FAQ

How long does GEO take to show results?

Two horizons: technical fixes (crawler access, machine-readable surface, consistent entity facts) show up in audit scores and crawl logs within 2–4 weeks; actually being mentioned and cited in AI answers typically takes 2–3 months or more, because it depends on accumulating third-party mentions and on engine re-crawl cycles. Any promise of being recommended by ChatGPT within two weeks is not credible — mention rates can only be verified by periodic sampling against a fixed question set.

Figure 01 · GEO's three clocks

  1. 01

    Technical surface · verifiable in 2–4 weeks

    Allow AI crawlers, ship llms.txt and Markdown mirrors, fix structured data, unify entity facts across the web

    Judge by: audit scores rise; GPTBot, ClaudeBot and PerplexityBot visits and success rates rise in server logs
  2. 02

    Content and entity · 1–3 months

    Question-shaped content from real buyer questions; one consistent version of company facts across the web; directory and third-party profiles completed

    Judge by: Q&A content appears in citation sources of relevant answers; entity-structure and machine-readability scores climb from the 30-site means (71.1 / 69.2)
  3. 03

    Mentions and citations · 3–6 months and up

    Press coverage, partner pages, reviews, rankings and community discussion accumulate over time

    Judge by: fixed question set, one consistent method, at least a quarter of periodic samples — read the trend line, not a single point
Ranges are Negency's judgement, not measured promises; they depend on starting point and investment. The flagship 30-site measurement covers website technical visibility only, not mention rates. Three clocks: engineering by the week, content by the month, mentions by the quarter. Data: Negency Flagship Report 2026 (Issue 1).

First, the headline: three phases, three clocks

GEO results do not arrive on a single date. They come in three phases on three clocks: the technical surface verifies in weeks, content and entity work accumulates over months, mentions and citations are judged over quarters. The ranges below are Negency's judgement — typical rhythms from our audit and diagnostic work, not measured promises; your pace depends on your starting point and investment. The 30-site flagship measurement covers website technical visibility only, not brand mention rates, so no 30-site data supports a mentions timeline — anyone promising mentions to the week should be treated with suspicion.

Phase one: the technical surface (verifiable in 2–4 weeks)

Allowing AI crawlers, shipping llms.txt and Markdown mirrors, fixing structured data and unifying entity facts across the web are deterministic engineering — done is done, and re-testing proves it. The evidence is not rankings: it is audit scores going up and AI crawlers (GPTBot, ClaudeBot, PerplexityBot) visiting more and succeeding more in your server logs. Each crawler's identity and behaviour is documented by its vendor — user-agent strings and robots controls are public — so "we allowed the crawlers" is checkable in the logs, not a matter of trust.

Speed here depends on the starting point. The Negency Flagship Report 2026 (Issue 1) — a sample of 30 Chinese exporters' global website home domains, measured with the Negency AI visibility audit engine (around 30 signals, five categories) in a single-day sample on 30 October 2026; the report measures website technical visibility only (whether AI can crawl and read the site), not brand mention rates in AI answers — found 17 of 30 sites with no JSON-LD at all, 15 of 30 with sitemap problems, and 13 of 30 with a missing or non-unique H1. Each is deterministic work finishable in weeks, and re-measurement shows the change the same day. For reference, Negency's own website scored 97/100 in an audit on 11 November 2026 — a finished technical surface.

Phase two: content and entity (1–3 months)

Once the technical surface is fixed, content and entity work follows: question-shaped content from questions buyers actually ask, company facts reconciled into one version across the web, directory and third-party profiles completed. This phase runs in months, for three reasons: content needs fact-checking time; aligning facts means changing and verifying them one place at a time; and engines re-crawl on their own cycles, which nobody controls.

Judge it by three signals: whether the question-shaped content is indexed and appears in the citation sources of relevant answers; whether the same fact now reads identically on the website, in directories and in the press; and whether entity structure and machine readability are climbing from the 30-site means (71.1; 69.2, median 68.5). Publishing is not the same as taking effect — after publication you wait for the next crawl cycle, and that wait cannot be skipped.

Phase three: mentions and citations (3–6 months and up)

Whom AI answers cite is driven mainly by accumulated third-party mentions — a variable your own website changes cannot move overnight. Press coverage, partner pages, reviews, rankings and community discussion take real time to accumulate, and engines refresh their indexes on their own cycles. Read this phase as a quarterly trend, not a weekly wobble.

