How much does AEO cost in 2026?

AEO costs most mid-market B2B companies somewhere between roughly $2,000 and $10,000 a month when it is bought as a monthly engagement, which is the range Digital Elevator's 2026 pricing guide gives for credible answer engine optimisation retainers. Treat that as the shape of the market rather than a quote, because it comes from an agency describing its own category.

The band is wide for a structural reason: the same three letters are sold as at least five different products. A monthly service, a fixed-scope project, a one-off audit, a software subscription and an internal hire all get called AEO, and their costs are not on the same scale. A monitoring subscription and a full monthly engagement are not the expensive and cheap versions of one purchase.

What makes the market hard to read is that almost nobody publishes a figure. The same Digital Elevator guide notes that established agencies rarely put out a rate card, which means the first number most buyers hear has been set after the supplier learned the size of the company asking. Three quotes written against one scope document is the only reliable way to find out where the middle of the market actually is.

Spend is already normal at companies like yours, which matters in a room where being early is treated as the risk. Conductor surveyed more than 250 enterprise executives and digital decision makers for its 2026 State of AEO and GEO report and found they had put an average of 12% of digital budgets into answer engine optimisation in 2025, with 94% planning to increase that share in 2026. Read the absolute numbers with the sample in mind, since enterprise digital budgets are not a guide to what a 60 person company should spend.

For what SIGNALS charges, see the pricing page. Figures stay there and are maintained by hand, so an article never quotes a rate that has since moved.

What drives the price of an AEO engagement up or down?

The price of an AEO engagement is driven by five things, and none of them is the number of articles. Scope in this work is measured in questions tracked and pages read, which is why two proposals with the same page count can differ by a factor of four.

The size of the question set comes first. A fixed list of buyer questions run across the engines is the spine of the engagement, and the list grows with every product line, buyer type, region and language you sell into. One product in one language is a different engagement from five lines across three regions, and the cost moves roughly with the number of distinct questions somebody has to run, log and read.

Engines and competitors tracked come second, and they multiply rather than add. Tracking four engines instead of one quadruples the runs behind every question, and naming six competitors in the log rather than two is more reading for every answer. Query fan-out makes this larger than it looks: AirOps' 2026 retrieval study found 89.6% of the 15,000 original prompts it tested generated two or more internal sub-queries, so one tracked question is already several searches behind the scenes.

The third is how many of your pages get read in detail, which is the difference between a crawl report and a diagnosis. A machine pass over 400 pages is cheap. Reading twelve pages against the questions they are meant to win is not, and it is the part that produces specific instructions rather than a list of warnings.

The last two are about labour split. A quote where the supplier writes and edits the pages is higher than one where they hand your team a brief, and a quote where somebody implements the rendering and heading fixes is higher than one where they file a ticket for your engineers. Neither is wrong. They are different purchases, and a cheaper quote often means your own team is carrying the part that was priced out.

What are the AEO pricing models, and which one fits your situation?

AEO pricing models fall into five shapes, and picking the wrong shape costs more than paying too much inside the right one. The table sets out what each one buys and the situation it suits.

Model	What you are buying	Who it suits	What it leaves to you
Monthly engagement	Repeated measurement, page fixes and new coverage, reported against a fixed question set	Companies where AI answers already name competitors and nobody internally owns the problem	Subject-matter time, publishing approval, and a named owner for the question set
Fixed-scope project	A defined piece of work with an end date, such as a baseline plus a technical pass plus ten rewrites	Teams with capable content people who need the diagnosis, not the hands	All the recurring measurement, which is where decay shows up
One-off audit	A written diagnosis of where you stand and what to change, with no implementation	Anyone deciding whether this is worth a budget line at all	Every fix, and the re-measurement that proves a fix worked
Software subscription	A dashboard that runs prompts and tracks mentions across engines	Teams with someone to read it weekly and the capacity to act on it	The reading, the deciding and all of the page work
Internal hire	Permanent capacity, in salary rather than fees	Companies where AI search will be a standing channel, not a project	Tooling, ramp-up time, and cover when that person leaves

The honest comparison between models is not price per month. It is price per decision you can actually make. A dashboard that nobody reads costs its subscription plus the staff hours it absorbs and returns nothing, while a one-off audit that gets implemented can be the cheapest useful thing on this list. Our guide to doing AEO in-house versus buying an agency or a tool works through that trade-off with the staffing numbers attached.

Mixing models is normal and usually sensible. A common pattern is a one-off audit first, a fixed-scope project to clear what the audit found, then a monthly engagement once there is something worth measuring repeatedly. Whether the recurring part is genuinely necessary is a fair question, and whether AEO is a project or a retainer sets out which parts of the work recur and which are done once.

Is an AI visibility tool cheaper than hiring an agency?

An AI visibility tool is far cheaper in cash and frequently more expensive in total, and the variable that decides which is true for you is whether anyone on your team will run it. Tool subscriptions are a fraction of a retainer, and our comparison of the best AEO tools in 2026 carries current published prices for the main products.

