Skip to content
Learn · AEO fundamentals

Best AEO agency: why the honest answer is a shortlist, not a name

There is no best AEO agency, because the six providers we audited in August 2026 sell four different business models at published prices ranging from $99 a month to figures two of them decline to publish at all, and the useful output of a search like this is a shortlist of two or three providers that will take your kind of business, checked against their own published method.

Vignette: a name being marked inside a generated answer. Illustration, not measurement: no figure appears in the loop.
In short

Key takeaways

  • Two of the six providers we audited publish best-agency rankings on their own sites, and both place themselves first. One of them also publishes competitors' pricing and cons, including its own, which is the more defensible version of the format.
  • We are not publishing a ranking with ourselves at the top. An argument with the conclusion pre-filled is discounted by readers and by engines, and we have less published proof than several of the providers we would be placing below us.
  • Provider fit is mostly decided before quality: one publicly refuses local service businesses, two are aimed at companies with an in-house marketing team, one is software with paid execution credits.
  • Run four checks: will they take your kind of business, is the deliverable a number or shipped work, can you verify one of their published figures, and what do they refuse to do.
  • Caldrin's own disqualifiers are listed here in full, including that we have no revenue-attributed win and hold every client citation figure we have measured.

Why best-agency pages exist, and who writes them

A search for the best AEO agency mostly returns pages written by AEO agencies. That is not a scandal, it is the economics of the format: the page ranks, the page is quoted by assistants, and the company that wrote it appears at the top of its own list.

We inventoried two of the six providers in our August 2026 audit deeply enough to see the mechanic. Breaking B2B runs roughly fourteen of its forty-seven blog posts as agency rankings that include itself, and in the ones we read it ranks itself first while publishing each competitor's services, pricing, pros and cons, including its own cons. aeoengine.ai runs the same play across a large comparison surface, sixty-four pages of them at the time of our crawl, and keeps competitor pricing behind weighted criteria rather than printing it. Of the two, Breaking B2B's version is the more defensible: a reader can at least see what is being compared and what the author admits about itself.

We are not going to publish our own version, and it is worth being precise about why, because the reason is not modesty. First, a ranking on our own domain that concludes with us at number one is an argument with the conclusion filled in before the evidence, and readers and engines both discount those, which makes it a poor investment as well as a poor argument. Second, and more directly: on published proof we would be placing ourselves above providers with more of it. Breaking B2B publishes nineteen named case studies. GrowthX publishes five customer stories and sixty-five pages of product documentation. We have no revenue-attributed win at all. A list that ranked us first would be an assertion we cannot support, and this site's entire position is that we do not publish figures we cannot support.

What follows instead is a routing table. It is built only from what each provider publishes about itself, captured on 2026-08-15, plus observations we made from their own pages that any reader can repeat. No competitor's client results appear anywhere below, because those are their claims about their clients and we have not verified a single one.

An answer enginenot captured
Commonly cited options include Provider A, Provider B and Provider C. Provider A publishes lower monthly tiers, Provider B focuses on business software clients, and Provider C sells a platform rather than a service.
prompt
“what is the best AEO agency”
captured
not captured

illustrative, not a capture A shortlist assembled this way is largely assembled from pages the providers wrote about themselves, several of which rank their own author first. The fit constraints that actually decide the outcome, such as a provider publicly refusing an entire segment, sit below the summary. Written to show the shape of an engine answer. It has no session, no capture date and no sample size, so it is not evidence about any category, including yours.

The field, as they describe themselves

aeoengine.ai publishes monthly tiers at $1,597 and $2,997 with a custom enterprise option on 90-day rolling terms, and describes its delivery as a human strategist working alongside a large set of execution agents. It is horizontal, covering many industries with a heavy weighting toward legal, local services and ecommerce, and it runs the largest content surface in the audited set: 777 URLs discovered in our crawl, including roughly 45 free tools and 381 blog posts. It publishes a performance guarantee. Fit: an owner-operator who wants the lowest published managed price and a guarantee. Things a buyer should check for themselves: their published Answer Index defines appearance rate as a share of 300 tracked questions, which makes the possible values multiples of roughly 0.333 percent, yet the tail of the published National 100 carries values such as 0.9, 0.8, 0.6, 0.4 and 0.1 percent, most of which are not reachable under that definition; the same month's editions show contradictory context columns for the same firms; the sitemap covers roughly 70 of those 777 URLs and the homepage last-modified date was still 2024-11-04 when we crawled. Set against that, their methodology page freezes and versions the question set and keeps a public correction log, which is better practice than most of the category manages.

