GEO Tools: What They Do and How to Judge Them

GEO tools look alike on their landing pages, but they measure very different things. Sort any tool you're evaluating into one of five groups before you look at price. If you need the term itself defined, start with what generative engine optimization means.

Many products bundle several of these. That's fine. What matters is which features carry the price and whether those features rest on evidence.

Every AI answer passes through four stages. Knowing which stage a feature touches tells you whether it can work.

1. Parametric. What the model knows with no web access, frozen at a training cutoff. Nobody changes this in a quarter.
2. Access. Whether a crawler may fetch and store your page. You and your CDN control this.
3. Retrieval. Whether your page enters the model's context for a given query. The engine decides, per query, and often rewrites the query first.
4. Generation. Which retrieved pages the model uses and cites. The model decides.

Here's the catch. Nearly every GEO technique with a published effect size was measured at stage 4, on pages already placed in the model's context. A critical survey of 45 GEO studies (arXiv, July 2026) concluded that "already-retrieved content can causally alter its citation or use, but no reviewed technique shows a stable, longitudinal, cross-platform causal effect on organic discoverability or downstream behavior."

So a tool that sells stage 4 tweaks as a way to get found is selling a result its own evidence never measured. A tool that checks stage 2 is checking something you can fix and verify.

These features show up across the market. Each one rests on a claim that primary sources contradict or never support.

Schema and FAQ schema boosts

Google says structured data isn't required for generative AI search. Ahrefs' matched-control test of 1,885 pages found AI Overviews citations fell 4.6% after schema was added, with no major uplift on any platform. Google deprecated FAQ rich results in May 2026. Two honest caveats: Google says existing markup does no harm, and every page in the Ahrefs test already had heavy AI Overview citations, so whether schema helps a page enter the set is untested. Neither caveat makes schema a lever.

llms.txt generators

Google says Search ignores llms.txt. No other engine documents reading it. A tool that scores you down for lacking one is grading you on a file nobody has shown is used.

AI rank or position tracking

Google's generative AI report in Search Console shows impressions only, with no clicks, CTR or position. There is no ground truth to check an "AI rank" against, so any such number is the vendor's model, not a measurement.

E-E-A-T and domain authority scores

Google says E-E-A-T "isn't a specific ranking factor", so there is no score to improve. Domain Authority is a Moz metric that no engine computes.

"Up to 40% more visibility"

This figure traces to the original GEO paper, which tested a lab engine the authors built, with sources already supplied in context. It says nothing about getting found on live engines.

AI-friendly writing and chunking

Google rejects both and says there is no ideal page length. The evidence on rewrites versus access fixes points the same way: C-SEO Bench (NeurIPS 2025) found most conversational SEO methods largely ineffective and often harmful to ranking.

The checks that hold up are mostly subtractive. They find what blocks you, rather than adding something that promises a lift.

Per-bot robots.txt

robots.txt rules apply bot by bot. OpenAI states that "each setting is independent of the others": GPTBot governs training, and OAI-SearchBot governs ChatGPT search. A good tool reads each rule and reports its documented consequence. This combination blocks training but keeps ChatGPT search eligibility:

User-agent: GPTBot
Disallow: /

User-agent: OAI-SearchBot
Allow: /

Firewall and CDN reachability

A firewall or CDN can block a crawler that robots.txt allows. OpenAI's help centre says eligibility needs both: allow OAI-SearchBot and confirm your host or CDN allows its published IP addresses. From 15 September 2026, Cloudflare blocks Training and Agent crawlers by default on ad-displaying pages for new domains and existing free-tier customers. A site can lose access without anyone editing it.

JavaScript rendering

Vercel found that ChatGPT and Claude crawlers fetch JavaScript files but don't execute them. Content that only appears after JavaScript runs may be invisible to those crawlers. Googlebot renders, so this gap is mainly a non-Google problem. A useful tool diffs raw HTML against the rendered page and lists what's missing.

Snippet and index controls

noindex, nosnippet, max-snippet:0 and data-nosnippet keep content out of AI answers. Google says a page must be indexed and eligible for a snippet to appear as a supporting link in AI Overviews or AI Mode. Templates often apply these by accident.

Accuracy, platform by platform

AI platforms get specific facts wrong about specific brands: prices, product names, ownership. A tool that shows which fact is wrong, on which engine, gives you a concrete fix target. Per-platform reporting matters. In Otterly's 2026 experiment, Microsoft Copilot produced 2,718 of the 3,934 citations tracked for one test site, while ChatGPT, Gemini and Google's AI surfaces almost never cited it. A blended score would have hidden that.