What the subscription does not include is the hour a week somebody spends running prompts, the judgement about which movement is real, and all of the page work that follows. A dashboard reports that you were named in three of twenty answers. Deciding which of your pages to rewrite first, and then rewriting it, is the expensive part, and no tool does it.

The measurement discipline is the specific thing that tends to fail in-house. Running the same prompts the same way every month and keeping a log you can compare is dull, and it is the first task dropped when a quarter gets busy. Any price comparison that assumes the internal runs keep happening is comparing a real retainer against an optimistic one.

There is a real case for tools over a service, and it is worth saying plainly. A company with a disciplined content team, a person who owns organic search, and the ability to publish inside a week will get more from a subscription plus their own labour than from a thin retainer, because they already have the part that is hard to buy. The split that usually works is buying the measurement and doing the fixing yourself, which keeps the comparable series intact and keeps the page work with the people who know the product.

The case for a service is the opposite situation: nobody owns it, the content team is booked, and the pages that need rewriting are the ones nobody wants to touch. In that company a tool becomes a tab nobody opens, and the subscription is a smaller number attached to a worse outcome.

What does a cheap AEO retainer usually leave out?

A cheap AEO retainer usually leaves out the measurement, which is the part that makes the rest checkable. The invoice says AEO and the work is article production, so you get volume with no series to tell you whether any of it was retrieved, let alone cited.

Four exclusions show up again and again at the bottom of the market. The monthly re-run of the same question set goes first, replaced by a one-time baseline nobody revisits. Access to the raw run log goes second, which means the summary cannot be checked and the baseline cannot be carried to another supplier. Rewriting pages that already exist goes third, because new articles are easier to show as output. Implementation goes fourth, so the rendering and heading fixes arrive as recommendations your engineers never prioritise.

The reason the measurement matters more than it sounds is that citations rotate on their own. AirOps' 2026 retrieval study tracked 548,534 pages ChatGPT retrieved and found only about 15% became a visible citation, with the set rotating between runs. Without repeated runs you cannot separate a change you caused from noise you watched.

None of this makes a lower price dishonest. A smaller engagement that says plainly it covers two engines, one question set of fifteen questions, monthly runs and four page fixes is a clear, buyable product. The problem is the quote that uses the same words as a full engagement for a third of the fee and resolves the difference silently. Our breakdown of what an AEO engagement includes month to month is written to be used as a checklist against a proposal.

How do you compare two AEO quotes that look the same?

Comparing two AEO quotes starts with forcing both suppliers to answer the same seven questions in writing, because proposals are rarely built on the same units. Until that is done you are comparing two documents, not two offers.

Ask for the number of questions in the tracked set, the engines they run against, the run frequency, who writes the pages, who implements technical fixes, who owns the raw run log, and what the review point is at which you can stop. Seven answers fit on one page, and a quote that resists putting them in writing has told you something useful about month four.

Normalise on cost per tracked question per month and the difference usually appears immediately. An engagement covering forty questions across four engines with page rewriting included is not more expensive than a cheaper one covering ten questions on one engine with briefs instead of pages. It is a different quantity of the same thing, and the lower invoice is sometimes the worse rate.

Guarantees deserve the opposite treatment to the one buyers expect. A promised citation count should lower your confidence rather than raise it, because the same prompt returns different sources on consecutive runs and nobody controls the engines' retrieval behaviour. Our guide to choosing an AEO agency lists the other claims worth pushing back on before signing.

The cheapest way to make every quote legible is to have your own baseline first. A free visibility assessment runs your buyer questions across ChatGPT, Claude, Perplexity and Google AI Overviews and reports who is named today, which turns a proposal from a promise into a claim about numbers you already hold.

What should you expect to spend in the first quarter?

The first quarter of AEO spending is front-loaded into work you only buy once, and a proposal that charges a flat monthly fee across it is averaging rather than scoping. Month one is mostly measurement and technical checks, and months two and three are where the recurring shape appears.

Three things get built before anything is published: the question set, at least two baseline runs a week apart, and a list of page-level faults ordered by what they cost you. Running the baseline twice is not duplication. It measures your own noise, which is substantial, and without it the first reported improvement cannot be told apart from sampling.

Budget for your own team's time as part of the number. An hour a month with somebody technical, somebody who can approve a page going live inside a week, and access to analytics and staging are what let a supplier verify that a change landed rather than assert it. Engagements that stall have usually lost one of those rather than hit a technical wall.

Where the money goes should shift visibly after the first quarter. Early on, most of the gain is in pages that already exist and nearly work, so the mix is weighted towards rewriting. Once those are done the remaining gap is coverage, and the work starts to look like a publishing operation with a measurement loop attached. A supplier whose mix never changes is probably not reading their own numbers.

If the spend has to be defended internally before it can start, the argument is easier with comparable-company data than with a forecast. Conductor's 2026 survey of more than 250 enterprise digital leaders found 32% named answer engine optimisation their single top priority for the year, more than any other option, which supports the narrow claim that this is a line item at firms you are compared against. Our guide on getting budget approved for AI search visibility sets out the rest of that case.

What else do people ask about AEO prices?