Breaking B2B publishes tiers at $4,000, $5,500, $10,000 and $20,000 and above per month, states an average client at roughly $7,000, and sells in 90-day sprints with no lock-in. It is a senior-led boutique run by one operator, Sam Dunning, and its proof surface is 19 named case studies against 222 total URLs, with 137 podcast episodes as the demand engine. Fit: B2B software companies who want a named senior operator rather than an account team. Disqualifier, in their own words: the homepage states they do not take dentists, restaurants or local plumbers. If you are a local service business, this is settled before any quality comparison begins, and a competitor publicly declining a segment is more useful information than most sales calls produce.

GrowthX publishes a from-$6,000 monthly price for a platform your team operates with a strategist in the loop, and positions explicitly against agencies on the argument that agency dependency is the thing to escape. Its site is shaped like a software company: of 144 URLs, 65 are product documentation and 58 are long-form essays, against three blog posts. It publishes seven legal pages including a master agreement, a data processing addendum and a subprocessor list. Fit: a company with an in-house marketing team that will actually operate a product, and a procurement function that wants that paperwork. Consequence to weigh: if nobody in-house runs it, you have bought software.

daydream and Graphite publish no price. daydream describes an AI-native agency model pairing proprietary agents with dedicated experts, aimed at B2B software and growth-stage technology companies. Graphite describes a research team publishing original work on AI answers, with enterprise clients and a fixed prompt benchmark re-run quarterly. Fit: companies large enough that a call before a number is normal. Consequence: you cannot compare cost before entering a sales process, which is a real cost of its own.

CrowdReply is not an agency. It publishes subscription tiers at $99, $299 and from $499 a month and sells execution as per-unit credits on top, including individual social comments and posts and a backlink marketplace. Fit: cost-sensitive self-serve buyers, and agencies white-labelling execution. One feature deserves a buyer's attention before anything appears under their brand: it posts through established persona accounts, which is undisclosed paid placement, and that is a platform-terms and disclosure question rather than a neutral feature.

One more entry belongs in any honest version of this list, and it is not an AEO provider at all. The large marketing agencies many businesses already pay do not sell this. We fetched two of them live on 2026-08-15. One lists roughly eighteen services, including a reporting dashboard it describes as carrying more than 300 KPIs, which is their figure, and mentions AEO, AI search, LLM visibility, answer engines, ChatGPT and citations nowhere; its only AI is a chat widget. The other splits its offering into technology, advertising and marketing with no AI-search line. If you already pay one of them, the realistic comparison is not agency against agency. It is whether to add a measurement layer over the work you already buy.

Four checks that build a shortlist

First, will they take your kind of business. This eliminates more of the field than quality ever does, and it is usually published. One provider names the segments it refuses. Two are aimed at companies with an in-house team. One is software. Read the disqualifiers before the case studies.

Second, is the deliverable a number or shipped work. A platform hands you measurement and expects your team to act. A managed service is accountable for the acting. Both are legitimate and they are priced differently, so buying the wrong one is a category error rather than a bad deal. The test is simple: ask who writes the page that fixes the gap the report identifies.

Third, can you verify one of their published figures. Take the definition a provider gives and check whether the number they printed is reachable from it. The Answer Index example above took five minutes with a calculator and it is the highest-yield check available to a buyer, because it tests the thing you cannot otherwise see. Run it on us too. Every number on this site carries its instrument, its sample size and its date precisely so that it can be run.

Fourth, what do they refuse to do. Ask for a figure they withheld in the last quarter and why. Ask which engines they cannot reach. Ask what they will not do to manufacture the appearance of being recommended. A provider with no refusals has not thought about the boundary, and in a category where the output is probabilistic and the audit trail is thin, the boundary is most of the product.

Our own entry, written to the same standard

Caldrin gives every engagement one owner: the person who scopes the work runs it and presents the findings. The front door is a free AI visibility report anchored at a struck-through $1,500 and delivered on a call rather than as an automated PDF, and the work after it is scoped per engagement rather than sold from a tier list. Five of the six providers we audited in August 2026 gate or charge for the initial assessment; CrowdReply is the exception, and only because it is self-serve software with no assessment to gate.

