Last updated: 17 August 2026

AI has stopped being a sidebar in search and become the main event. SparkToro's zero-click research found that 58.5% of US Google searches end without a click to the open web, a figure that climbs to 77% on mobile. ChatGPT reached 900 million weekly active users in February 2026, more than doubling in a year, according to TechCrunch. And according to AirOps' 2025 offsite signals research, brands are 6.5x more likely to be cited in AI answers through third-party sources than through their own websites. The statistics below cover how AI is reshaping search in 2026, who gets cited and why, and what the data says specifically for real estate.

Key AI SEO statistics at a glance

• 58.5% of US Google searches end without a click; 77% on mobile (SparkToro, 2024)
• Only 1% of users click a link inside a Google AI Overview (Pew Research Center, 2025)
• Clicks on the #1 organic result fall by 58% when an AI Overview appears (Ahrefs, December 2025)
• Brands are 6.5x more likely to be cited in AI answers via third-party sources than via their own domains (AirOps, 2025)
• Branded web mentions correlate with AI Overview appearances at 0.664, three times more strongly than backlinks at 0.218 (Ahrefs, 2026)
• Brands cited in AI Overviews earn 35% more organic clicks and 91% more paid clicks (Seer Interactive, 2025)
• ChatGPT hit 900 million weekly active users in February 2026, up from 400 million a year earlier (TechCrunch, 2026)
• 68% of Realtors now use AI in their business, but only 17% say it has had a significant positive impact (NAR 2025 Technology Survey, 2025)
• 82% of US adults interested in buying or selling a home use AI for housing-market information (Realtor.com, 2025)
• 74.2% of newly published webpages contain AI-generated content in some form (Ahrefs, 2025)
• Content with 5 to 7 statistics earns roughly 20% higher AI citation likelihood (AirOps, April 2026)
• There is less than a 1-in-100 chance ChatGPT surfaces the same brand list across 100 identical queries (SparkToro, January 2026)

Zero-click search is now the default

The single most important number in search is how often nobody clicks anything. SparkToro's landmark study put the zero-click share of US Google searches at 58.5% as of 2024, meaning that for every 1,000 searches, only about 360 clicks reach the open web. On mobile devices, 77% of searches end without a click.

AI Overviews accelerate the pattern. Pew Research Center's 2025 analysis of real user browsing behavior found four things worth memorizing. When an AI summary appears, only 8% of users click a traditional search result, compared with 15% when no summary is shown. Just 1% of users click any of the links cited inside the AI Overview itself. And 26% of searches that trigger an AI summary end with the user abandoning the session entirely, versus 16% on traditional results pages. The summary is not a preview of the answer; for most users, it is the answer.

The click erosion is heaviest at the top. Ahrefs' December 2025 update measured a 58% reduction in clicks on the #1 organic result when an AI Overview is present, up sharply from a 34.5% reduction in April 2025. Ranking first still matters, but the reward for that ranking is shrinking on any query an AI can answer directly, and the erosion accelerated through 2025.

Who gets cited by AI, and why

If clicks are drying up, the new competition is for citations: being the name an AI mentions when it answers the question. The data here is remarkably consistent, and it points away from your own website.

AirOps' 2025 research found that brands are 6.5x more likely to be cited in AI answers through third-party sources, such as press coverage, directories, reviews, and independent rankings, than through content on their own domains. Ahrefs' 2026 correlation study reinforces the mechanism: branded web mentions show a 0.664 correlation with AI Overview appearances, roughly three times stronger than backlinks at 0.218. The same Ahrefs research program found that only 38% of AI Overview citations now come from top-10 ranking pages, down from 76% in mid-2025. Read those three numbers together and the conclusion is hard to avoid: AI systems are decoupling citation from ranking, and the signal they trust most is other sites talking about you, not you talking about yourself.

Content structure matters too. AirOps' April 2026 analysis quantified what AI-quotable writing actually looks like: comparison pages with three tables earn 25.7% more ChatGPT citations, shortlist pages averaging ten or fewer words per sentence earn 18.8% more, and content containing five to seven statistics earns roughly 20% higher citation likelihood. AI systems reward pages that are easy to parse, quote, and verify.

Winning the citation pays off in traffic as well. Seer Interactive's 2025 measurement found that brands cited inside AI Overviews earn 35% more organic clicks (a 0.70% versus 0.52% CTR) and a striking 91% more paid clicks than brands absent from the answer. Visibility in the AI layer and traffic from the traditional layer are not competing goals; the first drives the second.

The scale of AI search in 2026

The audience asking AI for answers is no longer a niche. ChatGPT reached 900 million weekly active users in February 2026, up from 400 million in February 2025, per OpenAI figures reported by TechCrunch. That is a user base rivaling the largest platforms on the internet, and a meaningful share of those sessions are searches in everything but name: recommendations, comparisons, and "who should I hire" questions.

At the same time, AI is reshaping the supply side of the web. Ahrefs found that 74.2% of newly published webpages now contain AI-generated content in some form. The open web is filling with machine-written pages while AI assistants get pickier about which sources they trust, which is exactly why third-party corroboration and consistent entity data are pulling ahead of raw publishing volume as visibility signals.

