rate card.
  • Sell presence on those pages to the SMBs you already know. The relationship is the distribution; the citation is the product.
  • Keep the margin: a native placement on your own cited page carries publisher economics, a resold Google or Yelp campaign carries the platform's.

Three signals that the local answer is becoming a market

First, directories are being licensed straight into the answer: Yelp announced on July 23, 2026 that its 330 million reviews, ratings and listings now feed ChatGPT, with a Request a Quote action inside the chat, which turns the directory layer into a platform transaction the publisher never touches. Second, OpenAI opened a self-serve Ads Manager in July 2026 and is testing ads that open a business agent instead of a landing page, so the demand side of local AI has a buying pipe. Third, inclusion in local answers remains a coin flip per attempt, and instability is what creates a paid market. Whoever packages the cited local page as buyable inventory first sets its price.

Conclusion

Hold on to this: AI local discovery turns the local publisher from the SMB's reseller into the SMB's source, and the reseller model prices the wrong asset. Smalk AI's answer is Generative Engine Advertising, native ads for AI agents placed on the publisher pages AI engines cite, so local advertisers buy presence where the recommendation forms and local publishers earn media revenue from the pages they already own. Watch what happens after Yelp's Request a Quote: the moment a local AI answer completes the transaction, the publisher page that shaped it either carries a price or carries nothing.