AI SEO for Marketing Agencies: How to Get Recommended by ChatGPT and Google AI Overviews
A marketing agency gets recommended when AI tools can find clear, checkable facts about who it serves, what it charges and what it has achieved. Publish plain-text service, pricing and proof pages, then earn mentions on third-party lists and review sites. Buyers ask 'how do I choose' questions, so write the honest selection criteria yourself and show where you fit.
- In Ahrefs' test, a brand's own blog list was recommended in 17.2% of agency prompts and its landing page in 30.4%. 1
- In 6sense's 2025 survey, 95% of the time the winning vendor was already on the buyer's Day One shortlist. 4
- Profound found brand-owned sites supply 72% of citations in marketing, against about 57% overall, so your own pages matter. 5
- Publish pricing, who you serve, process and dated proof as plain text. Name the clients you may name.
- Never write your own reviews or promise results: both create legal and trust risk. 89
How does a marketing agency get recommended by ChatGPT and Google AI Overviews?
Pricing, niche and proof are what a buyer checks first, so those are the facts an agency needs AI tools to find. 'AI engines' here means ChatGPT and Google's AI Overviews and AI Mode. To appear as a link in Google's AI features, a page must be indexed and eligible to show with a snippet 6.
Google also says no special markup or AI text files are needed 6. Its generative AI guide asks for original content based on your own knowledge and experience 7. For an agency that means your own cases, your own method and your own prices, not a summary of what others say.
Agencies are a special case, because you sell the skill you are trying to prove. Buyers expect you to be findable. Being invisible in the very channel you sell is a credibility problem as well as a lead problem.
- Buyers use AI to research vendors: 45% in Gartner's survey and 94% in 6sense's, though neither is agency-specific 34.
- Winners usually come from the Day One shortlist (95% of the time in 6sense's study), so early naming matters 4.
- Brand-owned sites supply 72% of citations in marketing, against about 57% overall, so your own pages carry weight 5.
- An agency that sells search or AI skills but cannot be found by AI loses credibility with buyers.
How do buyers use AI to choose a marketing agency?
They use it to research and shortlist, then check with people. A Gartner survey of 645 B2B buyers (August to September 2025) found 45% used generative AI in a recent purchase, mainly to research vendors, and 69% prefer to validate AI-generated insights with a sales rep 3. Buyers also used an average of seven information sources 3.
6sense surveyed nearly 4,000 buyers in 2025 and found 94% use large language models in the buying process, mostly mid-journey to compare vendors 4. In that study the winning vendor came from the Day One shortlist 95% of the time 4. If AI helps build that shortlist, being named early matters.
Read these figures carefully. Neither report is about agencies. 6sense sells buyer-intent software, its data is self-reported, and it says services purchases run about two months longer than software 4. Gartner and 6sense cover B2B purchases in general.
What do AI engines look at when they recommend an agency?
They read your own pages, and they read lists written by others. Profound analysed 11.84 billion citations across 29 industries from April to July 2026. It found brand-owned sites supplied 72% of citations in marketing, and about 57% overall 5. That suggests your own site carries more weight than many agencies assume.
Third-party lists still matter. Ahrefs tested 750 top-of-funnel prompts across software, products and agencies, such as best web design agencies in London. 'Best X' blog lists made up 43.8% of the page types ChatGPT used. Non-blog lists such as G2 and Clutch, and social posts, were also counted as page types across all three prompt categories 1.
Recency also showed up. Of 1,100 cited lists, 79.1% had been updated in 2025, and 57.1% had been updated since first publication 1. Both studies are published by firms that sell related tools, and Ahrefs notes limits, such as prompt wording that may not match real chat behaviour 1.
Should your agency publish its own 'best agencies' list?
Only if you can make it honest. In Ahrefs' 750-prompt test, a brand's own blog list was among the first-party pages recommended in 17.2% of agency prompts and a general landing page in 30.4% (the denominator is not stated, and one answer can cite several page types) 1. The study's author concludes agencies should publish such lists, and finds no ranking or citation downside, but he also quotes agency owner Wil Reynolds, who says listing yourself at the top of your own list does not build trust with buyers 1. Trust is a human-reader risk the data does not measure.
A fair list names real rivals, states criteria, shows where each is strongest and carries a date. A list that always puts the publisher first reads as an advert to a human buyer. Ahrefs itself publishes comparison posts sparingly, links to alternatives and highlights third-party wins 1.
