The AI platform Marketing Services use to get recommended on
Your next client is asking a model which agency to shortlist. LBOX measures what the engines say about you, shows which agencies they name instead, and gives you the same report to run for the clients who are about to ask you this question.
Measured across
Perplexity
Gemini
Google AI
Claude
Copilot
Meta
DeepSeekGrok... and moreEvery one of them is competing for the same answer. One name comes back when a buyer asks.
What makes Marketing Services different
Not general marketing advice. These are the constraints your category actually operates under.
The shortlist is built before the brief ever reaches you.
Best agency for B2B SaaS in Chicago goes to a model now, and it returns three or four names. If you are not among them you never learn the brief existed. There is no lost pitch to review, because there was no pitch.
You report on channels that no longer cover the first answer.
Rankings, sessions, impressions, cost per lead. None of them describe what a model said when someone asked which agency to hire, and that conversation leaves no entry in anyone analytics.
Your best work is locked behind a form.
The case study that would win the pitch is a gated PDF. Gating it is how you capture the lead, and it is also why the engines cannot read it, so the proof you are proudest of is the proof they never cite.
Every client will ask you about this, and the answer has to be yours.
Why are we not showing up in ChatGPT is already arriving. You either bring the measurement to the quarterly review yourself or you sit in a meeting where someone else brought it.
What you get back
The console, measuring a company in Marketing Services against the questions its buyers ask. Every brand named, every domain cited, every engine separately.
A worked example on an illustrative company, shown to display the console. Your own numbers come from measuring your questions through the engines’ own APIs.
See which agencies the engines name
Visibility and position on the unbranded who-should-we-hire questions your prospects ask, measured per engine.
- Ranked against the agencies the engines actually named
- See where a model reads your pages without naming you
You against the field, engine by engine
unbranded questions| Brand | Gemini | AI Overviews | OpenAI | Perplexity | Claude | ChatGPT |
|---|---|---|---|---|---|---|
| Thorne & Meadow | 30% | 27% | 25% | 21% | 13% | 29% |
| Kingfisher Collective | 24% | 22% | 20% | 17% | 10% | 23% |
| Brightwell Marketingyou | 19% | 10% | 16% | 14% | 7% | 3% |
| Rivet & Co | 11% | 8% | 8% | 10% | 4% | 9% |
Turn what the answers get wrong into work that ships
Findings arrive as a worklist sorted by who can act: what LBOX writes, what needs your decision, what needs your own studio.
- Scored against your services and sectors, not a generic SEO checklist
- Nothing publishes without your sign off
The worklist
Prove the change, quarter after quarter
Every measurement is stored, so the line lengthens each cycle and you can put movement in front of a client rather than a claim.
- The same questions, asked again and checked
- Sources broken out per model, per page
Visibility over time
every measured runKnow when the models come to read your work
Crawler activity captured at the edge, so you can see which models fetch your case studies, service and sector pages.
- Every model, every page, logged over time
- Robots and content signals kept current
Crawler activity
hits by model, last 30 daysWhere you disappear as the buyer gets closer to hiring
Presence falls away as intent rises. The who-should-we-hire questions are the ones that become a brief, and the ones directories win.
- Visibility split by where the question sits in the journey
- Names the stage that costs you the pitch
Visibility by buying stage
where you show up in the journeyThe sources the engines trust when someone asks who to hire
Review directories and community threads decide which agencies get named. Your own site is a small share of what the engines read about you.
- Owned versus earned, so you know what to build and what to earn
- Every citation traced to the answer it supported
Which domains the engines cite
629 citationsWhat this is worth to your agency
Most verticals buy this to be found. You can also sell it, pitch with it, and use it to show existing clients a result they are not getting anywhere else.
Put your name on it.
The dashboard, the reports and the client proposal all run under your brand and your colours. Your client sees your agency delivering the measurement, and the platform stays behind you.
A line item you do not have yet.
Sell it as a retainer add-on, a standalone audit, or the first phase of a bigger engagement. Agency pricing is built for running many client accounts rather than one.
Walk into the pitch already holding their number.
Run the test on a prospect before the meeting and open with what the engines say about them today, next to the competitors they care about. It is a specific finding about their business, so it is a very hard meeting to cancel.
Give existing clients something new at the next review.
Bring a measurement nobody else is showing them and a worklist that makes the following quarter obvious. Renewal conversations get easier when you arrive with a number that moved.
Make the content and creative you already sell defensible.
Your work is now read by models as well as people. This shows which pages the engines actually quote and which they ignore, so the content programme has a measurable outcome attached to it instead of an impression count.
The program for Marketing Services
Everything below is included. Measured through the engines’ own APIs, so the same questions can be asked again and the result checked.
- Measure what every major AI engine says about your agency today, through the engines own APIs
- The hiring questions your prospects actually ask, mapped by service, sector and city
- Which agencies and directories the engines name instead of you
- Whether the engines describe your services, sectors and results correctly
- Which sources the engines cite when someone asks who to hire, and which of those you control
- Case study, service and sector pages built so the engines can read and quote your proof
- The same report to run for your own clients, under your brand
- Re-measurement on the same questions, so the change is provable rather than asserted
What this does not show: what any individual user saw. LBOX measures what the engines answer through their own APIs, which is repeatable and checkable. It is not a record of any one person’s session, and it does not predict revenue.