Most AEO agency pitches in 2026 fail at least one basic honesty test, and the eight red flags below will disqualify the majority of vendors you talk to. The category is young, the demand is loud, and rebranded SEO shops have flooded in: serious AEO engagements generally run $2,000 to $25,000 per month depending on scope, which is exactly the price range where an invoice can hide six months of nothing. The core problem is that no agency controls AI outputs. ChatGPT, Perplexity, Google AI Mode, Gemini, and Copilot decide their own citations, so every legitimate promise is about influence, evidence, and measurement, and every guarantee of a specific result is a tell.
The good news: the fakes are easy to catch if you know what to listen for. Each red flag below has a specific phrase or behavior you can test for in a single sales call, before any contract is signed.
Before you evaluate any agency, get your own baseline. Run the free AI visibility audit and walk into every sales call already knowing which queries cite you and which cite competitors.
The 8 red flags, in the order you will hear them
1. They guarantee specific AI rankings or citations
“We will get you cited in ChatGPT within 30 days” is the category’s loudest lie. No agency controls what OpenAI, Anthropic, Google, or Perplexity retrieves and cites; the honest version is “we can influence the inputs and show you the movement.” Any guarantee of a specific citation, position, or timeline is a sign the seller either does not understand the systems or assumes you do not.
2. They offer to “submit your site to AI engines”
There is no submission process for ChatGPT, Claude, Gemini, or Perplexity. Nothing to register, no directory, no priority queue. Vendors selling “AI search registration” are charging for a thing that does not exist. The real access points are unglamorous: Bing Webmaster Tools for the index behind ChatGPT search and Copilot, IndexNow for fast crawling, and Google Search Console for AI Overviews. All three are free.
3. They cannot explain where AI answers come from
Ask one question: “Where does ChatGPT get its answer when someone asks for the best firm in my category?” A real practitioner will walk through retrieval versus training data, the Bing dependency, which sources each engine favors, Reddit for Perplexity, Wikipedia for ChatGPT, YouTube for AI Overviews, and how corroboration across directories, reviews, and press shapes recommendations. A fake will say “our proprietary AI methodology.” Buzzwords in place of mechanics is the whole diagnosis.
4. Their strategy is schema and FAQs, full stop
Schema markup and FAQ pages are table stakes, a few weeks of work, not a retainer’s worth of strategy. If the scope never mentions the off-site half, third-party coverage, review platforms, directory depth, Reddit and community presence, YouTube, digital PR, the agency is optimizing half the problem. Engines cite corroborated entities, and corroboration lives on domains you do not own.
5. They report “exact ChatGPT rankings”
AI answers are probabilistic: the same prompt returns different citations across sessions, locations, and model versions. Honest measurement reports citation share across a fixed prompt set over time, with dates and screenshots, plus sourced data like the Bing Webmaster Tools AI Performance report, Google Search Console, and referral traffic from AI surfaces in GA4. A vendor reporting a single fixed “ranking number” for ChatGPT is reporting fiction. Real tools in this space, Profound from around $399 per month, Otterly.ai from $29 to $989, Semrush’s AI toolkit, Ahrefs Brand Radar, all report share-of-voice ranges across repeated sampling, because that is the only honest shape the data has.
6. No baseline audit before the proposal
An agency that quotes a scope without measuring your current AI visibility is guessing, or worse, running the same deliverables list on every client. The legitimate sequence is baseline first: which prompts cite you today, which cite competitors, what your crawl access and index coverage look like, then a scope targeting the actual gaps. Skipping the baseline also conveniently removes the before-and-after evidence you would use to judge them later. The full evaluation checklist is in how to vet an AEO agency.
7. Contract terms that outrun the evidence
Twelve-month lock-ins with no exit clause, on a service where meaningful movement shows in eight to sixteen weeks, shift all risk onto you. Reasonable 2026 terms look like three to six month initial commitments, monthly reporting against the baseline, and defined deliverables per month. Be equally wary of the opposite tell: prices dramatically below the market floor. Real AEO work, content, technical fixes, PR outreach, measurement, has a labor cost, and a $500 per month “AI visibility package” is an automated report at best. Where your budget actually lands on the spectrum is covered in how much does an AI visibility company cost.
8. They cannot show their own visibility
The simplest test in the category: ask the agency which AI queries cite them. An AEO shop that cannot demonstrate its own citations, its own share of voice data, or even explain how it would check, is selling a service it has never successfully performed on the easiest client it will ever have: itself. Ask for a live demo on their own brand during the call. Watch what tools they open and how fluently they read the results.
What does the honest version of each promise sound like?
