We read 524 reviews of 36 GEO agencies. Only 61 describe GEO work.
Clutch's Generative Engine Optimization directory lists 36 agencies. We read every review on all 36 profiles and tagged each one. 11.6% describe AI search work, seven of the nine agencies claiming 100% GEO focus have none at all, and two reviews in the whole set put a number on a commercial result.
On this page
Note on verification: we read all 524 reviews by hand on the live Clutch site between 29 and 31 August 2026. Four things we could not verify or control. First, Clutch's "GEO focus %" is a number each agency types into its own profile. Clutch does not check it against reviews. Second, we sampled every review on 24 of the 36 profiles and the most recent 10 to 20 on the other 12, so long-tailed profiles are under-sampled at the older end. Third, a review counts as "GEO" only when the reviewer names AI search, ChatGPT, LLMs, or generative engines in the scope or the outcome. A project that quietly included GEO work reads as SEO here. Fourth, one agency's review set has an unusual posting pattern that we describe in full below. We counted it in headline totals and left it out of every buyer breakdown. The dataset behind every number is described at the foot of this post. Check our coding.
The short version: Clutch lists 36 agencies under Generative Engine Optimization. We read every review on all 36 profiles, 524 in total. Just 61 of them, 11.6%, describe any AI search work. Seven of the nine agencies claiming 100% GEO focus have zero GEO reviews. And across all 524, exactly two reviews put a number on a commercial result.
Every GEO agency says the same thing. We do AI search. We get you into ChatGPT. Trust us, we have done this before.
We wanted to know if anyone had actually done it before. Not claimed it. Done it, with a client who wrote about it afterward.
So we went and counted.
What we did
Clutch runs a directory called Generative Engine Optimization. On 29 August 2026 it listed 36 agencies. We opened all 36 profiles and read the reviews.
For each review we recorded 19 fields: the client, their industry, their country, their headcount, the reviewer's job title, what they bought, the project size, the duration, and the date. Then we tagged every review one of three ways.
- GEO-specific. The reviewer names AI search, ChatGPT, LLMs, or generative engines.
- Traditional SEO or adjacent. Search work with no AI mention.
- Non-search service. Web design, ads, branding, everything else.
That gave us 524 tagged reviews. Here is what they say.
Finding 1: 11.6% of reviews at GEO agencies describe GEO work
| Review type | Count | Share |
|---|---|---|
| Traditional SEO or adjacent | 275 | 52.5% |
| Non-search service | 188 | 35.9% |
| GEO-specific | 61 | 11.6% |
Nine reviews in ten at a "GEO agency" describe something other than GEO.
This is not a scandal. It is a repositioning. Most of these firms are SEO shops that added a line to their profile in the last 18 months. The work in the reviews is the work they were already doing.
But it does break a thing buyers do. You land on a Clutch profile, you see "100% Generative Engine Optimization", you see 38 reviews and a 4.9 rating, and your brain does the maths. Those reviews must be GEO reviews.
They are not.
Finding 2: seven of the nine agencies claiming 100% GEO have zero GEO reviews
Nine agencies in the directory list their GEO focus at 100%. We checked each one against its own review page.
| Agency | GEO reviews found | Reviews we read |
|---|---|---|
| Akravo | 14 | 14 |
| ShowUpWithAI | 6 | 6 |
| 97 Switch | 0 | 20 |
| CSP Agency | 0 | 11 |
| Digital Growth Core | 0 | 5 |
| Hop AI | 0 | 24 |
| MacRAE'S | 0 | 19 |
| RevenueZen by Eastern Standard | 0 | 20 |
| blimpp | 0 | 4 |
Hop AI has 24 reviews and a name with AI in it. None of the 24 mention AI search.
Across the whole directory, 19 of the 36 agencies have no GEO review at all.
We are not calling anyone a liar. The focus percentage is self-reported and Clutch never promised it matched the reviews. But if you are picking an agency off that number, you are picking off a number nobody checks.
Finding 3: the market has a birthday, and it is 2026
We dated every GEO review.
| Year | GEO reviews |
|---|---|
| Before 2024 | 0 |
| 2024 | 1 |
| 2025 | 7 |
| 2026 (to 31 Aug) | 53 |
Zero before 2024. One in 2024. Fifty-three in the first eight months of 2026.
