Costs & pricing
Is AEO worth it? A skeptic's honest answer.
No pitch: the concessions the sellers should make first, the verdict by business situation, ten signs you need this, four signs you don't yet, and the proof standard to hold every vendor to — including us.
Updated 2026-08-11 · Prices and data verified 2026-08-11
Sometimes. Answer engine optimization is worth paying for when three things are true at once: your buyers already ask AI assistants the questions your business should be the answer to, the answers leave you out, and one customer is worth enough to cover the work many times over. When any of the three is false, the honest answer is not yet. This page is the skeptic's version of the question: what the sellers should concede up front, the verdict by situation, and the proof standard to demand from anyone who wants your money — including us.
Four concessions before any argument
First, the conflict of interest.This page is published by AEO Action, a company that sells AEO software — one of the options weighed here is ours, so read everything below as a vendor's account and check it against your own free measurements. (If the term itself is new, start with what AEO isbefore deciding whether it's worth anything.)
Second, nobody can put you into ChatGPT. There is no portal, no payment, no placement. Every vendor in this market — agency, tool, or us — can only build the pages and signals engines read, submit them for indexing, and measure what the engines then decide. Anyone who talks about your future AI visibility in the language of certainty is describing a decision that is not theirs to make.
Third, no citation date can be promised. The timeline evidence that exists is observational — medians across tested pages, quoted later on this page — and plenty of pages take longer or never get cited at all. Treat any promised date as a sales artifact.
Fourth, a single check proves very little — in either direction. The screenshot a vendor shows you goes stale, and so does the reassuring answer you got when you checked your own brand last Tuesday:
40–60%
of the sources AI engines cite change every month
Source: eMarketer, 2026
This cuts both ways. It's why one flattering answer isn't safety and one bad answer isn't doom — and it's the strongest honest argument that AEO is a cadence, not a purchase.
The verdict, three ways
Buyers in your category ask AI, the answers omit you, and your customer value is high. This is the clear case. The demand already exists, the gap is measurable for free, and the math is forgiving — when one client, patient, or contract covers months of cost, being one of the names an engine gives doesn't need to happen often to pay. Measure the absence first; spend second.
There's nothing for the engines to read yet. No real website, no reviews, no third-party record, pre-product — then AEO spend idles, because engines corroborate before they recommend. Build the foundations first. The manual playbook costs hours, not fees, and its early steps double as those foundations.
You're already named accurately, or you have in-house hands. If the engines already describe you well, what you need is watch-duty and defense, not a rebuild — worth tool prices, not retainer prices. If an SEO or writer already ships for you, measurement may be the only missing piece. Match the spend to the gap, not to the category.
The numbers a skeptic should actually weigh
Strip out vendor decks and there are four load-bearing facts. The first two establish that the surface matters:
48%
of Google searches now return an AI Overview above the organic results
Source: Similarweb, March 2026
And when that AI answer appears first, Similarweb's 2026 data shows organic clicks fall by 61% — the answer absorbs the attention the links used to get. The third fact establishes that existing SEO strength does not carry over on its own:
77%
of businesses ranking on Google page 1 are invisible in ChatGPT
Source: AI Search Engineers visibility audit, 2026
A skeptic should discount this one before using it: it is a vendor study — an SEO firm's audit of its own sample, not independent research, published by a firm that sells the fix. We keep it because the direction is consistent with every audit we run and with the two Similarweb figures above, and because the alternative is citing nothing. Use it as evidence that 'we already do SEO' is an incomplete defense, not as a measurement of your business.
The fourth fact is about the traffic itself — the reason absence from these answers is worth estimating rather than shrugging off:
7% vs 5%
conversion rate of AI-referred visitors versus average visitors — they also stay ~15 minutes versus 8 and view 12 pages versus 9
Source: thestacc.com, 2026
One study, so hold it loosely — but the direction makes sense: a visitor who arrives pre-recommended by an assistant behaves like a referral, not a browser.
What these numbers do not tell you: whether your buyers ask, whether your name is absent, and what that absence costs you. Those three are the worksheet below — and they're the difference between a real decision and buying a category because its charts point up.
Ten signs it's worth it for your business
Count honestly. In our reading, three or more of these is a strong case; six or more means the question has already answered itself.
- Your buyers' questions are recommendation-shaped. “Best,” “which,” “who should I hire” — engines answer those with a short list of names, and every list is somebody's new customer.
- You checked, and the names weren't yours. The five-prompt manual check takes minutes and settles whether this page is theoretical for you.
- You rank on Google, but the engines skip you. The 77% pattern above. If this is you, the diagnosis guide finds which specific failure applies.
- One customer covers months of the cost. Law, dental, B2B services, home services with large tickets — the worth-it math barely needs a calculator.
- Prospects have started saying “I asked ChatGPT.” Once that phrase shows up in your intake calls, the channel has arrived in your category whether you invest or not.