The AI Crawler Checker requests a page as each AI crawler and shows which ones your firewall turns away.

Be clear about this before you pay for anything.

In your control:

• Which bots your robots.txt allows, bot by bot.
• Whether your firewall or CDN lets those bots through.
• Whether key facts are in the HTML before JavaScript runs.
• Snippet and index directives on your pages.
• Stating specific facts (a price, a number, a one-line definition) on fetchable pages. That gives an engine a sentence it can quote, and it's the only positive experimental signal for on-page content.
• Being named on the third-party sites engines cite. The evidence shows a strong association here, not a promised effect.

Not in your control:

• What the model learned before its training cutoff.
• Whether an engine searches the web for a given question at all.
• How the engine rewrites the question into search queries.
• Where your page lands in the model's context and whether it gets cited.

Any tool that promises citations, rankings or a score is promising things in the second list. Treat that as a warning sign.

Run the checks in the order the stages run. A later fix can't help if an earlier gate is shut.

1. Check crawler access per bot. Read robots.txt for OAI-SearchBot, PerplexityBot, Claude-SearchBot and Googlebot. Then test each crawler against your firewall.
2. Check rendering. Confirm your pricing, product names and core claims appear in raw HTML.
3. Check index status engine by engine. Look at Google Search Console and Bing Webmaster Tools separately. One being healthy tells you nothing about the other.
4. Audit accuracy per platform. Find which engines state wrong facts about you, and which facts.
5. Fix your own pages. Remove contradictions between your pricing page, about page and footer. State the correct facts plainly.
6. Then add tracking. Prompt tracking is worth having once the gates are open. Judge it by whether it reports each engine separately and discloses its prompt set.

Many of these checks are free. Free GEO tools cover most of steps 1 to 3, and the free checker on this site shows a snapshot of whether AI engines name you for three buyer questions. When you do pay, compare features against what GEO and AI visibility tools cost per month.

• Buying a single score. Any one AI visibility score averages answers given with and without live search, across engines that behave differently. Ask for the per-platform numbers underneath.
• Trusting an AI rank. No position data exists for Google's AI surfaces to check it against.
• Paying for generators as levers. Schema and llms.txt generators are fine utilities. They are not ways into AI answers.
• Blocking all AI bots at once. A blanket block on AI crawlers can remove ChatGPT search eligibility when you only meant to stop training.
• Reading a citation as influence. OpenAI exposes retrieved and cited sources separately, and the retrieved set is larger. Citation counts are a noisy proxy.
• Reading prompt-tracker swings as cause and effect. Engines change models and retrieval behaviour often. A drop after a page edit may have nothing to do with the edit.
• Falling for injection tricks. Some tactics hide instructions or cloaked text for AI crawlers. Research shows such content can sway search agents once retrieved, which is why it's a risk: it's manipulation, it can be detected, and it can damage trust in your brand. Never use it.

For a sharper shortlist, see which GEO tools are worth paying for.

Are free GEO tools enough to start? For the first pass, usually yes. Access checks are deterministic: a robots.txt parser, a crawler request test and a raw HTML check give clear pass or fail results. Paid tools earn their price on repeated accuracy audits across several engines and on tracking change over time. Start free, fix what blocks you, then decide whether ongoing monitoring is worth a subscription.

Should I remove the schema markup I already have? No. Google says existing structured data does no harm, and it still powers some search features. The point is narrower: schema isn't a lever for getting into AI answers, so don't pay for a tool on the promise that adding more will raise citations. Leave working markup in place and spend effort on access and accuracy.

Does allowing GPTBot help ChatGPT find my site? No. OpenAI documents GPTBot as a training crawler only. ChatGPT search uses OAI-SearchBot, and OpenAI states each setting is independent. You can block GPTBot to keep content out of training and still allow OAI-SearchBot to stay eligible for ChatGPT search answers. A GEO tool should report these two bots separately.

How often should I re-run access checks? Re-run them after any change to your CDN, firewall, hosting or robots.txt, and on a regular schedule regardless. Third parties can change your access without you touching anything. Cloudflare's September 2026 default, which blocks Training and Agent crawlers on ad-displaying pages for some customers, is one example. A monthly check catches most silent changes.

Can any GEO tool guarantee AI citations? No. Citations depend on retrieval and generation, which the engine controls. A tool can confirm your pages are reachable, readable and accurate, and it can measure what engines say about you. It can't make an engine search, pick your page or cite it. Treat a guaranteed citation, ranking or score as a reason to look elsewhere.