GAC logo

Case 03 · Automotive retail · 🇷🇺 St. Petersburg, Russia · 6 weeks

GAC auto retailer — 9 articles, 9,042 reads, named in 73 of 160 AI prompts

GAC · 🇷🇺 St. Petersburg, Russia

Six weeks of cost calculations, comparisons and buying guides brought 9,042 reads. In the 5 October 2026 measurement, AI engines named GAC in 73 of 160 buyer prompts (45.6%).


Engines lifted

  • ChatGPT
  • Gemini
  • Google AIO
  • Google AI Mode
  • Perplexity
  • Copilot

Before → After · 6 weeks

Before

No baseline measurement

The 160-prompt panel was not measured before the start, so the case has no before-and-after coverage comparison.

After · 6 weeks

9,042 reads · named in 73 of 160 prompts

9 articles published, 6 of them cited by AI systems. 7,250 reads from the top 4 articles alone. Now appearing as the default recommendation for the city's auto-buying queries.

Price$1,497/mo · Growth plan
Timeframe6 weeks
Articles9
BeforeNot measured
AfterNamed in 73 of 160 prompts; 9,042 reads

Price is today’s list price of the plan whose article allowance covers this scope, not the client’s contract. All plans

“Now ChatGPT recommends us by name.”

— Head of Marketing, GAC, St. Petersburg

The numbers in detail

What moved — GAC


160

Prompts in the panel

Regional buyer prompts, Hermes Project 35.

9,042

total reads · 6 weeks

Across the 9 published articles.

7,250

reads from top 4

Concentration in the best-performing formats: calculations, comparisons, and 'should I buy?' guides.


6 / 9

articles cited by AI

Two-thirds of the editorial set entered the AI answer layer.


Method

How this was measured

When5 October 2026, Hermes run #504. A later run on 12 November covered only two engines.
Base160 fixed buyer prompts for St. Petersburg auto retail. The figure of 73 of 160 counts prompts where at least one engine named GAC.
HowEach prompt went to 10 engines: ChatGPT, Gemini, Google AI Overviews, Google AI Mode, Perplexity, Copilot, Yandex Neuro, Yandex AI Mode, Alice and GigaChat. 1,536 of 1,600 answers came back. Perplexity returned nothing for 60 prompts.
LimitsThere was no measurement before the work began, so this page makes no growth claim. 1,434 of the 1,600 answers in run #504 came from earlier collection on the same prompts. Not all of them date from 5 October. On 12 November, run #886 on Alice and Yandex AI Mode only named GAC in 5 of 160 prompts. Almost all of its answers came from the lighter Yandex model, so it cannot be compared with run #504.
What you can checkThe 73 of 160 figure is shown as a count over a total, and the 12 November run is named here with its own count.

Section 01

Key takeaways

  • Built a panel of 160 commercial prompts for St. Petersburg auto retail.
  • Tested one hypothesis: editorial format matters more than editorial volume.
  • Shipped 9 articles in 6 weeks — calculations, competitor comparisons, "should I buy?" guides.
  • Reached 9,042 article reads.
  • Top 4 articles drove 7,250 reads — concentration in chunk-extractable formats.

Section 02

Why this case matters

Auto buying still happens at the dealership, but the research starts on the phone. In 2026 that phone increasingly runs an AI assistant. If a dealer is not in the AI's recommendation list, the local customer never walks in.

GAC had a strong on-the-ground reputation in St. Petersburg but zero presence in the AI answer layer. Competitors were already being recommended by name when buyers asked for the best dealer in the city.


Section 03

What we did, step by step

  1. Picked nine formats AI systems extract most readily. Calculations, competitor comparisons, and "should I buy?" guides — each format has the structure AI needs to answer "who should I choose?" questions.
  2. Structured every article for chunk extraction. Direct answers near the top. Question-led headers. FAQs. Self-contained paragraphs that survive retrieval without surrounding context.
  3. Published one topic per week for six weeks. Each topic mapped to a real St. Petersburg buyer query, not a content-calendar idea.
  4. Measured against the prompt panel. Tracked which articles entered which AI answers, by week, by engine. Format-driven outperformance was visible by week three.

Section 04

What this changes for the operator

Three lessons that rewrite how a local auto dealer should think about content:

  1. Format trumps volume. Six of nine articles entered the AI answer layer. The top four articles drove 7,250 of 9,042 reads. Calculations and comparisons outperformed everything else.
  2. Local commercial intent reads everywhere. The same articles got cited by ChatGPT, Yandex Neuro, Alice, and Gemini — buyers research in whichever engine they default to.
  3. The result shows at prompt level. In the 5 October 2026 measurement, AI engines named GAC in 73 of 160 prompts (45.6%), with no paid distribution.

Section 05

Source


Want a case like this?

The strategy call shows your closeable gap. The audit shows your AI mention rate.

Free thirty-minute strategy call.

An engineer from our AI search team runs your brand through Hermes before the call. You see exactly where each engine names you today. We tell you whether the loop that ran for this case applies to yours.