A 3.3-star car wash just won the AI response.
When Annie Jackson, director of revenue operations and growth at GatherUp, asked Google for a contactless car wash suitable for an SUV in Norfolk, Virginia, Google returned a 3.3-star business and answered questions about permits and hours of operation above the star rating itself. Query match exceeded ranking.
This example of frames the session Jackson introduced Jason Wertham, VP of Review Defense Operations at GatherUp: AI tools gather their description of each location from reviews, listings, and public mentions across the web, then repeat that description to customers who never access your website.
Jackson and Wertham crossed the four-step emergency audit they race to surface what ChatGPT, Google AI Overviews and Ask Maps are saying today about a multi-location brand, then the build, manage, defend deployment this changes the answer.
The summary below covers the results. The prompts, deployment, and audit document are available in the full session. Watch it on demand.
How are customers using AI to find local businesses now?
They ask a complete question and accept the summarized answer. In GatherUp consumer data collected in fall 2025, 55% of consumers had viewed AI summaries from Google or Bing, 48% had asked ChatGPT about a local business, and 31% had asked repeatedly.
Jackson’s car wash query shows what this looks like in practice. Instead of “car wash near me,” the search was “contactless car wash for my SUV in Norfolk, Virginia.” Google leveraged its 300 million places and 500 million review contributors to return a business, with clearance height and 24/7 online response.
“Google answered my questions, but this company actually shows up with a 3.3 star rating,” Jackson said. “The context of my query has surfaced above the star rating.”
Wertham added that these tools now take into account who you are and when you ask. “The time of day that you actually perform that query in Google Maps could impact which businesses are returned in those results,” he said. An LLM that knows you own an SUV or a big dog applies that context to every future local query, whether you rephrase it or not.
Audit of the session begins with this change in behavior: four prompts, from a brand query to a spot check location by locationthat show you the answer your customers are already seeing. See the four audit prompts.
Do Google Reviews Power AI Responses?
Not from the listing itself: Google, Yelp, and other major directories prevent LLM crawlers from reading review content on business profiles. Reviews still support local rankings and on-list conversion; they only capture AI responses via surfaces that LLMs can explore.
“The major directory service providers, Google, Yelp and others, do not allow LLM tools like ChatGPT and Claude to scrape or crawl business listing review data,” Wertham said. “You’ll notice that they don’t cite specific reviews from these platforms.”
The same reviews become searchable as soon as you repost them. Post them on public social media or embed them in review widgets on your own site and, in Wertham’s words, “they’re now fair game for LLM tools.” »
This determines which queries you can win. When a customer asks for “popular” or “highly rated” companies, the LLM searches for review text it can access. Examination of the content confined to the directory does not contribute anything to this answer.
“If you rely on review platforms to do it for you, that’s not going to be enough,” Wertham said.
Action item: Repost your reviews where AI can crawl them. The session covers which social widgets and placements make review content readableincluding how to accompany the business response with the notice.
Wertham also broke down capturing first party reviewsthe survey responses that never make it to Google and why one customer’s 11,000 responses mattered. This segment lives in on-demand recording.
Does your star rating still matter in AI search?
This matters less than the recency and timeliness of the review. No response from the AI in audit examples of the session cited an average star rating; everyone cited the content of the reviews.
Consumer data supports this: 45% of users favor review recency over star ratings, 60% trust detailed written reviews over purely rated reviews, and 70% prefer a review request within 72 hours of the transaction.
Wertham noted that consumers regularly override Google’s default sorting of “most relevant” reviews and move to “most recent” because the most recent review predicts the experience they will get. A high average based on years-old reviews carries less weight than a current, consistent feed.
“I’d rather go to a company with 1,000 reviews and a 3.9 or 4.2 rating than 30 reviews and a 5.0 rating,” Wertham said.
The session organizes the response into three areas of work, constructs, manages and defends: creation of coherent lists and examination volumes, management of responses and monitoring within a 72 hour response windowand defending the grade you earned against counter-policy opinions and burying tactics such as see suffocation again. Walk through construction, management, and defense deployment.
Why does AI give a different answer about your business every time?
Since LLM answers behave like a slot machine, a query is never a reliable read. Jackson cited research from SparkToro in which different people asked LLMs the same question across all devices and accounts, and the results were never returned in the same order.
“Asking the AI a question is a bit like a slot machine,” Jackson said. “It’s going to return similar data, but each time it’s going to be a little different.”
Position is not a good metric for AI visibility. The total number of citations, the breadth of sources feeding into the response, predicts whether or not your brand will appear. Your brand can completely miss the response of one device and take the lead on the next.
Google published its own generative AI optimization guide and updated it this month. Wertham pointed out a change with bite: the AI Slope Penalty. Google now detects low-value AI-generated content and, in its words, “essentially penalizes businesses” for it. Generic AI blog posts and glorified FAQ scraping targets now cost you rather than being ignored.
Action item: Run your audit prompts in incognito mode or temporary chat mode so that stored context stops shaping your results, and rerun them on a schedule. The session shows the monthly re-execution method to measure whether your visibility work is changing the response.
Q&A: Most useful questions from the webinar
Q: What’s the quickest thing I can do this week to change what AI says about my business?
“Process your listings. Make sure your listings are all correct and consistent, no matter what platforms you’re on. And then make sure that you’re evangelizing your reviews from the third-party directory where you receive them. Post them on your social media platform, post them in a section of your website.” Jason Wertham, 49:23 in On-Demand Recording.
“Make sure you have the basics. Get the basics, make sure they’re set, and then you can move on to more elaborate things.” Annie Jackson, 50:30 in on-demand recording. Jackson illustrated this point with a local restaurant whose Facebook page listed the owner’s personal cell number; he never knew why the calls kept coming.
Q: How long does it take before content changes actually appear in AI responses?
Annie replied: small facts evolve quickly, positioning evolves slowly. Store hours and phone numbers are updated quickly, but “what you’re known for will take a little longer,” typically two weeks to a month, with a long tail beyond that. She emphasized that your own website is the quickest lever: a new offer must first appear on your own channels, because reviews will not announce it for you. Full response at 54:19 in on-demand recording.
Q: My weakest location has old bad reviews that continue to appear. Should I wait until they get older?
Jason replied: Age naturally dampens a review’s relevance, but keyword-heavy reviews and Local Guides reviews retain their rankings longer, and emoji reactions keep a review from slipping even when they don’t add any upward momentum. Reviews that violate policy remain questionable at any age; her review the defense The team regularly deletes reviews that are more than ten years old. The reliable solution lies in volume and speed, because recency trumps content in terms of long-term relevance. Full response at 47:33 in on-demand recording.
Q: How should franchisors handle this when each franchisee controls their own profile?
Jason replied: Franchise reputation management fails due to lack of consistency, as each franchisee owns their listing while the brand absorbs the AI response. Establish best practices, offer white label or partner tools for franchisees to use, and put a playbook in their hands. He also suggested running the audit prompts on behalf of franchisees and coaching them on the results, because one poor AI response from one location costs the entire brand. Full response at 55:46 in on-demand recording.
Watch the full webinar
The on-demand session contains the four emergency audit prompts and downloadable document, the full build, manage, and defend deployment, the monthly measurement method, the review defense walkthrough for challenging policy-violating reviews, and Jackson’s guide. AI Narrative Audit offer. Watch the full webinar on demand.