One caution for anyone tracking their own AI visibility: SparkToro's January 2026 study found there is less than a 1-in-100 chance ChatGPT surfaces the same list of brands across 100 identical queries. AI recommendations are probabilistic. The goal is not to win one prompt; it is to be present across the full distribution of answers.

AI SEO statistics for real estate

Real estate is one of the clearest examples of AI search changing who gets hired, and it is the industry where the professional and consumer datasets diverge most sharply.

On the professional side, the NAR 2025 Technology Survey, drawn from a random sample of 49,233 active Realtors, found that 68% of agents now use AI in their business; for the first time, adoption has crossed the majority line. ChatGPT is the most common tool at 58% of surveyed agents, ahead of Gemini at 20% and Copilot at 15%. Usage is habitual for many: 20% of agents use AI daily and another 22% weekly, while 32% have not used it in their business at all. The most common application is AI-generated content such as listing descriptions, used by 46% of agents, and 21% use a CRM with AI-powered insights. Only 7% have deployed chatbots for lead capture or client communication.

But adoption has not translated into an advantage for most. In the same survey, only 17% of agents said AI has had a significant positive impact on their business, 33% reported a moderately positive impact, and 46% reported no noticeable impact. Two-thirds of the industry is using AI; fewer than one in five can point to a meaningful business result. The pattern in the data is that most agents use AI to produce content faster, while far fewer have addressed the other side of the equation: whether AI recommends them when a consumer asks.

That consumer side is where the shift is starkest. Realtor.com's 2025 AI and Housing Survey of 1,000 US adults interested in buying or selling found that 82% already use AI for housing-market information. The same survey found real estate agents still rank as the most trusted and accurate source of housing information, which cuts both ways: consumers arrive with AI-generated answers in hand, and the agents who appear in those answers inherit that trust before the first phone call.

Put the two datasets together, and the gap is obvious. Two-thirds of agents use AI as a writing tool. Four out of five consumers use AI as a research tool. Very few agents are optimizing for the moment those consumers ask an AI who to hire.

The mechanics of that moment follow the citation research above. When a buyer asks ChatGPT for the best agent in a market, the model leans on the same signals AirOps and Ahrefs measured: third-party mentions, consistent entity data across directories and press, and pages structured plainly enough to quote. An agent with a polished website but a thin third-party footprint is invisible in exactly the channel where 82% of their future clients are now doing research. The 6.5x third-party citation multiplier is not an abstraction for this industry; it is the difference between being the recommendation and being absent from it.

Emerging trends and what's new in 2026

Three developments define the 2026 data. First, the citation economy has matured, and it has detached from rankings: mentions correlate three times more strongly than backlinks with AI Overview appearances (Ahrefs, 2026), third-party sources drive 6.5x more citations than owned domains (AirOps, 2025), and the share of AI Overview citations coming from top-10 pages has been cut in half in under a year, from 76% to 38% (Ahrefs, 2026). The winning playbook now looks more like PR and entity building than traditional link acquisition.

Second, being cited is measurably profitable. Seer Interactive's finding that cited brands earn 35% more organic clicks and 91% more paid clicks turned AI visibility from a branding argument into a performance argument, and AirOps' structured data (three tables, short sentences, five to seven statistics per page) gives that argument an executable spec.

Third, the consumer behavior change is outrunning professional adaptation, and nowhere more than in local services like real estate, where 82% of interested buyers and sellers use AI for market information (Realtor.com, 2025) while only 17% of agents report a significant business impact from their own AI usage and 32% have not used it at all (NAR, 2025). The demand side has moved. Most of the supply side has not.

How FlyDragon helps

FlyDragon makes real estate agents the answer when buyers and sellers ask AI who to hire, using the same third-party citation and entity signals described in the research above. The free AI Visibility Scorecard shows what ChatGPT actually says about you in about a minute.

Before hiring an AI SEO agency, ask the agency to prove five things:

• that it has produced real results
• that it can explain how those results were achieved
• that it can measure what matters
• that its strategy fits your business and market
• and that it can distinguish established search practices from AI SEO experiments.

Before you hire, go beyond technical competence. Ask what you're buying, what it costs, who will do the work, how the agency will communicate, what you own, how long you're committed, what happens if you cancel, how success will be measured, and what the agency will do if visibility improves without producing a meaningful business result.

If an agency can't answer those questions clearly, with evidence, you’re taking an unnecessary risk.

You need another agency.

You should understand what you're buying, why the agency believes its work will make your business more discoverable, and how it intends to prove the work made a difference.

This guide helps you move from researching agencies to comparing proposals, conducting due diligence, and making a hiring decision.

How to Hire an AI SEO Agency in 7 Steps

Hiring an AI SEO agency is a procurement decision, not a technical quiz. Use the same sequence for every agency so you can compare like-for-like proposals and decide based on evidence, scope, and commercial fit.