More than half of every GEO review ever written landed this year. The category is roughly 20 months old in review data.
So when an agency tells you they have years of GEO experience, ask them where the clients went. Nobody in this directory has a track record, including us. That cuts both ways, and it is worth saying out loud.
Finding 4: IT services buys GEO more than any other industry
Here is the part that surprised us.
We took every industry with at least 15 reviews in the dataset and worked out what share of those reviews were GEO.
| Client industry | GEO reviews | Total reviews | GEO rate |
|---|---|---|---|
| IT services / MSP | 9 | 32 | 28.1% |
| Agency / marketing services | 7 | 33 | 21.2% |
| Consulting / professional services | 8 | 42 | 19.0% |
| Legal | 2 | 18 | 11.1% |
| B2B SaaS | 2 | 25 | 8.0% |
| Manufacturing / industrial | 3 | 43 | 7.0% |
| Fintech / financial services | 3 | 44 | 6.8% |
| Ecommerce / retail | 2 | 37 | 5.4% |
| Healthcare | 1 | 19 | 5.3% |
| Real estate / construction | 1 | 26 | 3.8% |
Managed IT firms buy GEO at more than five times the rate of ecommerce.
That is not where the industry is selling. Read the average GEO agency homepage and you will see ecommerce, SaaS, and enterprise. The reviews point somewhere else entirely: mid-size B2B service firms whose buyers already research vendors by asking a chatbot.
Two more cuts from the confirmed GEO set, minus the flagged agency (n=47).
- 63.8% are US companies.
- 76.6% have under 200 staff. The three biggest bands are 1 to 10, 11 to 50, and 51 to 200, at 12 reviews each.
Finding 5: not one CMO
We normalised the job title on every confirmed GEO review.
| Buyer role | Share |
|---|---|
| Founder, CEO, owner, president | 34.0% |
| Marketing manager or specialist | 14.9% |
| Directors (all functions) | 10.6% |
| Sales / business development | 6.4% |
| VP | 4.3% |
| CMO | 0% |
Across 47 confirmed GEO projects, not a single reviewer is a CMO.
The buyer is the founder at small firms and the marketing manager at slightly bigger ones. Nobody with "Chief" on their business card has written one of these reviews.
If you sell GEO and your outreach targets VPs and CMOs, you are writing to people who are not buying it.
Finding 6: this is a small-cheque product
Clutch asks for project size. Twenty-six of the confirmed GEO reviews disclosed one.
| Project size | Count | Share of disclosed |
|---|---|---|
| Under $10,000 | 6 | 23.1% |
| $10,000 to $49,999 | 16 | 61.5% |
| $50,000 to $199,999 | 4 | 15.4% |
| $200,000+ | 0 | 0% |
Nothing above $200,000. Nearly two thirds sit in one band, $10k to $49,999. And 31 of 47 projects are marked ongoing rather than completed, so most of these are retainers, not one-off audits.
Finding 7: nobody buys GEO on its own
Read the review titles and a pattern jumps out. They are nearly all bundles.
"SEO & GEO Services for..." "SEO and AI Search Optimization for..." "Digital Marketing & GEO for..."
Standalone AI-visibility products barely appear. GEO shows up as a line item bolted to an SEO retainer that already existed.
That matters if you are selling. A GEO-only pitch asks a buyer to create a new budget. A GEO attachment asks them to widen one they already approved.
Finding 8: buyers arrive already convinced
We coded the stated reason each buyer went looking.
Roughly six in ten confirmed GEO reviews describe the same moment. The buyer typed their own category into ChatGPT, saw a competitor, did not see themselves, and went hunting for help.
Almost nobody mentions being persuaded. Under 5% cite a competitor outranking them. Under 2% cite missing in-house skills. Not one review says an agency's content convinced them the category was real.
The lesson for anyone selling this: you cannot argue somebody into caring. You can show them a prompt result.
Finding 9: two reviews out of 524 contain a number
This is the finding we did not expect and cannot stop thinking about.
