Professional services
Marketing and advertising
Creating demand and attention for a brand.
Marketing firms are the one part of this category that sells marketing for a living. Their own clients now put the brief to an engine before they put it to an agency, and most of the proof an agency would answer with sits under an NDA. So we treat this as a peer set: white-label capacity, entity work at relaunch, and partnership where earned coverage meets citation.
Where the answer is being lost
Agencies get their clients cited and stay uncited themselves.
A growth lead at a D2C brand types "How do you lower blended CAC when paid channels plateau" and reads a considered answer with a named source. It did not come from the performance agency that has solved that problem many times over, because that agency's evidence is a client dashboard it cannot show anyone. Down the corridor, a studio hands over a rebrand on a Friday and every engine spends the next quarter describing the company by its old name, a product line that was discontinued and the positioning the studio was hired to replace. One of those is a brief that never arrived. The other is a client question on Monday morning that nobody in the room was hired to answer.
How we win this
The programme for marketing and advertising
Peers first, prospects second
A peer spots a pitch at forty paces, so there is no point dressing this up. We are not going to teach a performance team how to buy media or a studio how to name a company. What we sell is the surface they never staff, the technical work at relaunch, and delivery under their name where that is what suits.
Method is the publishable proof
The account name, the spend and the result all belong to someone else. What belongs to the agency is how it decided: the test it ran before the budget moved, the threshold at which it changes creative, the aggregate across a whole book of accounts. Published with the working shown, that is material a rival cannot copy and a model can quote.
Rebrands break the record
Rename a company and the engines keep the old one. Old descriptions, retired products, an executive who left two years ago, all still attached to an organisation that no longer exists under that name. Somebody has to reconcile the two records and check what the redirect map dropped. It is unglamorous, it is billable, and no studio has that person on staff.
Measurement you can resell
A PR firm defending a retainer and an agency justifying a media plan need the same thing: evidence that ties their work to what a model now says about the client. We run the prompt set, track which sources get pulled, and hand over the output as data for their own reporting. Our name does not need to be on it.
The mix that carries it
Content
Answer and comparison pages
Cost, process, eligibility and comparison pages built for direct extraction, not for a reader who scrolls.
Foundation
Entity and schema engineering
Structured data and entity definition so engines know exactly what you are, where you operate, and what you are credible in.
Measurement
AI Presence tracking
Standing measurement of inclusion, share of answer and competitor movement as models update.
Measurement
AI Visibility Diagnostic
We query the live models with your buyers' real questions and document exactly who gets named today.
Content
GEO blogs and authority content
The definitive written answer to the questions your buyers put to an engine, structured so it can be lifted and attributed.
Foundation
Technical fixes
Crawlability, render, speed and the machine-readability faults that keep an engine from reading you at all.
Foundation
Website build and optimisation
Sites architected for SEO and GEO from the ground up: fast, structured, hardened, and legible to a model.
Authority
Original data and benchmarks
Proprietary numbers, surveys and benchmarks — the most-cited asset class there is, because nobody else has them.
Distribution
Practitioner and executive content where B2B buyers and the models watching them both look.
Authority
Digital public outreach
Earned mentions, trade coverage and third-party citations — the corroboration a model checks before it names you.
The constraint we work inside
Almost everything that would prove an agency's case belongs to a client: the account name, the spend, the numbers on the dashboard. We publish nothing that names an account or restates a result we cannot source. What is left is method, aggregated benchmarks and the reasoning behind a decision, which is what gets cited anyway.
Specialisations
4 total
The pitch is different for each one, because the buyer, the trigger and the rules on what may be published are different for each one. Open the one that is yours.
Your account manager takes a call on Monday morning: the client has seen a rival named in an AI answer, wants to know why they are not, and the renewal is already in the diary.
The question deciding this today
“Which agencies do generative engine optimisation”
- Who they sell to
- Companies losing organic visibility as AI answers absorb search traffic
- Who signs
- CMO, head of demand generation, founder
- What starts it
- Traffic decline, AI Overviews rollout, a competitor appearing in AI answers
- Cost of staying invisible
- Pipeline erodes quietly while the brand vanishes from the answers buyers see
You do this for a living, so we are not going to sell you visibility. When a CMO types "Which agencies do generative engine optimisation" you and we come back inside the same answer set, and neither of us is going to publish the other out of it. The gap that is real here is capacity. Your clients are now asking a GEO question inside an SEO retainer, and answering it properly takes entity work, an answer layer and a monitoring habit that nobody on your team has a free fortnight to build. So the offer is delivery, on your accounts, under your name. Where the work is bigger than that or sits outside what we do, we would rather refer it than compete for it.
What we would run
- 01AI Visibility Diagnostic
We run the diagnostic on your client's category rather than yours: their buyers' real questions put to live models, the named set that comes back, the reasoning attached to each name. Delivered in your template with our name nowhere on it.
It gives your account manager something to open Monday's call with, and you can sell it before you have hired anyone to deliver the rest of the retainer.
- 02Entity and schema engineering
Entity and service definitions built out on client domains: what the client sells, which entity owns which product, how the same organisation is named across its own pages and its directory records. Built to your spec and handed back as your team's work.
This is the half of a GEO retainer that needs an engineer and a fortnight nobody has. It is also the half that decides whether the client is retrievable at all.
- 03Answer and comparison pages
The answer and comparison layer on client sites, written against the prompt set the diagnostic returned: cost, process, eligibility, one option against another. Drafted, sourced and shipped to your editorial standard.
You already know this is the layer that gets quoted. What you do not have is writers who can hold that pace across a book of accounts without you editing every draft.
- 04AI Presence tracking
A standing record per account: which sources the models pull for that category, who is named this month who was not last month, and what shifts once a new page has been indexed a while. Exported as data for your own reporting.
Renewals here turn on showing movement. This is the number you put in the review, and it goes in with your logo on it.
What we would not recommend
- GEO blogs and authority content. You are not going to hand your own blog to a rival firm, and we would not take the brief if you did. Anything we write goes out on a client domain.
- Reddit. The marketing subreddits remove vendor self-promotion on sight, and agency-recommendation threads there are settled by people who will never sign a retainer.
- Quora. Agency-selection answers on Quora are written by agencies. Adding another one puts you in a pile the engines already discount.
- Instagram. None of this work photographs. A schema block and a render fix leave nothing to post, and the agency accounts that do post there are recruiting rather than selling.
What a lead looks like
The enquiry runs the other way in this one. A search agency founder writes with three screenshots attached, having spent the weekend putting his clients' category prompts into the models himself. He wants to know what a white-label retainer costs per account, whether the diagnostic can go out in his template, and how firmly our name stays off the deck in front of his client. Some of these end as a referral arrangement rather than a retainer, and often enough that is the right answer.
What we measure
- Delivery shipped to your template, unattributed
- Client entity and answer work completed inside your retainer
- Movement reported per account, sourced and dated
- Work referred out where a retainer is the wrong fit
What changes
Everything that arrives is from a practitioner, so it opens at scope rather than at credentials. A performance agency founder who needs a benchmark built and a writer he does not want on payroll. An account director at a PR firm who has been asked for a citation line in a client's quarterly review and has no way to produce one. A studio principal calling in launch week because the renamed company still resolves to the old name. Fewer enquiries than a consumer category would produce, and not one of them price shopping.
Start here
See who gets named in marketing and advertising today
We put your buyers' real questions to the live models and come back with the businesses they name, the sources behind those answers, and the gap between that list and yours.