The contrast is audible once you know it. Fake: “guaranteed ChatGPT placement.” Honest: “we target these 40 prompts, here is your citation share today, here is the movement we would expect by week 12.” Fake: “proprietary AI submission network.” Honest: “we fix crawl access, build the content engines quote, and earn the third-party coverage they corroborate against.” Fake: “exact rankings dashboard.” Honest: “sampled share of voice, with dates, alongside Bing and Search Console data you can verify yourself.”
The pattern behind every honest answer is falsifiability. Real practitioners hand you evidence you could check without them. Sellers of nothing hand you dashboards only they can generate and metrics only they can define. The same falsifiability test applies to the software side of the market, which we covered in the best GEO and AI visibility tools: tools give you data, agencies should give you data plus the work, and neither should give you mystery.
What green flags should you look for instead?
Invert the list and the buying criteria write themselves. A credible agency opens with a baseline audit and hands you the raw data, the prompt set, the citation record, the crawl findings, in a format you keep even if you never sign. Its proposal names the engines it will target and the mechanism for each: Bing indexing for ChatGPT search and Copilot, structured content and passage optimization for Google AI Overviews, third-party corroboration for Perplexity and Gemini recommendations. Its reporting sample shows dated, repeated sampling with the variance visible, not a single triumphant screenshot. And its team can name the sources engines actually favor in your vertical without checking notes, review platforms for local services, G2 and Capterra for software, directories and press for professional services.
Two subtler green flags separate the top tier. First, they tell you no: a good vendor will say a query set is unwinnable at your budget, that your site needs technical work before content spend makes sense, or that a cheaper scope fits your stage. Vendors who agree with everything are selling everything. Second, they connect the work to revenue instrumentation from day one, AI referral tracking in GA4, intake attribution questions, lead-source tagging in the CRM, because they expect to be judged on pipeline, not screenshots. An agency that volunteers to be measured is making the opposite bet from one that hides behind proprietary dashboards, and that bet tells you most of what the reference calls would.
Pricing posture is the final tell. Honest shops quote ranges tied to scope and explain what moves the number. Evasive shops quote a single figure fast and resist itemizing, because the itemized version would reveal how little sits underneath.
FAQ: AEO agency red flags
What is the biggest red flag when hiring an AEO agency?
Guaranteed AI rankings or citations. No agency controls what ChatGPT, Perplexity, Gemini, or Copilot cites; the engines decide retrieval and recommendation themselves. Legitimate agencies promise influence over inputs, content structure, crawl access, third-party corroboration, and measurable movement in citation share, never a specific placement on a specific date.
Are “submit your site to ChatGPT” services legitimate?
No. There is no submission process for any major AI engine, nothing to register and no queue to join. The real, free access points are Bing Webmaster Tools, which feeds ChatGPT search and Copilot, IndexNow for faster crawling, and Google Search Console for AI Overviews. Any vendor charging for “AI search registration” is billing for a fiction.
How much should real AEO work cost in 2026?
Serious engagements generally run $2,000 to $25,000 per month depending on scope, competitiveness, and how much content and PR work is included. Self-serve tracking tools like Otterly.ai start around $29 per month and platforms like Profound around $399 per month. Prices far below the service floor usually mean an automated report with no actual work behind it.
How do I verify an AEO agency’s claims before signing?
Three checks in one call: ask them to explain where AI answers come from mechanically, ask for a live demo of their own brand’s AI visibility, and ask what your baseline would be measured with. Then confirm their reporting uses dated, sampled citation share plus verifiable sources like the Bing AI Performance report and GA4 referral data.
What contract terms are reasonable for AEO in 2026?
Three to six month initial commitments with monthly deliverables and reporting against a documented baseline. Meaningful movement typically shows within eight to sixteen weeks, so twelve-month lock-ins without exit clauses shift all the risk to you. Refuse any contract where the measurement methodology is undefined or proprietary to the agency.
Can an SEO agency do AEO work?
Sometimes, the disciplines overlap heavily, but verify rather than assume. Strong SEO fundamentals, crawlability, structured content, and authority building all feed AI visibility. The gaps to probe are engine-specific mechanics, citation measurement, and off-site corroboration work like PR and community presence. The full comparison is in AEO agency vs SEO agency.
The bottom line
Every red flag on this list reduces to one principle: real AEO work is falsifiable and fake AEO work is not. Guarantees, submission services, proprietary rankings, and baseline-free proposals all share the same design, they cannot be checked. The vendors worth hiring hand you a dated baseline, name their sources, demo their own visibility, and sign terms short enough that the evidence has to show up. Hold every pitch to that standard and the market gets very small, very fast, which is the point. If you would rather see what disciplined delivery looks like scoped against your own gaps, our services page shows how we structure it.
Either way, do not walk into sales calls blind. Get the free AI visibility audit first, and measure every agency’s promises against a baseline you already own.
Tagged