Which AEO tools actually move brand citations: what the data shows
No AEO tool can be shown to move brand citations, because no vendor publishes a method behind its results and no independent study has measured one. So we measured the next best thing and say so plainly: what each of ten vendors publishes about itself. Eight publish a price. Nine publish customer results. None publishes how those results were measured.
We publish this page and are one of the ten vendors audited on it, including the row where GetXEO alone scores No. Every reading was taken on 29 August 2026.
In short
The questions this page answers, and the short answers.
- Is there data showing which AEO tools move brand citations?
- No. Nine of the ten vendors publish customer results, and none publishes the method behind them, so the figures cannot be compared or verified. No independent study has tested these products against each other either. Any ranking claiming outcome data is describing marketing material.
- What did this audit of AEO vendors measure?
- What each vendor publishes about itself, read on one day across ten vendor sites. Price, plan limits, engine coverage, metering unit, content output, compliance record, evidence for results, and each vendor's own robots.txt policy toward named AI crawlers.
- How many AEO vendors publish a price?
- Eight of ten. AirOps and Conductor publish none at all. One more, Scrunch AI, serves different plan sets to a browser and to a plain fetch, so a price read without a browser can be wrong for that vendor.
- Do AEO vendors allow AI crawlers on their own sites?
- Almost universally, and almost none of them say so deliberately. Nine of ten robots.txt files name no AI crawler at all and fall back to a catch all rule. GetXEO is the exception, naming each agent group explicitly and blocking one scraper.
- What is the most useful published number when comparing AEO tools?
- The metering unit and its ceiling. No two of the ten vendors meter on the same set of units, and the ceiling behavior differs by vendor. Those two facts predict a real monthly bill far better than any headline price or published customer result does.
What can be measured here, and what cannot
The question as asked has no evidence behind it.
The honest answer to the question in the title is that nobody knows, and the reason is worth stating before any table appears.
Nine of the ten vendors in this audit publish customer results. Those results are impressive and mutually incomparable: different time windows, different baselines, different definitions of a mention, and no published method anywhere. A lift figure from one vendor cannot be set beside a lift figure from another, and neither can be checked.
No independent study has tested these tools against one another either. What exists is research on what makes any page citable, not on which vendor produces such pages. Pages carrying figures, quotes and citations get named noticeably more often, while off site brand signals rank ahead of on page craft, which is useful and is not a product comparison.
So this page measures what is measurable: what each vendor is willing to publish about its own product. That is a proxy for evaluability rather than for performance, and it is the strongest checkable signal available today.
Method: one day, ten sites, published pages only
How the audit was run, so it can be repeated.
The method is deliberately dull, and it is written out so anybody can run it again and get the same answer or a different one.
- Scope. Ten vendors: Profound, GetXEO, AirOps, Peec AI, Writesonic, Conductor, the Semrush AI Toolkit, Ahrefs Brand Radar, AthenaHQ and Scrunch AI
- Sources. The pages each vendor publishes itself, and nothing else. Pricing, plan comparison, the relevant feature pages, and the robots.txt file at the root of the domain
- Date. All readings taken on 29 August 2026, so a change after that date is not reflected here
- Rendering. Pages were read in a browser where a vendor renders figures only in the browser, which applies to at least three of the ten
- Exclusions. No review site, no roundup, no analyst summary, and no figure quoted from any vendor about a different vendor
One methodological note. A vendor withholding a figure is recorded as not published rather than as absent, because those are different claims and conflating them is how a competitor page becomes wrong.
What ten vendors publish, and what they hold back
The audit result, one row per vendor, one day.
The pattern in this table is more interesting than any individual row: the vendors publishing most about limits are not the ones publishing most about results.
| Vendor | Price | Results | Method | Metering unit |
|---|---|---|---|---|
| Profound | Yes | Yes | No | Prompts, responses, credits |
| GetXEO | Yes | No | No | Questions, blogs, workspaces |
| AirOps | No | Yes | No | Tasks, tracked prompts |
| Peec AI | Yes | Yes | No | Prompts, models, projects |
| Writesonic | Yes | Yes | No | Prompts, answers, articles |
| Conductor | No | Yes | No | Drafts a year, pages, credits |
| Semrush | Yes | Yes | No | Prompts daily, reports daily |
| Ahrefs | Yes | Yes | No | Checks, prompts, credits |
| AthenaHQ | Yes | Yes | No | Credits, per AI response |
| Scrunch AI | Yes | Yes | No | Prompts, audits, optimizations |
Three findings come out of that grid. Eight of ten publish a price, and the two that do not are both enterprise platforms. Nine of ten publish customer results, and GetXEO is the exception because it publishes none at all, which is a gap rather than a virtue. And zero of ten publish a method behind a result.
