Why your coverage matters more than it did
When someone asks ChatGPT, Gemini or Perplexity which brand to choose, the answer is assembled mostly from sources the brand neither owns nor pays for, and journalism alone is about a fifth of all citations. That part is your work. You keep the relationship, the brief and the credit. We are the production and technology arm behind it, as we are for branding studios, event agencies and consultants and communities. The full offer is on the partners page.
| Number | What it says | Source |
|---|---|---|
| More than 84% | of links cited in AI answers come from earned and third-party sources: journalism, government and NGO sites, research, encyclopedias, other companies' pages. Stable across four editions since July 2025 | Muck Rack, October 2026 |
| 22% | of citations are journalism, down from 27% in May | Muck Rack, October 2026 |
| 0.3% | of citations are paid or advertorial content, unchanged since May | Muck Rack, October 2026 |
| 3.5x | how much more often press releases are cited in industry-trend answers than in "best-of" answers | Muck Rack, May 2026 |
| 96% / 82% | share of ChatGPT and Gemini answers that cite sources. If you are not in the sources, you are not in the answer | Muck Rack, May 2026 |
| 4% to 28% | regular ChatGPT use in Spain, 2023 to 2025 | Funcas, January 2026 |
In plain words: the thing you already sell became more valuable, and nobody is measuring where it goes. The clipping report tells the client where the story ran. The share-of-voice report tells them what the assistants did with it.
The 3.5x figure matters for planning. A release does one job in industry-trend answers ("what is happening in this sector") and a different job in best-of answers ("which one should I buy"). The first needs news; the second needs comparisons, reviews and a page that says what the product is and costs.
What you sell, in three lines
- AI Visibility Audit, 5 days. Thirty category questions across ChatGPT, Gemini, Perplexity and Google AI, share of voice against five competitors, the sources each model cited, a technical check and a ten-step plan. The natural first sale, see below.
- AI Visibility Program, monthly. Citable content, entity structure, your coverage aligned with the questions people actually ask, and a share-of-voice report the client presents to the board. Details on AI visibility.
- Launch Page 72h, one per release. Message, copy in Spanish and English, design, build, analytics and a baseline of AI visibility in the page itself. Two rounds of revisions included. Details on Launch Page 72h. When a campaign needs a mechanic (a quiz, a calculator, a promo game), it becomes a Campaign Microsite in 1 to 2 weeks.
A page for every release
The rhythm follows your news calendar, not ours.
| When | What | Who |
|---|---|---|
| Day 0, 30 min | Intro: you tell us the brand and the next news moment, we tell you the model and the partner price | You and us |
| Day 1, 45 min | Brief with the brand, one call, one checklist. Materials land, the clock starts | You lead, we ask the technical questions |
| Day 4 | Page live on the client's domain, collecting leads. The release goes out with somewhere to land | Us |
| Week 1, then monthly | Audit, then the share-of-voice chart | Us, presented by you |
Partner prices are fixed per package and quoted within 24 working hours. EU invoicing, prices ex-VAT.
The co-branded audit as the first sale
The audit is the easiest thing to sell because it does not ask the client to believe anything. Five days later they see how their brand appears in four AI models against five competitors, which sources the models trust, and which of your placements are already doing the work. From that report, the Program is a logical next line on the same retainer, and the release calendar becomes a page calendar.
It is also why we push for the co-branded model with PR agencies: the product is literally a joint one. Your publications plus our structure and measurement. One proposal, two suppliers, one team in front of the client.
What the monthly report looks like
One document, your logo on the cover in co-branded and white-label, ours in referral. Inside:
- Share of voice against five named competitors, per model: ChatGPT, Gemini, Perplexity, Google AI.
- The sources each model cited that month, and which of them are your placements.
- Movement since the audit baseline, so the before and after is a chart, not a claim.
- What we changed on the brand's pages and what we recommend for the next news cycle.
You present it as your result. Because it is.
Models and margins
| Model | How the brand sees it | Money | When it fits a PR agency |
|---|---|---|---|
| Referral | You introduce us. The contract is between lobod.rocks and the brand; you stay in the loop as much as you want | 15% of the first contract, 10% of renewals | Big brands with procurement, or when you would rather not carry production risk |
| Co-branded (our pick) | "Communications: you. Digital and AI: lobod.rocks." One proposal, two suppliers, one team | Each invoices its part; the brand gets 10% off the joint package | Campaigns where PR and digital run together. The strongest and the most honest format |
| White-label | You sell it as your own service. We are the invisible subcontractor working through your contact | Your 20–30% margin on our partner price | When you are growing into a full-service shop. Your name stays in front; we ask only to publish the results anonymously |
Mix them per client. No exclusivity, and never two agencies for the same brand: the first one that brings it keeps it.
Who talks to the brand, who owns the code
Who talks to the brand: in referral and co-branded, both of us, with you leading. In white-label, only you; we work through your contact and join calls as your team if you introduce us that way.
Who owns the code, the content and the data: the brand, always. Repositories, accounts and access are handed over at delivery, GDPR applies by default, and no personal data is used to train models.
Objections you will hear, and the answers
"We already report coverage." You report where the story ran. This reports what ChatGPT and Gemini did with it, against five competitors, every month.
"Can you guarantee we appear in ChatGPT?" A guaranteed spot in an AI answer is something nobody controls, so nobody can sell it. What we sell is the measurement and the movement: the rate at which the brand appears across four models, reported every month against five competitors and worked up from the audit baseline. Your client gets a number with a direction, not a promise.
"Our client has no budget for digital." The audit costs the client nothing the first time, and a Launch Page 72h is a fixed fee.
"We do not want another supplier in the room." Then white-label: your contact, your logo, our build. Or referral: we take the contract, you take the fee and stay in copy.
First step
A message is enough. A reply the same working day, in English or Spanish, from one person: Dmytro Lobod.