1. Define the business outcome you want the agency to influence, not just the visibility metric you want to improve.
2. Establish your starting point so every agency is responding to the same baseline.
3. Set a realistic budget, internal resource commitment and decision timeline before you request proposals.
4. Shortlist agencies whose experience, case studies, and working model are relevant to your business.
5. Ask every shortlisted agency the same core questions about proof, methodology, measurement, pricing, contract terms, reporting and ownership.
6. Compare the proposals on scope, assumptions, exclusions, team, communication, commercial terms and the first 90 days — not just on the number of deliverables.
7. Choose the agency that can explain why its plan fits your business, document the agreed baseline and success criteria, and put the scope and exit terms in writing before you sign.

The 12 Questions to Ask Before You Hire an AI SEO Agency

You don't need to understand every technical detail of AI search before hiring an agency. You do need enough information to compare agencies on the same criteria, understand the commercial terms, and decide which proposal best fits your business.

These 12 questions are designed to do that.

The questions move in a deliberate order — from defining what you need and checking proof, into strategy, measurement and business context, then pricing, scope, contract terms, implementation, reporting, ownership and commercial accountability.

An agency that answers them well should leave you with enough information to decide whether to hire it: what it believes, what it intends to do, what you will pay, who is responsible, how the relationship works, how the work connects to your business, and what happens if you leave. An agency that answers them badly is not giving you useful information before you sign.

1. Can the AI SEO Agency Show You Results It Has Produced?

An AI SEO agency should be able to show real client results, explain what was measured, and demonstrate what happened before and after its work.

This is the first question I'd ask. Not how many articles you receive, not whether the agency uses proprietary software, not whether it has invented its own three-letter acronym.

Ask it to show you something it has done.

A credible case study should make it possible to understand where the client started, what problem the agency identified, what work was carried out, and what changed afterward.
If you're close to hiring, ask whether you can speak to a current or former client whose situation resembles yours.

A reference call helps you check communication, responsiveness, whether the promised team actually did the work, whether the case study matches the client's experience, and how the agency behaved when something didn't go to plan.

You should also ask whether the result lasted. AI responses aren't conventional fixed rankings; answers vary between platforms, between sessions, and over time.

A screenshot showing that a client was recommended by ChatGPT is useful evidence that the recommendation occurred, but it doesn't prove the agency created a durable competitive advantage.

Ask what happened a month later. Ask what happened three months later. Ask whether the business began appearing across a meaningful group of related questions, or whether the entire case study depends on one carefully selected prompt.

I'd also ask the agency to show me something that didn't work. An agency working seriously in a discipline this young should have hypotheses that failed. What matters is whether it can explain why it believed the test was worth running, what happened, and what it learned from the result. If every experiment an agency has ever run apparently succeeded, I get more skeptical, not less.

Third-party research — Seer Interactive's 2026 GEO RFP guide makes a similar recommendation: ask agencies to show live wins, explain uncertainty, discuss failed tests and connect measurement to business impact.

What Do Real FlyDragon AI SEO Results Look Like?

FlyDragon observed — We tell prospective clients to demand evidence, so FlyDragon should be held to the same standard.

Nate Clark's AI SEO case study is one example. Nate already had genuine subject-matter experience in a specific local service area in Austin, so rather than trying to associate him equally with every possible local query, the campaign concentrated on an area where he had a legitimate reason to be considered.

Our published case study reports that Nate became a leading recommendation across multiple AI platforms for probate-related local searches. The more useful evidence came afterwards: a prospect contacted Nate and told him she had found him after asking ChatGPT for the best local probate specialist.

Richard Berman's Reno case study shows a different shape of result. Our published case study reports that Richard received his first AI-sourced inbound opportunity within roughly two weeks and later averaged around two inbound opportunities per month, including one tied to a property valued as high as $1.6 million.

Ben Lang's case study documents a third variation: FlyDragon reports that Ben received his first AI-sourced inbound opportunity within roughly 30 days, and that one resulting client opportunity was associated with an approximately $890,000 transaction.

These are FlyDragon's own published client case studies. They aren't controlled scientific experiments, and they aren't guarantees that another business will produce identical results.

They're evidence that named clients experienced identifiable business outcomes while working with us — no more, no less.

2. What Exactly Did the AI SEO Agency Do to Produce Its Results?

This question exposes the difference between a search strategy and a content package faster than any other. Suppose an agency tells you your plan includes 20 articles, five backlinks, schema markup, and two press releases every month. You still don't know whether any of those things are what your business needs.

So ask why.

• Why does this page need to exist?
• Why is this existing page being changed?
• Why is the agency targeting this question rather than another one?
• Why is it chasing coverage from that particular publication?
• Why is it creating a new page rather than improving one you already have?

The answer should relate to a problem the agency has identified: your website lacks a satisfactory answer to an important customer question, a competitor is repeatedly recommended where you aren't, your business has strong evidence in one market and almost none in another, third-party sources contain conflicting information about you, important content is difficult to crawl, or several existing pages answer essentially the same question and compete with one another.

3. What Does the Agency Know, What Has It Observed, and What Is It Still Testing?

An AI SEO agency should distinguish established search requirements from its own observations and experimental tactics.