- An AI Overview now sits above your best queries. The clicks those rankings earned are being absorbed by an answer box you're not in.
- Your third-party record is thin or stale.Old directory entries, few reviews, no roundups — the material engines corroborate against either misrepresents you or doesn't exist.
- The same competitors recur across engines. The sources feeding them compound over time, which is exactly why reverse-engineering their citations is cheaper now than later.
- You already publish content nobody quotes. The assets exist; reshaping them into answer-shaped, schema-marked pages is the smaller lift.
- Nobody would notice a wrong answer. If an engine told buyers your old prices or a dead product line tomorrow, no one on your team would know for months. That blindness is its own cost.
Four signs it isn't worth it — yet
This is the section vendor pages skip. If any of these describes you, keep your money for now:
- No real website. A placeholder or a social-media-only presence gives engines nothing to read, quote, or cite. Ship the site first.
- No third-party footprint at all. Zero reviews, no listings, no press, pre-product. Engines corroborate before they recommend; a few months of ordinary reputation-building comes first.
- Your demand doesn't research.Walk-in trade, referral-only contracts, procurement lists — if buyers genuinely never ask an assistant or a search box, AI answers can't route them to you or away from you.
- Nobody will own even the approvals.Every version of this work — DIY, tool, software, agency — needs one human who reads a report and clicks approve. If that person doesn't exist, the spend idles regardless of vendor.
What invisibility costs: a worksheet, not a number
The favorite move of this category's marketing is the invented dollar figure — “you're losing $X a month!” — with no inputs shown. We won't print one. The loss from AI invisibility is real but it is your number, built from three inputs only you can supply:
One more honest wrinkle: your analytics cannot do this worksheet for you. Many AI recommendations arrive with no clickable link — the buyer reads the answer, then shows up later as a branded search, a phone call, or a “direct” visit, or quietly goes to the competitor the answer named. The part of the channel your dashboard can see, and the larger part it can't, are mapped in how to track AI traffic in GA4.
Input B is the free part: the audit measures your absence rate across 30 live queries in about three minutes. No account, no email, no card.
How to demand proof — from any vendor
Whether you end up talking to an agency, a tool, or us, worth-it stops being a matter of opinion the moment you insist on four questions. Good vendors of every type can answer them; the answers are where the money hides.
- “Show me the answers, not a score.” A visibility number without the verbatim, dated engine answers behind it can't be audited. Ask to read exactly what ChatGPT said, on what date, to which question.
- “For every deliverable, two dates.” The date it went live, and the date an engine first cited it. If a vendor can't connect those two dates for their own work, their report measures activity, not outcome.
- “What's the re-check cadence?” With 40–60% of cited sources turning over monthly (eMarketer, 2026), a quarterly PDF describes a market that no longer exists. Weekly or daily re-checks are the honest minimum.
- “What happens when nothing moves?” The trustworthy answer contains the word observation — medians and percentages across pages, like the ~6.8-day median to a first ChatGPT citation once indexed and the ~42% of pages cited within 30 days that 2026 practitioner testing (Semrush) reported — and never a promised date. Walk away from certainty.
Where we fit, honestly labeled: AEO Action was built around question two. Every landing page approved in the product is registered against the buyer question it targets and re-checked on each following scan for up to 120 days, producing a dated line from shipped to first cited — we call it the fate receipt. Nothing publishes without your approval, and the trial runs 7 days with no card. To be clear about what that is: a mechanism you can inspect, not evidence about your business. Your receipts would be your own.
And the proof standard doesn't point at one vendor. If you have in-house hands and only lack measurement, monitoring tools at $29–399/month published list prices (checked August 2026) are the cheaper right answer. If your citations are decided by journalists and relationships, a $2,000–6,000/month agency retainer buys judgment no software has — what those retainers include is its own honest read. The full ladder, rung by rung with every number sourced, is in how much AEO costs in 2026. Hold whoever you pick — us included — to the four questions.
08 · FAQ
The questions that decide it.
Is AEO worth it for a small business?
Does AEO actually work?
How do I measure AEO ROI?
What happens if I stop paying for AEO?
Is AEO just hype?
Keep reading
Costs & pricing
How much does AEO cost
The full price ladder — DIY, tools, software, agencies — with every number sourced and dated.
Diagnosis
Not showing up in ChatGPT
The six real reasons — and a ten-minute self-diagnostic to find out which one is yours.
Costs & pricing
AEO services
Retainer anatomy — what $2,000–6,000/month buys, when an agency is right, and the software alternative.
Diagnosis
Does ChatGPT know you
The five-prompt manual check, the three outcomes, and why a single check can mislead you.
Answer it with your own numbers, not ours.
Input B of the worksheet — your absence rate — is the only one you never have to estimate, and it costs nothing: ten of your buyers' questions, 30 live queries, about three minutes, no account or card. If the absence turns out to be real, what we charge to close it is at /#pricing.
Free audit first — see exactly where you stand before paying anything. 7-day trial, no card.