Our decision rule is below. It is our judgement, not a tested result.
| Situation | What to do | Why |
|---|---|---|
| You serve a clear niche, such as B2B SaaS startups | Publish 'how to choose an agency for [niche]' with criteria first | Buyers ask this exact question, and criteria are quotable |
| You would rank yourself first of ten | Do not publish a ranked list | Some buyers distrust self-ranking, as one agency owner quoted by Ahrefs says 1 |
| You can name three real rivals and cite their strengths | Publish a dated comparison with a row where a rival wins | It reads as evidence, not advertising |
| You cannot back a claim with a file | Cut the claim | Objective claims need a reasonable basis 9 |
What searches does an AI run before it names an agency?
ChatGPT searches only some of the time. Ahrefs found 14.4% of 66.6 million free, logged-out responses in August and September 2026 contained fan-out queries, about 1.6 each 2. 'Fan-out' means the extra searches an AI runs behind one prompt. In September, '2026' appeared in 22.9% of English fan-out queries, 'official' in 13.1%, 'price' in 8.9% and 'reviews' in 3.6% 2.
The list below is modelled on those patterns, not captured from a real run, and shows what to prepare for. Test your own prompts with the Query Fan-Out Explorer.
- Prompt: how do I choose a marketing agency for a B2B startup
- Likely search 1: how to choose a B2B marketing agency 2026
- Likely search 2: B2B startup marketing agency pricing retainer
- Likely search 3: questions to ask a marketing agency before hiring
- Likely search 4: best B2B marketing agencies for startups reviews
- Likely search 5: [Agency name] official case studies
- What to publish: a selection guide, a pricing page, a process page, dated case studies and a review-site profile
What should a marketing agency publish first?
Publish the pages a cautious buyer needs to build a shortlist. Services purchases are large and slow, so buyers look for reasons to rule you out. Give them plain answers before they ask, and put each answer in the first line of its section.
Start with what you charge, whom you serve and how you work. Then add proof that a stranger could check, with the client's permission.
- A pricing page with ranges or starting prices, what is included and what is extra.
- A 'who we serve' page per niche, stating team size, stage and budget fit, and who we do not suit.
- A selection guide: how to choose an agency, with criteria that favour no one.
- A process page: first 90 days, reporting rhythm, who does the work.
- Dated case studies: starting point, actions, measured result, limits and a named source.
- Team pages with real names, roles and verifiable backgrounds, as Google recommends for bylines 11.
What rules apply to agency claims, case studies and reviews?
In the US, the FTC expects advertisers, including ad agencies, to have a reasonable basis for objective claims before publishing them 9. The policy statement is dated 23 November 1984. 'We grew traffic 300%' is an objective claim, so keep the data, the date range and the client's consent.
The FTC's rule on consumer reviews took effect on 21 October 2024. It targets fake reviews, incentives tied to a positive opinion, undisclosed insider reviews, and false claims that a site you control gives independent reviews 8. Do not write your own testimonials, and do not run a 'best agencies' site while hiding that you own it.
Pages that use LocalBusiness or Organization markup are ineligible for star ratings when the agency controls reviews about itself, including through a widget 10. Google warns against fabricated authorship, fake headshots and false credentials 11. For an agency, a made-up case study is both an ethics failure and an easy one to disprove. This is general information, not legal advice.
How do you measure whether AI is sending agency leads?
Track two things: what AI says about you, and what leads say about how they found you. Write ten buyer prompts, such as the seed prompt above, and run each three times on two engines every month. Record whether you are named, how you are described and which sites are cited. The AI Prompt List Builder can draft the list.
Add a free-text field asking how the buyer found you to every enquiry form, and ask the same on discovery calls. AI visits often show up as direct or unknown traffic, so the buyer's own words fill the gap. Google reports traffic from AI features inside the Performance report under the Web search type 6.
Compare month to month rather than run to run, because AI answers vary between runs. See how to measure AI share of voice.
What have Chalk Labs and Rahil Jain done for marketing agencies, and what have they not?
Chalk Labs is itself a marketing agency, so we compete with many readers and apply this method on our own site. We have not published AI-citation or lead results for it, so we make no claim about them here.