Professional opinion — This may be the most revealing question in the guide. Based on the agencies, audits, and sales material I review, I see the line between documented evidence, observation, and speculation handled poorly far more often than it should be.

There are parts of AI search we can speak about with relatively high confidence because the platforms document them. For Google, we know traditional SEO remains relevant to generative search.

Google says AI Overviews and AI Mode use its Search index and existing ranking and quality systems, and it has described techniques such as query fan-out as part of how its generative experiences locate supporting information.

Then there are things SEO practitioners observe repeatedly but can't responsibly describe as universal ranking factors. And then there are hypotheses worth testing because generative search keeps changing under our feet. Those three categories should never collapse into one category called "AI SEO best practices".

If an agency says a particular tactic is essential, ask where the conclusion comes from.

• Is Google documenting it?
• Has OpenAI said it?
• Has the agency measured it repeatedly?
• Is it an inference? Or is the team currently testing it?

A serious AI SEO agency should be comfortable saying:

"We think this is worth testing, but it isn't a documented requirement."

That's not uncertainty to be embarrassed about. It's precision. And in a field moving this fast, precision is the scarcest thing an agency can sell you.

Has AI SEO replaced traditional SEO?

No, AI SEO hasn’t replaced traditional SEO.

AI search changes how information can be discovered, combined, and presented to a user, but it doesn't make the foundations of search disappear. Google's current documentation is explicit that existing SEO best practices remain relevant to AI Overviews and AI Mode.

That doesn't make AI SEO and conventional SEO identical, though.

A conventional search sends the user to a list of pages to investigate themselves, while a generative system can do more of that investigation inside the interface and may present particular brands, sources, or recommendations directly in the answer.

That introduces genuinely new things worth measuring. What it doesn't do is turn someone with no understanding of crawling, indexing, relevance, authority or search intent into an expert because they learned the word GEO.

4. How Will the Agency Decide Which AI Searches Matter to Your Business?

An AI SEO agency should choose the questions it tracks and targets according to real customer demand, commercial relevance, and the decisions people make before buying.

Prompt tracking becomes meaningless surprisingly quickly. Give a model a service, a city, and a company, and it can generate hundreds of syntactically plausible questions in under a minute.

Professional opinion — Many AI visibility dashboards and agency reports rely heavily on synthetic prompt sets. Those sets can be useful for controlled tracking, but they become misleading when a generated prompt list is presented as evidence of actual customer demand. The better question is where the agency's target questions came from.

This gets more important when a business has several audiences, products, services, locations, or buying conditions.

"What is the best CRM software?" is an obvious search, and it isn't the whole decision.

One buyer needs a CRM for a five-person sales team. Another needs enterprise permissions and security. Another requires a specific integration. Another cares most about implementation support, and another needs evidence that the software works for a particular industry or workflow.

Each additional condition changes which business or product is the most relevant answer. Your agency should identify the circumstances in which you have a legitimate reason to be recommended, which is worth far more than trying to attach your brand to every question a model can imagine.

5. How Will the Agency Measure AI Visibility and Business Impact?

An AI SEO agency should establish your starting position before major work begins and should measure visibility separately from citations, website traffic, leads, and revenue.

Without a baseline, almost any future movement can be packaged as improvement, and it will be.

Before substantial work begins, I want to know which businesses are being recommended, how often my client appears, which questions produce those recommendations, which sources are influencing the answers, and whether incorrect information about the business is already circulating. Then, and only then, can we compare against something real.

Platform documented — Google now provides a dedicated Generative AI performance report in Search Console for eligible sites, giving site owners first-party visibility data for Google's generative Search features.

The language used in reporting matters just as much. Being mentioned and being recommended are different outcomes.

Being cited and receiving a website visit are different outcomes. AI traffic and a qualified lead are different outcomes, and a lead and a won customer are different again.

Those things influence one another without being interchangeable, and an agency that blurs them in reporting is either sloppy or hoping you won't notice.

The eventual question is simple: are more of the right customers discovering and considering this business in AI search?

6. What Experience Does the Agency Have in Your Industry, Business Model and Market?

Industry, business model, and market experience matter because the customer journey, sales cycle, evidence sources, competitors, conversion paths, and entities involved change substantially between businesses.

Platform documented — Google recommends asking prospective SEOs about their experience in your industry and in your country or city.

Professional opinion — I think this matters even more in AI search because the brand, product, founder, parent company, locations, software stack and customer segments can all become separate entities or retrieval contexts. An agency needs to understand which one the campaign is actually trying to strengthen.

A strategy designed for a SaaS company can't simply be copied onto a local service business, ecommerce brand or professional firm by swapping a product keyword for a different commercial term.

The entity you want recommended may be a product, founder, location or service line, while the website may primarily represent the parent brand.

• That product or business unit may operate under a parent company.
• The parent company or platform vendor may control portions of the website.
• A key conversion experience might run through a separate application, subdomain or third-party platform
• Reviews, product data, or brand information might be split across profiles and systems,
• And the business may serve several products, locations or customer segments that buyers treat as completely separate markets.

