FAQ

Choosing a GEO provider: 12 due-diligence questions to ask first

GEO has no official leaderboard and its results lag by weeks, which leaves vendors a lot of room for talk. These 12 due-diligence questions — grouped into measurement, method, evidence and contract risk — each come with what a credible answer sounds like and the red flags to walk away from, so you can pin a GEO provider down before you sign.

Figure 01 · The 12 due-diligence questions at a glance

  • Set 1 · Measurement
  • 1Which question panel and engines, and how often?
  • 2How are mention and citation rates defined? Can we see raw answers?
  • 3What is our baseline, and how are targets set?
  • Set 2 · Method and limits
  • 4Which layers do you actually work on?
  • 5Can you guarantee placement in AI answers, ranked first?
  • 6What role does llms.txt play in your method?
  • 7How is content produced? Hundreds of generated articles?
  • Set 3 · Evidence and deliverables
  • 8Can we see your own website's GEO data?
  • 9What exactly is delivered, and how often?
  • 10How long until results, judged by what?
  • Set 4 · Contract and risk
  • 11How is pricing structured? Can we pilot first?
  • 12Who owns the data, accounts and content assets? How does exit work?
  • Red flags

    Any one is an automatic disqualifier

    Guaranteed placement or ranking; refusal to show raw samples; method kept secret; a single file or tag sold as the secret; bulk content farming; reporting that never mentions a baseline

  • Green flags

    Tie-breakers when two firms are close

    Fixed panel and definitions; raw answers inspectable; their own website as the first sample; layered timelines; limits stated up front

12 questions in four sets: measurement (3), method and limits (4), evidence and deliverables (3), contract and risk (2). Any single red flag is an automatic disqualifier.

Why due diligence comes first

GEO is a new category — no official leaderboard, no shared standard, results lagging by a sampling cycle: ideal for sales talk. Judge a provider by its method, data and stated limits — not by its self-description. Twelve questions follow — each with a credible answer and a bad-answer signal. Mismatched answers: do not sign yet.

Set 1: Measurement

1. Which question panel and engines do you measure on, and how often? Credible: fixed question list, engine list and cadence, compared over time. Red flag: the questions change every report, or proof is a single screenshot.

2. How exactly are mention rate and citation rate defined? Can we see the raw answers? Credible: definitions written down, applied consistently, raw sampling records available. Red flag: one composite score, definitions going fuzzy when asked.

3. What is our baseline, and how are targets set? Credible: baseline first; targets layered by crawling, mentions, citations. Red flag: a promise to double results before any baseline exists. Note: Negency's flagship 30-site data (30 Chinese exporters' global website home domains; audit engine, around 30 signals, five categories; single-day sample, 30 October 2026) measures website technical visibility only, not mention rates — a technical score is not a mention-rate baseline.

Set 2: Method and limits

4. Which layers do you actually work on? Credible: crawl reachability, machine readability, entity consistency, citation ecosystem, each with its own deliverables. Red flag: one vague "proprietary playbook".

5. Can you guarantee our brand appears in AI answers, ranked first? Credible: no — a citation in a natural answer cannot be bought; outcomes depend on engine sampling, so the work raises probabilities and verifies with data. Red flag: guaranteed placement or ranking, fees per "ranking achieved". Google's docs confirm there are no additional technical requirements for AI Overviews or AI Mode.

6. What role does llms.txt play in your method? Credible: an optional aid, not the core, not a pass — and they can state Google's official position. Red flag: llms.txt or some "AI-only tag" sold as a secret weapon. The proposal is public at llmstxt.org — a convention, not a standard.

7. How is content produced? Will hundreds of generated articles land on our site? Credible: content built page by page around real customer questions, with fact-checking and standards. Red flag: hundreds of auto-generated articles a month.

Set 3: Evidence and deliverables

8. Can we see your own website's GEO data? Credible: their own site first, with measured mention, citation and crawl records. Red flag: polished client screenshots only, their own data always confidential. Negency's own website scored 97/100 in an audit on 11 November 2026 (performance and basics 80, homepage response 3.08s) — our own sample first.

9. What exactly is delivered, and how often? Credible: audit reports, fix lists, sampling scorecards, re-measurement comparisons; samples available, cadence fixed. Red flag: the only deliverable is a monthly "work summary".

10. How long until results, judged by what? Credible: layered timelines — engineering in two to four weeks, mentions and citations as quarterly trends — naming the engine sampling cycle as a variable nobody controls. Red flag: "results in 7 days". The full three-clocks explanation: How long does GEO take to show results?

Set 4: Contract and risk

11. How is pricing structured? Can we pilot first? Credible: diagnostic, pilot and ongoing service separated, with pilot acceptance metrics. Red flag: annual-only contract, full prepayment, a pilot with no acceptance criteria.

12. Who owns the data, accounts and content assets? How does exit work? Credible: sampling data, content and accounts belong to the client; handover checklist on exit. Red flag: data locked in their system, deleted when the contract ends. Put it in the contract.

Red flags and green flags

Red flags: guaranteed placement or ranking, refusal to show raw samples, a secret method, one file or tag sold as the secret, bulk content farming, no baseline. Green flags: fixed panel and definitions, inspectable raw answers, their own website as the first sample, layered timelines, limits stated up front.

After the questions

Three buckets. Disqualifiers: guaranteed ranking, refusal to show raw samples, data not belonging to the client — any one ends it. Negotiables: price, cadence, pilot scope — write agreed terms into the contract. Tie-breakers: their own website as the first sample, limits stated up front, published samples — they settle close calls.

Negency's positions are public on this site: What is GEO?, How do you measure a brand's visibility in AI answers?, Going Global from China: How to Choose an AI Consulting and GEO Partner.

Pick the kind first

Settle the kind first: four kinds of GEO provider, compared puts all four on one yardstick. Weighing software against a service? Start with GEO tools vs a GEO agency.

See where your brand stands in AI search

Start with the free AI visibility audit: your scores for crawler reachability, machine readability and entity structure in minutes. For the full competitor comparison and fix list, see the AI search visibility diagnostic package.

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.