The metering column carries the most practical finding. No two of the ten meter on the same set of units, and seventeen distinct units are named across the table, which means a price comparison across this category is an arithmetic exercise rather than a lookup. AthenaHQ publishes its formula openly, at one credit per AI response, five for heavier models, and four times for fan out, and the transparency there is unusual enough to be worth crediting.
The crawler finding, which nobody advertises
Ten robots files read the same day, one exception.
Every vendor in this category sells a product that checks whether AI crawlers can reach your site. So we read what each one does on its own domain.
Nine of the ten robots.txt files name no AI crawler at all. Each falls back to a catch all group, so GPTBot, OAI-SearchBot, ClaudeBot, PerplexityBot, Google-Extended, CCBot and the rest inherit whatever that group allows rather than getting a decision of their own. Two of those groups carry substantial disallow lists, including one with a crawl delay attached, and one file carries no rules at all beyond a sitemap line.
GetXEO is the exception, and it is our own file. Its file names the search crawlers, the training and caching crawlers, and the live browsing agents in separate groups. It allows the search and live browsing groups everywhere, keeps the training and caching group off five policy pages, and blocks one scraper by name. That is a deliberate posture rather than a default one, and it is the only file in the set that reads as a decision.
None of this says a vendor is careless. A permissive catch all is a reasonable choice for a marketing site, and every operator crawls under its own user agent whether a file names it or not. What it does say is that agent by agent control is rarer in practice than the category talks about it, even though a robots change is picked up in about a day.
What a buyer should take from this
Three conclusions supported by what was actually measured.
Three things follow from the audit, and nothing about outcomes follows from it at all.
- Price the meter, not the plan. Seventeen distinct units across ten vendors mean the headline figure predicts very little. Work out what your configuration consumes and ask the vendor to confirm it in writing
- Treat every published result as a case study. Nine vendors publish them and none publishes a method, so they establish that something happened somewhere, and nothing more
- Weight what a vendor volunteers. A platform publishing its price, its limits and its formula is a platform you can evaluate without a sales cycle, and that is a real difference in cost of decision, with most B2B buyers now using AI somewhere in the buying process
The last one cuts against us in places and we have left it standing. Buyers already fact check what AI tells them about vendors, and the same instinct applied to a vendor page is the most useful research habit in this category.
Frequently asked questions
Longer tail questions that did not need a section of their own.
Has anyone independently tested AEO platforms against each other?
No published study compares these products head to head on outcomes. Research exists on what makes pages citable in AI answers, which is a different question. Anyone presenting a vendor ranking as outcome data is presenting judgment or marketing material as evidence.
Why do AEO vendors publish results without publishing a method?
Because a case study is a marketing asset and a method is an invitation to be checked. Different time windows, baselines and definitions of a mention make every figure look strong. None of this is unique to AEO, and none of it makes the numbers comparable.
What does not published mean in an AEO vendor comparison?
That the vendor states no figure on its own pages, which is different from the capability being absent. Treating those two as the same is how a comparison page becomes wrong, so an honest audit records the silence rather than filling it.
How often should an AEO vendor audit be repeated?
Quarterly at most, and before any renewal. Plan structures in this category changed materially within a single quarter during 2026, including plan names, prices and which features sit at which tier. A figure six months old should be treated as unverified.
Does a permissive robots.txt help a vendor get cited?
It removes one obstacle rather than creating an advantage. Allowing AI crawlers lets a page be fetched and considered. Whether it then gets quoted depends on what the page says and how it is structured, which is the part no robots rule can influence.
Sources
The 10 records behind every external claim on this page.
All were published or last updated within the past twelve months. A competitor page appears only as a record of that vendor’s own published terms.
- Machine Relations, AI search citation factors research
- MarketScale, on the TrustRadius finding that buyers fact check AI
- Cloudflare, the AI crawler bot reference naming each operator, category and user agent
- OpenAI, the crawler and user agent documentation
- Machine Relations, how B2B buyers research vendors with AI
- AthenaHQ, the published credit calculator, read 29 August 2026
- GetXEO, the published robots.txt, read 29 August 2026
- AirOps, the published robots.txt, read 29 August 2026
- Ahrefs, the published robots.txt, read 29 August 2026
- Semrush, the published robots.txt, read 29 August 2026
Read next
The rest of this cluster, in the order it makes sense to read.
- What are the best answer engine optimization (AEO) tools in 2026
- AEO tools to avoid and the warning signs to look for
- Profound vs Peec AI vs AirOps: how the leading AEO tools compare
- Reporting only or reporting plus execution: the real divide in AEO tools
- AEO tools compared: engine coverage, citation tracking, and content execution