An agency that doesn't understand those relationships will end up producing content around a business it never properly defined.

7. What Will This Cost, What Are You Committing To, and What Happens If You Leave?

Before hiring an AI SEO agency, get the commercial scope in writing: the fee, what is included, what is excluded, contract length, renewal terms, cancellation rights, ownership, pass-through costs, implementation responsibilities, and whether the agency can work with a direct competitor.

This is where many AI SEO buying guides are incomplete.

They explain what the agency should know, but not what the buyer is actually agreeing to. A good hiring decision requires both technical confidence and commercial clarity.

What will this cost, and what is not included?

Ask for the full commercial picture, not just the headline monthly fee. You should understand setup fees, recurring fees, software costs, pass-through expenses, implementation charges, content or PR costs, minimum spend, and which requests are considered out of scope. Two proposals with the same monthly price can create very different total commitments.

How long am I committing for, and how do I cancel?

The proposal should state the initial term, renewal structure, cancellation notice, early-termination terms, refund policy and what happens to unfinished work if the relationship ends. Read the contract against the sales conversation; anything important enough to influence your decision should exist in writing.

What guarantees are you making?

Treat guarantees as commercial claims that need precise definitions. Ask exactly what is guaranteed, what is merely a target, what assumptions the promise depends on, and what remedy exists if the guarantee is not met. A promise of activity is not the same thing as a promise of visibility, and a promise of visibility is not the same thing as a promise of revenue.

What exactly is included in the scope I'm buying?

Ask the agency exactly what the fee covers. A strategy retainer, content production, technical implementation, digital PR, reporting, software access, and consulting are not automatically the same scope, and the contract should not leave the answer to interpretation.

If you sign a contract for "AI visibility," you should know which platforms are included, which business units or markets are covered, how much implementation is included, whether off-site work is part of the fee, and which deliverables create additional charges.

The same principle applies to any scope boundary.

Two products, locations, service lines or customer segments may be related without being the same commercial target. The agency needs to understand the distinction for its strategy. You need to understand it for your contract.

Which brand, product, person, location, or business entity are you trying to make more visible?

Suppose Acme Group owns Acme Software and a separate consulting division. Acme Group, Acme Software, and the consulting business are related, and they are three separate entities.

If the goal is for Acme Software to become the answer to questions about a particular software category, the online evidence has to make the product's relationship with the parent company, other business units, target customers, and market unambiguous.

That decision cascades into product pages, company information, reviews, third-party coverage, authorship, structured information, and how the business is described everywhere else online.

An agency selling "entity optimization" should be able to tell you which entity it's trying to strengthen.

Does my website or technology stack create technical limitations?

Business websites range from highly flexible custom builds to hosted platforms where the company controls almost nothing.

Important content may be rendered in JavaScript, search or product experiences may live on another subdomain, templated pages may be duplicated widely, key page types may use restrictive templates, and a platform vendor may control the site's technical elements outright.

None of that automatically makes a website unsuitable, but it does mean the agency should inspect it before promising anything.

Platform documented — Google says crawlability and indexability remain fundamental to its generative Search experiences. OpenAI says publishers should not block OAI-SearchBot if they want site content discoverable and cited in ChatGPT Search.

An AI SEO agency has to understand both the marketing strategy and the infrastructure it's being asked to work with, because one regularly breaks the other.

How will you prove that my business is genuinely relevant to the customers and topics we're targeting?

Adding a target phrase to hundreds of pages isn't expertise. A business's strongest evidence comes from real work: customer outcomes, product or service data, client reviews, original video, subject-matter knowledge, case studies, first-party research, and reputable third parties that independently associate the business with the topic or market.

The strongest evidence varies by business: one company may have a long performance record in a narrow vertical, another may have unusually deep technical expertise, and another may have served hundreds of customers with a specific problem or use case.

The agency's job is to identify those real attributes and make the underlying evidence easier to find and understand. It should never manufacture expertise the business doesn't possess, and you should walk away from any agency that offers to.

How will you handle different products, services, audiences, or specialties?

"Best accounting software" doesn't represent every accounting-software search.

A buyer might need software for a solo operator, a multi-location company, an enterprise team, a regulated industry, a specific integration, or a particular workflow, and each situation involves different criteria and produces a different "best."

The underlying principle is the same in every industry: the agency should identify the conditions under which your business has a legitimate reason to be the answer.

The opportunity is rarely to compete for every broad query in a category. Existing expertise, product strengths, customer outcomes, geography, integrations, price point, or use case can create a far more meaningful area around which to strengthen relevance.

A good AI SEO agency finds those legitimate areas of differentiation before it writes a single word.

Will you work with a direct competitor, and how do you handle conflicts?

Not every AI SEO agency needs to offer exclusivity. You simply need to know what you're buying. If an agency offers category, territory, market or account exclusivity, the contract should define that boundary precisely. If it does not offer exclusivity, ask how it handles direct competitors, confidential research, shared tactics and conflicts of interest before work begins, not after the first invoice.