We looked for any quantified commercial outcome in a GEO review. Leads, revenue, traffic, visibility, anything with a figure attached.
We found two. Both at the same agency, ShowUpWithAI in Austin.
"Within the first two months, we've seen our overall search visibility improve by almost 20%"
a "20%–30% increase in monthly lead flow"
That is it. Two numbers across 36 agencies, 524 reviews, and every GEO project any of them has ever been reviewed for.
Everything else is warm and vague. Responsive. Knowledgeable. Great communication. Real praise from real clients, and worth nothing to a buyer trying to work out if the thing works.
An entire category is selling a service and nobody has published a defensible result. That is either the biggest problem in GEO or the biggest opening in it, depending on where you sit.
The one agency we set aside
We owe you this one in full, because we counted it in the totals and then removed it from every breakdown.
Akravo, a Netherlands firm, has 14 reviews. All 14 are tagged GEO. All 14 were posted inside a 13-day window between 28 March and 9 April 2026. Three of the 14 carry Clutch's Verified badge. Several client names read like placeholders and the industry tags sit oddly against the review text.
We are not going to tell you what that means. We do not know. We will tell you what we did about it: we left Akravo out of every buyer-characteristic table above, because 14 reviews from one 13-day window would have moved the role, size, and industry numbers on their own.
If you are evaluating agencies, the general point is more useful than the specific one. Look at posting dates. A review set that arrives in a cluster tells you something a five-star average does not.
What to do with this
If you are buying GEO:
Ignore the focus percentage. Ask the agency to send you the review URLs where a client describes AI search work. Not the profile. The individual reviews. Nineteen of 36 agencies cannot produce one.
Then ask what number they moved and how they measured it. Two reviews in this entire dataset survived that question. The rest of the questions worth asking, and the answers that should worry you, are in our buyer's framework for choosing a GEO agency.
If you are selling GEO:
Your buyer is a founder or a marketing manager at a US B2B service firm under 200 people, most likely managed IT, and they already pay someone for SEO. They found the problem themselves before you called. They will spend $10k to $49k on a retainer, not a project.
And the first agency to publish a controlled, attributable result owns the proof gap in a category where 522 out of 524 reviews contain no number at all.
Data and sources
The full 524-row dataset and the 36-agency profile sheet are available on request, and the coding rules are described in the method section above. Every number in this post comes from those two files.
- Clutch Generative Engine Optimization directory — clutch.co/seo-firms/generative-engine-optimization (read 29 August 2026)
- 36 agency profiles at
clutch.co/profile/<agency>(read 29 to 31 August 2026) geo_agency_buyer_dataset.csv, 524 rows, 19 fieldsgeo_agencies.csv, 36 rows, directory metadata
If you find a coding error, tell us and we will publish the correction.
Frequently asked
How many GEO agencies are there on Clutch?
Clutch's Generative Engine Optimization directory listed 36 agencies on 29 August 2026. Nineteen of them have no review describing AI search work.
What is Clutch's GEO focus percentage?
It is a service-mix figure each agency enters on its own profile. Clutch does not verify it against the agency's reviews. Seven of the nine agencies listing 100% GEO focus have zero GEO reviews.
Who buys generative engine optimization?
On this evidence: B2B service firms under 200 staff, most often managed IT and IT services, bought by the founder at smaller firms and the marketing manager at larger ones. Not a single confirmed GEO review came from a CMO.
How much does a GEO engagement cost?
Of the 26 confirmed GEO projects that disclosed a budget, 84.6% came in under $50,000 total and 61.5% fell between $10,000 and $49,999. None exceeded $200,000. Two thirds are ongoing retainers.
Does GEO get sold separately from SEO?
Rarely. Almost every GEO review in this dataset is titled as a bundle with SEO or wider digital marketing.
Is there proof that GEO works?
In public client reviews, almost none. Two of 524 reviews quantify a commercial outcome, both at one agency. That is not evidence GEO fails. It is evidence the category has not published its results yet.
Written by
Lourdes Paul Agilan runs GeoCited, a generative engine optimization agency. He works with B2B companies on how AI assistants describe and recommend them. He writes about that work here, and shares the experiments behind it. The tests that worked, and the tests that did not, written up the same way.