What a GEO audit shows
The audit captures the current visibility picture in generative engines across three layers. One — citability: for 20 niche queries we check whether your brand is mentioned and which competitors appear; the result is a "query → engines → named companies" table. Two — content signals: whether key pages have self-sufficient answers, dated facts, consistent descriptions of the company and services. Three — the technical layer: robots.txt (GPTBot, YandexBot, PerplexityBot access), llms.txt presence, server response speed for bots, Schema.org correctness. The output is a clear map: where you lose mentions and which steps bring you back into answers.
How the testing runs
Queries follow a formula: 8 transactional ("order CRM development"), 6 comparative ("best website support agency"), 4 question ("how much does promotion cost"), 2 branded. Each query runs through ChatGPT, Yandex Neuro and GigaChat in a fresh session — answers are captured with screenshots and dates. We also test phrasing variants: models are sensitive to rewording, and it matters which variants surface you. All raw data ships with the report: you can re-verify any answer yourself. A follow-up measurement in 2–3 months shows dynamics under identical conditions.
The technical layer: llms.txt and crawler access
Many sites lose citability at the technical stage: the AI crawler cannot reach the content. We check robots.txt: whether GPTBot, ClaudeBot, PerplexityBot and YandexGPT crawlers are blocked (a common "just in case" legacy). We check llms.txt presence and quality: the model-facing map of key pages. We review server response speed and stability for bots, aggressive antibot walls on public sections, meta tags and entity markup correctness. Every finding gets a priority and a fix instruction — most fixes take your or our developer a single day.
The content layer: why you are not cited
If tech is fine, the causes are usually content-related. Four typical ones: no self-sufficient answers (beautiful copy, but the answer to a specific question must be assembled across paragraphs); facts without dates and numbers (models prefer verifiable data); inconsistent entities (the company described differently across pages and directories); missing definition pages and FAQ (nothing to cite for question queries). The report shows which causes apply to you and gives a prioritized plan: which pages to rewrite first, which blocks to add, where an answer lead is needed. The plan is self-sufficient: execute it in-house or order it from us with the audit cost credited.
GEO audit cost and timelines
The base audit — from 18 000 RUB, 5 working days: 20 queries × 3 engines, a citation report, tech and content layers, a plan of 10–15 actions. Extended (40 queries, 5 engines with Perplexity and Copilot, analysis of top-cited domains, 3 competitors reviewed) — from 35 000 RUB, 7–8 days. A follow-up dynamics measurement — from 9 000 RUB. When ordering GEO promotion (from 60 000 RUB/mo), the audit cost is credited to the first month. The deliverable: a document plus a 40-minute walkthrough session — what to do, in what order, with what expected effect.
Why order an audit from ZXY
We practice what we audit: our own site is optimized for generative engines — test its citability in ChatGPT right now; the methodology is proven on ourselves. The audit is manual, in fresh sessions, not a "script checker": engines change daily, a live test is the honest one. Our plan stays SEO-compatible: the fixes strengthen classic SERPs too, so the budget is not split. And we stay after the audit: we can execute the plan, run monthly monitoring and repeat measurements — AI visibility becomes a managed metric, not a lottery.