8. What Will the AI SEO Agency Change on Your Website?

Website changes should follow an audit of technical accessibility, existing content, search intent, information architecture, and factual consistency rather than starting with a predetermined article quota.

Technical SEO still matters. Google's current guidance for generative Search still emphasizes the same basic conditions that matter everywhere else in Search: content needs to be crawlable, indexable, and useful.

Generative AI has made content production close to free, and that has broken a lot of agencies' incentive structures.

More pages don't automatically improve a website. If your site already contains four pages answering substantially the same question, adding a fifth because the agency owes you another article that month creates more duplication, not more value.

Sometimes you need a new page. Sometimes an existing page needs more depth. Sometimes several weak pages should be consolidated into one strong one.

A strategic agency can explain the difference and show you which decision it made, page by page.

Where will the unique information come from?

This is one of my highest-value questions, and it's disarmingly simple: if every article can be produced by typing its title into ChatGPT and publishing the response, what information has your business contributed?

Google's current generative-search guidance specifically encourages unique, non-commodity content rooted in expertise and first-hand experience — which is exactly the content a title-into-ChatGPT workflow can't produce.

For most businesses, that information already exists inside the operation. It's in sales calls, support tickets, product questions, implementation lessons, deals that went wrong before they went right, customer objections, recurring problems and the specialist knowledge the team uses every day.

It's in client conversations, market or product data, photographs, video, pricing discussions, support logs, customer research, internal documentation and professional opinions formed through years of doing the work. A strong agency knows how to get that knowledge out of the business and onto the page.

9. How Will the Agency Build Authority Outside Your Website?

A good off-site AI SEO strategy should identify credible third-party sources relevant to your topic and market rather than assuming that every backlink or brand mention has equal value.

Your website isn't the only source describing your business. Search and AI systems can find you through publications, professional profiles, news coverage, business directories, customer reviews, video, industry sites, forums, and dozens of other places, which makes external information genuinely important.

It does not make "get mentioned everywhere" a strategy. And Google now specifically warns businesses against pursuing inauthentic mentions as a tactic for generative Search, which should tell you where that game ends.

So I'd ask an agency why each external source matters.

1. Does it cover the topic?
2. Is it relevant to the market?
3. Does it already appear around the kinds of questions we're researching?
4. Is it trusted by real users?
5. Does it contain meaningful information about competitors?
6. Would a potential customer plausibly encounter it while researching the decision?

A high authority score from an SEO tool answers none of those questions by itself. For a regional or specialist business, a respected publication that genuinely covers that market or industry can provide far more contextual value than an unrelated national website.

The same principle applies to communities. Real customer discussions and legitimate professional participation can help with ranking in AI search models.

10. Which AI SEO Claims, Guarantees or Sales Promises Should You Ask the Agency to Prove?

The strongest claim in an AI SEO sales pitch is often the best place to begin your due diligence.

Several claims are repeated online so often that people assume they're established facts. Some aren't, and the gap between what gets repeated and what gets documented is where a lot of retainers go to die.

Common AI SEO claims worth challenging

Professional opinion — The correct response to uncertainty isn't to stop experimenting. It's to label experiments accurately. I expect agencies that present unverified tactics as requirements to move on to the next fashionable tactic when the evidence changes; that is exactly why the evidence label matters.

11. Who Will Perform the Work, What Will You Own and What Happens After You Sign?

Before hiring an AI SEO agency, you should know who makes the strategic decisions, how the work is produced, which assets remain yours, and what happens once the contract begins.

There's nothing inherently wrong with contractors, automation, or AI-assisted production — I use automation extensively myself, and I'd be a hypocrite to pretend otherwise.

The important thing is knowing where responsibility sits.

• Who decides what needs to be published?
• Who reviews the strategy?
• Who catches factual contradictions?
• Who decides whether an experiment failed?
• Who handles technical implementation, who determines which external sources are worth pursuing, and who is accountable if the work doesn't achieve what was intended?

Those decisions need names attached to them. Google's own hiring guidance encourages businesses to ask prospective SEOs how they communicate, how they'll explain changes, and whether they'll share the reasoning behind recommendations, and the same standard should apply to AI SEO without exception.

You should also understand what happens if you leave. The contract should make it clear who owns the website, content, research, accounts, tracking, data, and every other asset created during the relationship. AI search may be new.

Vendor lock-in isn't.

Who will actually work on my account?

Ask who owns strategy, who performs implementation, who writes or edits content, who handles technical work, who manages off-site activity and who is your day-to-day contact. The salesperson who wins the account is not necessarily the person who will run it.

How often will we communicate, and what will the reporting include?

Agree the meeting cadence, reporting frequency, decision-making process and escalation path before you sign. Reports should show what changed, why it changed, what happened afterwards, what the agency learned, and what it plans to do next — not just a dashboard full of metrics.

What work is done by people, AI, automation or contractors?

There is nothing inherently wrong with AI-assisted work, automation or contractors. The hiring question is where human judgment enters the process, who reviews the output, who is accountable for errors, and which parts of the work are outsourced or automated.

What access and permissions will the agency need?