What we can point to is client work in other fields. Chalk Labs reports that its programme took Meru Trophy, a Bengaluru trophy shop, to 27,300 Google clicks and 2.3M impressions in 16 months, and a new personal site to the top result for its niche phrase within a month 12. Those are Google search numbers, not AI-citation counts, and not agency results.
Rahil Jain runs Chalk Labs' AI SEO programme, and Chalk Labs sells AI SEO services. We have a commercial interest in agencies caring about this topic.
Your first 90 days
Days 1-30
- Write ten buyer prompts by niche and run them on two AI engines, three times each.
- Log which agencies are named and which sites are cited.
- Check that service, pricing and case-study pages are plain HTML and AI crawlers are allowed.
- Add 'How did you hear about us?' to enquiry forms and discovery calls.
Days 31-60
- Publish a pricing page and one 'who we serve' page for your best-fit niche.
- Rewrite two case studies with starting point, actions, dated results and limits.
- Claim or update your profiles on the review and list sites your buyers use.
Days 61-90
- Publish a selection guide with criteria that do not favour you.
- Rerun the prompt log and compare it with day one.
- Decide which pages to refresh monthly, given how often AI favours updated lists 1.
The first pages to publish
- Pricing page with ranges and inclusions. 'Price' appeared in 8.9% of ChatGPT fan-out queries in September 2026 2.
- How to choose an agency for [your niche]. It answers the selection prompt, and criteria are easy for an AI to quote.
- Who we serve, and who we do not. Niche pages match 'for a B2B startup' style prompts.
- Process and first-90-days page. Services buyers want to see how work runs before they sign.
- Case studies with dated, sourced results. Proof you can verify beats claims that need a reasonable basis 9.
- A fair comparison or alternatives page. Comparison questions pull in outside sources more than factual ones 2.
- Team and expert pages. Google recommends clear bylines and warns against false credentials 11.
- A reviews page linking to third-party profiles. Self-controlled reviews on Organization or LocalBusiness markup are not eligible for star ratings 10.
Trust signals and risks
What AI looks for
- Named clients and dated case studies with permission.
- Public pricing or starting prices.
- Named team members with verifiable backgrounds 11.
- Profiles on third-party lists and review sites such as Clutch and G2.
- A written method and reporting sample.
- Honest 'who we do not suit' statements.
What can go wrong
- Case-study numbers without a reasonable basis or client consent 9.
- Self-written or incentivised reviews, and hidden ownership of 'best agency' sites 8.
- Self-ranked lists that put the publisher first, which one agency owner quoted by Ahrefs says costs buyer trust 1.
- Mass-made 'agency in [city]' pages that Google may treat as scaled content abuse 7.
- Stale prices and team pages that AI keeps repeating.
Why Rahil Jain for Marketing Agencies?
Direct experience
Chalk Labs has done work in this field and can show it.
Chalk Labs is itself a marketing agency and runs this method on its own site, so we apply it ourselves. We have not published AI-citation or lead results for our own site. Client proof is Meru Trophy, a Bengaluru trophy shop, and shilikajain.com, both measured in Google search.
Rahil Jain
Rahil Jain is an AI SEO (answer engine and generative engine optimisation) consultant and co-founder of Chalk Labs.
Results we can show
These are Chalk Labs' own search results. They are not a forecast for your site: results depend on your market, competition and starting point.
How to read these numbers
Chalk Labs reports them from Google Search Console data. We cannot publish the raw console. The shilikajain.com site belongs to Shilika Jain, a Chalk Labs co-founder, so treat it as a demonstration on our own network and not as an independent client result. Meru Trophy is a client of Chalk Labs.
Frequently asked questions
Can a marketing agency get recommended by ChatGPT?
Yes, but no agency controls it. ChatGPT searches only part of the time, and then uses a few queries 2. Give AI plain pages on pricing, niche and proof, keep third-party profiles current and earn honest mentions. We do not guarantee any recommendation.
Do AI engines prefer agency websites or third-party lists?
Is it a good idea to publish a 'best agencies' list with ourselves at number one?
It can appear in AI answers, but it may cost trust. Ahrefs found a brand's own blog list among first-party pages recommended in 17.2% of agency prompts, and its author advises agencies to publish them, while a quoted agency owner argues self-ranking hurts buyer trust 1. A fair list with real criteria and a dated comparison is the safer route.
Can I use client results in my case studies?