What we can show you. A 488-point local grid for 603 Basement Solutions, who have cleared us to use their name: 8 service keywords across 61 towns in New Hampshire, Massachusetts and Maine, showing them holding a local pack position at 10 of 488 points, or 2.0 percent, on 2026-08-09, while one competitor held 39.3 percent of the same grid. A page-by-page indexation audit on another account that found 26 of 52 pages indexed and only 3 of 22 articles. And a controlled experiment we lost: we changed page titles on a client site, the homepage average position moved from 16.8 to 34.5 overnight, and 34 untouched control pages moved 0.8 positions over the same window, which is how we knew it was us and not the market. We rolled it back the same day.

What we cannot show you, in full. We have no revenue-attributed win. We hold every client citation share figure we have measured, because each is a single-sample reading with roughly 25 points of margin and our rule is that nothing travels until it has run at least three times, which means a dashboard could show you a citation percentage tomorrow and we will not. We have no B2B software case studies, so if you are a growth-stage software company, three providers in the list above have more relevant proof than we do. We publish no guarantee while two of the six do. We do not have the procurement documentation an enterprise buyer expects. We do not measure Claude at all, because no ground-truth path to the consumer product exists, and we do not cover Perplexity either, because it has no scraper and automating its consumer site is disallowed in its robots.txt, so any coverage comparison against a provider claiming every engine will look worse for us on the page and better in the footnotes. and we do not cover Perplexity either, because it has no scraper and automating its consumer site is disallowed in its robots.txt,

Who we are wrong for: procurement committees, buyers who need a case study from their exact vertical with revenue attached, teams that want a dashboard to run themselves, and categories where AI answers do not name businesses yet. That last one is a measurable condition rather than a judgement, and the free assessment is designed to be able to return it. If the channel is not live in your market we will say so and you will have spent nothing, which is the one promise a seller who charges nothing for the diagnosis can actually make.

Questions

Questions, answered plainly.

Which is the best AEO agency?

The question does not have an answer, because the six providers we audited in August 2026 sell four different business models: retained agency, platform plus strategist, self-serve software with paid execution credits, and research-led enterprise consulting. Published prices run from $99 a month to figures two of them do not publish. The useful output is a shortlist of two or three that will take your kind of business, checked against their own published method.

Why do so many best-AEO-agency lists rank the author first?

Because the page is an acquisition asset. Two of the six providers we audited publish agency rankings on their own sites and both place themselves first; in one case roughly fourteen of forty-seven blog posts are self-including rankings. The more defensible version publishes each competitor's pricing, pros and cons, including the author's own cons. The less defensible version keeps competitor pricing behind weighted criteria.

Why does Caldrin not publish its own best-agency ranking?

Two reasons. A ranking on our own domain that concludes with us first is an argument with the conclusion pre-filled, and readers and engines discount those. And on published proof we would be ranking ourselves above providers who have more of it: one publishes nineteen named case studies, another publishes five customer stories and sixty-five pages of product documentation, and we have no revenue-attributed win at all.

How do I check whether an AEO agency's numbers are real?

Take their published definition and test whether their published number is reachable from it. One provider defines appearance rate as a share of 300 tracked questions, which makes the possible values multiples of roughly 0.333 percent, and then publishes tail values including 0.9, 0.8, 0.6, 0.4 and 0.1 percent. The definition and the numbers are theirs; the arithmetic takes five minutes. Run the same check on us.

Who should not hire Caldrin?

Procurement committees, buyers who need a case study from their exact vertical with revenue attached, teams that want a dashboard to run themselves, and businesses in categories where AI answers do not yet name companies. We also have no B2B software case studies, publish no guarantee, hold every client citation figure we have measured until it re-runs at three passes or more, and do not measure Claude at all.

Method

What is measured, and what this page is not.

This is an explainer. It carries no figures, and it is not a reading of your category. The disclosure below states the instrument that produces the numbers the essay refers to, so the distinction is on the page rather than assumed.

instrument
Caul
what was measured
Nothing on this page. Where the essay refers to citation share, that figure is produced separately, per account.
how
A prompt set written once for a category and then frozen, run against every engine in clean sessions, with each answer stored unmodified.
over what window
Reviewed on 2026-08-16. The engines change, so read the essay against the date on the byline.
what this cannot tell you
An explainer is not evidence about your category. Being named is not being recommended, and it is not traffic or revenue. Any figure about your own visibility has to come from a capture of your own category, carrying its sample size and its window.
Keep reading

The rest of the cluster.

All Learn guides · Glossary

See where you stand.

The audit is a real sweep of your category, benchmarked against competitors you name, delivered on a call so the findings get explained rather than emailed. You keep the report and the underlying data whatever you decide afterwards.

Get your free auditWhat is in the report