Ask which systems the agency needs to access, whether it requires administrator permissions, how credentials are handled, which third-party tools will receive your data, and how access is removed when the relationship ends. Technical access is an operational and security issue as well as an SEO one.

What should the first 90 days look like?

There is no universal 90-day plan that's correct for every website, and that's precisely the point.

A technically broken website shouldn't receive the same first-month priorities as a site with excellent technical foundations but poor topical coverage, and an established brand with hundreds of reviews and strong recognition shouldn't be treated the same as a newer business whose information is inconsistent across the internet.

The first phase should establish what exists now. The next phase should address the most important problems or opportunities identified during that diagnosis, and the agency should then evaluate what changed and adjust. If the exact same timeline appears in every proposal regardless of the website or client, ask why.

12. What Happens if AI Visibility Improves but Your Leads Don't?

If AI visibility increases without producing meaningful business results, the agency should investigate the quality of the visibility rather than declaring success because its reporting dashboard improved.

This is where vanity metrics get expensive.

The business may be appearing for queries with very little buying intent. The wrong audience, product, service line, or market may have improved.

The parent brand may be gaining visibility when the commercial objective was to promote a specific product or business unit. Recommendations may appear too inconsistently to influence many buyers; the website may receive more attention but convert poorly, or leads may be arriving without attribution, making the connection difficult to prove.

It's also possible the campaign simply isn't creating a commercially meaningful outcome and an agency has to be willing to investigate that possibility too, out loud, in your reporting.

This carries extra weight in high-value or long-sales-cycle businesses, because traffic volume can be a poor proxy for commercial value.

A business doesn't always need thousands of website visits for AI SEO to matter; a single qualified enterprise lead, high-value customer or large contract can be commercially significant, which makes relevance far more important than traffic volume.

The ultimate question is whether the strategy is making the right potential customer more likely to discover the business, trust it, and start a commercial conversation.

What Should You Ask FlyDragon Before Hiring Us?

You should ask FlyDragon the same difficult questions in this guide before hiring us. We shouldn't be held to a different standard because we wrote it.

If you're considering working with us, don't ask me only for FlyDragon's biggest success story.

• Ask FlyDragon to show you a client whose situation resembles yours. Ask FlyDragon what currently appears when customers research your category, and exactly how they define the audience, market, product or scope you're paying us to target.

• Ask FlyDragon whether they believe your website is a genuine limitation or whether it can be improved without rebuilding it.

• Ask FlyDragon which parts of our approach are based on established SEO principles, which are based on observations from the campaigns we're running, and which are still experimental.

• Ask FlyDragon how we'll know whether somebody contacted you because of the work, who does the work, whether another FlyDragon client can compete directly with you, and what we'll do if the metrics improve but the commercial outcomes don't.

Probably the best question you can ask FlyDragon is:

"Show me why you believe this will work for my business.”

FlyDragon should be able to answer that using your business, not ours. They should understand who you are, where you operate, what you're genuinely good at, which competitors are currently stronger, what your website currently communicates, what external evidence already exists, and where the biggest opportunities are.

Review our AI SEO case studies and our GEO and AI SEO services before speaking with us. Then ask us about them.

AI SEO Agency Hiring Checklist

Use this checklist during or immediately after an agency sales call. The purpose isn't to find an agency that gives you the answer you want to hear on every line it's to make sure you've received enough evidence to understand what you're buying.

• The agency showed me a real result rather than relying only on anonymous charts or screenshots.
• I understand what the agency changed to produce that result.
• The agency can show whether its results persisted over time.
• The agency can discuss experiments that failed as well as successful ones.
• It distinguishes documented platform guidance from observations and experimental tactics.
• I understand how it chooses the AI questions and prompts that matter to my business.
• The questions being tracked reflect real customer and commercial intent.
• A baseline will be established before major work begins.
• The agency distinguishes mentions, recommendations, citations, referrals and leads.
• It can explain how increased AI visibility is expected to contribute to business outcomes.
• It understands my industry rather than applying a generic AI SEO template.
• It understands the geographic market in which I'm competing.
• The agency can distinguish my brand, products, people, locations and business units when those entities overlap.
• My website and technology stack will be technically assessed rather than assumed to be suitable or unsuitable.
• The strategy will use genuine expertise and first-party information rather than producing commodity AI content.
• The agency has explained how it decides whether to create, update or consolidate content.
• Its off-site strategy focuses on relevant and credible third-party sources rather than raw mention volume.
• The agency does not present llms.txt, special AI schema or content chunking as universal requirements.
• I know who will make strategic decisions on my account.
• I know what I own if the relationship ends.
• I understand the total recurring cost, one-off fees, pass-through expenses and what is excluded from the scope.
• I understand the initial contract term, renewal structure, cancellation notice and early-termination terms.
• I know how often we will meet, how often I will receive reports, and what those reports are expected to explain.
• I understand which parts of the work are performed by employees, contractors, AI or automation and who reviews the output.
• I understand which systems and permissions the agency needs, how credentials and data are handled, and how access is removed if we stop working together.
• Any guarantee or performance commitment that influenced my decision is defined clearly in writing.
• I understand how the first phase of work will be prioritized.
• If exclusivity matters to me, the category, territory, market or competitive boundary has been clearly defined.
• There is a reasonable plan for attributing calls, forms or opportunities back to the work.
• The agency has explained what happens if visibility increases without leads.