Volatility is normal: AI answers are non-deterministic, so the same question asked twice can cite different sources, and any single sample is governed by engine sampling, not your performance. Judge results with a fixed question set, one consistent method, and at least a quarter of periodic samples — read the trend line, not a single point. Any report proving "results" with one screenshot is not worth believing.

Setting expectations

  • Month 1: technical surface fully fixed; baseline mention rate measured. Most companies start near zero — an observation from Negency's diagnostic experience, not a failure. Note the flagship 30-site data does not measure mention rates, so this judgement comes from our client work, not that report.
  • Months 2–3: content and entity layers live; third-party mentions start entering engine citation sets; your name appears for target category questions.
  • After one quarter: use sampling data to judge the trend and decide which engines and channels deserve more investment. Three to six months is the typical range for the mentions layer to form a readable trend, depending on starting point and investment — the lower the start and the thinner the spend, the longer the end you land on.

The three clocks mean: verify engineering in month 1, content in months 2–3, the trend at the end of quarter 1. Judging month 1 by mentions, or mentions by engineering speed, is using the wrong clock. Start with a free AI visibility audit to see where your technical surface stands before talking timelines.

Related questions

How is GEO different from SEO?

SEO optimises a page's rank in a list of links, aiming for clicks. GEO (Generative Engine Optimization) optimises how discoverable, understandable, trustworthy and citable a company is inside AI-generated answers — the goal is to be in the answer itself. They share technical foundations (crawlability, structured data, content quality), but GEO additionally requires governed entity facts, a machine-readable surface (llms.txt, Markdown mirrors) and a third-party mention ecosystem. A site that ranks well is not automatically visible to AI: engines cite the companies whose facts they can fetch, parse and verify. The concrete differences sit in four dimensions — goal, object, content structure and evaluation — compared item by item in the body below.

How can Chinese companies going global build AI search visibility?

For Chinese companies going global, the visibility gap is rarely content volume — it is the machine-readable surface. Our audit of 30 Chinese outbound companies found machine-readability the weakest category (69.2/100 average): 17 had no JSON-LD on the homepage, 15 had a missing or thin sitemap, 13 had a missing or duplicated H1. The path has four steps: audit bot reachability and machine-readability first; govern entity facts so Chinese and English statements agree; publish question-shaped content built on real overseas buyer questions; then earn third-party mentions (media, communities, partners) and verify with periodic resampling against a fixed question set.

What should a corporate website change for the AI search era?

Five changes. (1) Bot reachability: robots.txt explicitly allows AI crawlers and live fetches are not stopped by firewalls or challenge pages. (2) A machine-readable surface: an llms.txt fact file (optional aid), a complete sitemap, Markdown versions of core pages (optional). (3) Entity structure: JSON-LD naming entity types such as Organization, clean title/H1 discipline, hreflang pairing on bilingual sites. (4) Question-shaped content: pages organised around buyer questions, short answers first, FAQ schema attached. (5) Measurability: analytics that separate human and bot traffic and attribute AI-referred inquiries. Negency's own site was rebuilt to exactly this standard, and the public audit tool scores against it.

What is llms.txt, and should every company website have one?

llms.txt is a plain-text brief at the site root written for AI systems: who the company is, what it does, where the core pages live, and the canonical facts and definitions. It does not by itself raise citation rates — in Ahrefs' May 2026 study of 137,210 domains, 97% of published llms.txt files received zero requests that month, so its direct effect is near zero — but it is the cheapest fact-governance move available: one file anchors a single version of the truth that crawlers can rely on. Verdict: worth adding, useless alone; it works as part of entity-fact governance, structured data and question-shaped content. Negency's llms.txt is a full fact bible: company definition, offering boundaries, sourced cases and citation rules.

What should an enterprise prepare before deploying AI agents?

Four preparations, and we advise against starting without any of them. (1) Scenario selection: pilot on a workflow that is frequent, comparatively rule-clear and verifiable — not the most glamorous one. (2) Knowledge and data foundations: govern the knowledge base first (deduplication, versioning, permissions); a dirty knowledge base only scales errors. (3) Governance design: permission boundaries, human approval points and audit logs decided before launch, not after. (4) A measurement baseline: record the manual process's time and error rates before the pilot, or value can never be proven afterwards. Negency's method is a small controlled pilot, verified before scaling, with governance as a design input rather than a patch.