Only with consent and a reasonable basis. The FTC expects advertisers to hold evidence for objective claims before publishing 9. Keep the data, the date range and written client permission, and state limits such as what else changed in the period. This is general information, not legal advice.
How do I know if AI is sending leads?
Ask. Add a free-text 'How did you hear about us?' field and a call question, and log your tracked prompts monthly. Google reports traffic from its AI features in the normal Performance report 6, but ChatGPT visits often arrive without a clear source.
Sources
- Do self-promotional 'best' lists help you show up in ChatGPT? · AhrefsOpened October 20, 2026. Supports: Dated 4 December 2025: 750 top-of-funnel prompts across software, products and agencies; own blog list appeared in 17.2% of agency answers recommending the publisher, landing page 30.4%; 'best X' blog lists 43.8% of page types; 79.1% of 1,100 lists updated in 2025, 57.1% updated since publication; author concludes agencies should publish such lists and sees no ranking or citation downside, and quotes agency owner Wil Reynolds saying self-ranking does not build buyer trust. Ahrefs sells related tools.
- How ChatGPT's query fan-out works (Ahrefs study of 66.6 million responses) · AhrefsOpened October 20, 2026. Supports: Dated 12 October 2026: 14.4% of 66.6 million free ChatGPT responses contained fan-out queries, about 1.6 fan-out queries each; September modifiers 2026 22.9%, official 13.1%, price 8.9%, reviews 3.6%; official sites took 63.4% of citations on factual and 46.4% on comparison questions.
- Gartner Survey Finds 69% of B2B Buyers Turn to Sales Reps to Validate AI-Generated Insights · Gartner (press release dated 20 May 2026, read on Gartner's newsroom mirror)Opened October 20, 2026. Supports: Survey of 645 B2B buyers, August to September 2025: 45% used GenAI in a recent purchase, mainly to research vendors; 69% prefer to validate with sales reps; average seven information sources.
- 2025 B2B Buyer Experience Report · 6senseOpened October 20, 2026. Supports: Nearly 4,000 buyers: winning vendor from Day One shortlist 95% of the time; 94% use LLMs in buying, mostly mid-journey to compare vendors; services cycles about two months longer than software. Vendor-run, self-reported, not agency-specific; 6sense sells buyer-intent software.
- Where do AI citations come from? · ProfoundOpened October 20, 2026. Supports: Dated 30 July 2026: 11.84 billion citations, 16 April to 16 July 2026, 29 industries, 8 models; brand sites about 57% overall and 72% in marketing. Profound sells AI visibility tools.
- AI features and your website · Google Search CentralOpened October 20, 2026. Supports: Pages must be indexed and snippet-eligible for AI Overviews and AI Mode; no additional technical requirements, special markup or AI text files; AI feature traffic is reported in the Performance report under Web search type.
- Optimizing your website for generative AI features on Google Search · Google Search CentralOpened October 20, 2026. Supports: Last updated 10 July 2026: create original content from your own knowledge and experience; structured data not required; pages made per fan-out variation to manipulate results violate the scaled content abuse policy.
- Consumer Reviews and Testimonials Rule: Questions and Answers · US Federal Trade CommissionOpened October 20, 2026. Supports: Effective 21 October 2024: bans fake reviews, sentiment-conditioned incentives, undisclosed insider reviews, claims that a site you control gives independent reviews, review suppression and fake social media indicators.
- FTC Policy Statement Regarding Advertising Substantiation · US Federal Trade CommissionOpened October 20, 2026. Supports: Dated 23 November 1984: advertisers and ad agencies must have a reasonable basis for objective claims before disseminating them.
- Review snippet (Review, AggregateRating) structured data · Google Search CentralOpened October 20, 2026. Supports: Pages using LocalBusiness or Organization markup are ineligible for star ratings when the reviewed entity controls the reviews about itself, including via a third-party widget.
- Creating helpful, reliable, people-first content · Google Search CentralOpened October 20, 2026. Supports: Trust is the most important part of E-E-A-T; Who, How and Why framework; recommends accurate bylines and warns against fabricated profiles, fake headshots and false credentials.
- Chalk Labs case studies · Chalk LabsOpened October 20, 2026. Supports: Chalk Labs' own account: Meru Trophy 27,300 clicks and 2.3M impressions in 16 months; shilikajain.com ranking first for 'Fractional PR Manager Crypto' within a month; Google search figures, no AI-citation counts.