If you can't confidently check most of those boxes after speaking with an agency, you haven't been given enough information to make the decision yet. And if the agency gets uncomfortable because you're asking these questions, that's useful information too.

How Do You Choose Between Two Good AI SEO Agencies?

If two AI SEO agencies survive the questions above, compare them across four things: understanding of your business, quality of the proposed strategy, strength of the evidence, and commercial terms. I would not choose the agency that simply promises the most work.

A good agency should be able to explain why your existing visibility looks the way it does, which customer questions matter commercially, where competitors have stronger evidence, what problems are holding the website back, and why its proposed priorities make sense in that order rather than another.

Then compare the price, scope, exclusions, contract length, cancellation terms, reporting cadence, account team and implementation responsibilities side by side.

I'd put additional weight on whether the agency understands your business model, can define the commercial scope precisely, understands your website and technology stack, can distinguish the correct brand or product entity, and can show clients receiving real business outcomes rather than only improved visibility scores.

I’d also ensure the agency has an experienced SEO behind the strategy, with proven success that has continued for years.

Should You Hire an AI SEO Agency at All?

Not every business needs to hire an AI SEO agency immediately.

If your website has serious technical problems, your brand information is inconsistent, you don't have a clear offer, or customers can't validate basic facts about the business, those problems may deserve attention first.

AI search doesn't make weak fundamentals disappear. In some cases, it exposes them more clearly than classic Search ever did, because the model reads every inconsistency across the web and picks the version it trusts.

The same applies across industries.

A new business with almost no reviews, no meaningful website, unclear positioning, inconsistent brand information, and little evidence of expertise may need to establish those assets before investing heavily in sophisticated AI visibility work, and a good agency should be willing to tell you that on the first call.

Google makes a similar point in its own hiring guidance: many small businesses can handle meaningful parts of SEO themselves before deciding they require specialist outside help. The agency that tells every single prospect they urgently need GEO may be excellent at selling GEO.

Check What AI Says About You Before You Hire Anyone

Before you pay FlyDragon or any other AI SEO agency, establish your starting point.

Our AI Visibility Scorecard lets supported businesses see how their brand currently appears across important AI searches and compare that visibility with competitors in the same market or category.

Instead of entering an agency sales call asking:

“Can you get me ranked in ChatGPT?"

You can ask:

"This is what AI currently says about me. Why is this happening, what would you change, and how will you prove that your work improved it?"

That is a much harder question to bullshit.

And it's probably the best question you can ask when you’re hiring an AI SEO agency.

About the Author: Why I'm Qualified to Write This

I'm Ryan Darani, Co-Founder and Chief AI Strategist at FlyDragon. I've worked in organic search for more than 12 years — long before AI SEO, GEO, and AEO became services people could sell.

You don't have to take FlyDragon's word for that.
Business Insider has published my work on SEO and describes me as a search professional who has generated millions of dollars in revenue for online businesses. In a separate profile about my career, Business Insider stated that it had verified my consultancy revenue with documentation and recorded that, before going independent, I had already generated millions of dollars in revenue for brands in finance, travel and retail through organic search.

My work has also been examined publicly by other SEO practitioners — one of my SEO campaigns in the highly competitive health and YMYL space was documented as a two-million-click case study, showing the strategy behind growing a health website in one of the most difficult areas of organic search. In 2023, Rise at Seven appointed me Search Strategy Director and publicly referenced my decade of search experience and previous work with brands including Lloyds, the NFL, Aldi, AO, Superdry and Lastminute.
That history matters in the context of AI SEO. I didn't learn search after ChatGPT launched. I've spent more than a decade ranking websites, studying how people search, building organic-growth strategies, working across technical SEO and content, and connecting search visibility to commercial outcomes.

How FlyDragon Labels Evidence in This Guide

FlyDragon uses a formal evidence standard so readers can tell whether a claim comes from a platform, FlyDragon's own work, outside research, an active hypothesis, or professional judgment. Google's June 2026 guidance on third-party SEO advice recommends this same discipline: good advice should either qualify empirical or opinion claims as based on data or experience, or substantiate them with official Google Search guidance.

Platform documented — Directly stated by the platform or its official documentation. We link to the primary source.

FlyDragon observed — Observed in FlyDragon campaign data, prompt tracking or client outcomes. Where practical, we state the sample size, time period and measurement method.

Third-party research — Supported by research, reporting or testing from another organization. We link to the original research rather than repeating the conclusion without attribution.

FlyDragon hypothesis — A current working theory or test. It is not presented as a confirmed ranking factor or platform requirement.

Professional opinion — My interpretation based on more than 12 years in SEO and current AI-search work. It is explicitly opinion, not a claim about how a platform says its system works.

If we cannot explain which category a meaningful claim belongs to, we should not present it as established fact.