AI SEO Case Study: From 72 to 336 AI Mentions in Four Months

This AI SEO case study documents how PushLeads grew AI mentions across its managed local-business fleet from 72 to 336 over four months, and, more usefully, the repeatable method behind it. The headline number matters less than the process: a fixed prompt panel, answer-first content, consistent entity signals, and a steady review engine. Any local business can copy the method, so this focuses on what we did and why it worked rather than the result alone. The figure comes from our own internal tracking across the sites we manage.

We run this work from Asheville, NC, and treat our own fleet as the testing ground before recommending anything to a client.

Key Takeaways

• This AI SEO case study centers on method, since the 72 to 336 result came from a repeatable process.
• A fixed prompt panel made AI mentions measurable instead of anecdotal.
• Answer-first content and consistent entity signals drove most of the gain.
• Reviews and organic rankings moved alongside AI mentions rather than separately.
• The timeline ran in months, matching how SEO authority builds.

How We Measured AI Mentions

The first step was making AI visibility measurable, because you cannot improve what you do not track. We built a fixed panel of the questions customers actually ask AI tools about local services, then checked each one across Google AI Overviews, ChatGPT, and Perplexity on a set schedule. For each check we recorded whether a fleet business was named, whether a competitor appeared instead, and which sources the answer cited.

Counting mentions this way turned a vague goal into a monthly number. The starting count was 72 mentions across the panel. Four months later it was 336. Because AI answers vary day to day, we treated each check as one data point and judged progress on the monthly trend, an approach our note on content performance dashboards describes for any measurable SEO program. Our guide to AI Overview tracking covers why this method beats one-off spot checks.

Why AI Mentions Were Worth Chasing

We prioritized AI mentions because customer behavior has shifted. Roughly 58.5% of US searches now end without a click (Ekamoira analysis of Semrush data, 2026), which means a customer can read an AI answer, see a recommended business, and act without visiting a website. A mention inside that answer captures demand that a blue-link ranking alone would miss.

The overlap with traditional search made the work efficient. Google’s AI Overviews cite around five sources per query, and about 52% of those cited sources also appear in the top 10 organic results (Semrush, 2025). That told us the fastest path to more AI mentions ran through the same authority that drives rankings, so we did not build a separate program. Our overview of AI SEO for local service companies explains that connection in more detail.

The Content Changes That Moved the Number

The largest share of the gain came from rewriting content to be quotable. We opened key pages and sections with direct, 40 to 60 word answers to the exact questions on our prompt panel, then supported them with detail. AI models lift clear answers far more readily than buried ones.

We also leaned into specific, longer questions, since queries of eight or more words are roughly seven times more likely to trigger an AI Overview (QuickSEO analysis of WordStream data, 2026). Detailed how-to and comparison content aimed at real customer questions gave each site more citation surface. Because most AI questions are informational, about 88% of AI Overview queries by Semrush’s count (Semrush AI SEO statistics), explaining the work clearly mattered more than clever keyword targeting. Organizing that content into topic groups, as in our guide to SEO content silos, helped AI read each site as an authority rather than a set of stray pages.

Entity Signals and Technical Cleanup

The second driver was consistency. We standardized each business’s name, address, phone number, and service area across its website, Google Business Profile, and directory listings so AI systems could connect scattered mentions into one confident recommendation. We added or corrected schema so machines read the details as labeled facts, following the types in our local schema markup implementation guide.

Internal linking tied it together. Connecting related pages with descriptive anchors, using the approach in our internal linking blueprint for local SEO, helped both search engines and AI systems read each site as a connected whole. None of this was flashy, but the mentions did not climb until the entity signals were clean.

Reviews Moved With the Mentions

Reviews turned out to be a quiet engine behind the gain. As we improved review volume and response rates across the fleet, both rankings and AI mentions rose. That tracks with the research: BrightLocal found that about 75% of consumers always or regularly read online reviews, and 88% would use a business that responds to all of its reviews (BrightLocal, 2024), and review volume correlates with map ranking, where top-three positions average around 240 Google reviews (SurfSigma analysis, 2026). AI systems lean on the same review signals, so a stronger review engine lifted both at once, much like the process in our work on online reputation management.

The Timeline and What It Means

The four-month arc was not linear. The first month was mostly measurement and cleanup, with little movement in mentions. Gains accelerated in months two and three as content and entity work took hold, then compounded in month four. That curve matches how authority builds, and it lines up with what former Google spokesperson Maile Ohye described as a “four months to a year” window for SEO to show benefit (Maile Ohye, Google). Our SEO timeline guide frames why the early weeks look flat before results build.

The practical lesson from this AI SEO case study is that the method transfers. A single local business does not need a fleet to copy it: define a prompt panel, write answer-first content, clean up entity signals, and run a real review engine. The number will look different, but the process is the same.

What We Would Do Differently

Looking back, the biggest lesson was to start measurement earlier. We spent part of the first month setting up the prompt panel while content work was already underway, which made it harder to attribute early movement. If we ran it again, we would lock the panel and record a clean baseline before touching a single page, so every later gain had a clear starting point.

We would also front-load the entity cleanup. The mentions did not climb meaningfully until business details were consistent across each site, profile, and listing, so doing that work first would likely have pulled results forward. And we would publish original data sooner, since the pages with first-hand numbers earned citations faster than the ones with strong but generic content.

The through-line is that none of this required a bigger budget, only better sequencing. A single local business can apply the same order: baseline first, entity cleanup second, answer-first content and original data third, reviews running throughout. For owners planning their own version, our AI SEO strategy for small businesses guide lays out that sequence in a format built for one location rather than a fleet.

Frequently Asked Questions

What does this AI SEO case study actually prove?

That AI mentions can be measured and grown with a repeatable method. The 72 to 336 result came from a fixed prompt panel, answer-first content, consistent entity signals, and a stronger review engine, not from any single trick.

How were the AI mentions counted?

Through a fixed panel of customer questions checked on a schedule across Google AI Overviews, ChatGPT, and Perplexity, recording whether a fleet business was named. Because answers vary, progress was judged on the monthly trend.

Can a single local business replicate these results?

Yes, using the same method. Define your own prompt panel, rewrite key content to answer questions directly, standardize your business details, and build a review process. Your numbers will differ, but the process transfers.

Why did reviews matter in an AI case study?

Because AI systems lean on the same review signals that influence local ranking. As review volume and response rates improved across the fleet, both rankings and AI mentions rose together.

How long did the results take?

Four months, with little movement in the first month during measurement and cleanup, then accelerating gains as content and entity work took hold. That curve matches how SEO authority typically builds.

Is the 72 to 336 result typical for every business?

No single result is guaranteed, since outcomes depend on your market, competition, and starting authority. This figure came from our own fleet tracking and reflects the method more than a promise. What transfers is the process: measure with a fixed prompt panel, write answer-first content, clean up entity signals, and run a real review engine. Your numbers will differ, but the same sequence tends to move AI mentions in the same direction.