What a Cape Verde Footballer's LinkedIn Message Teaches Founders About Being Findable

Summary: Roberto Lopes deleted a LinkedIn message in 2018 because he assumed it was spam. Seven years later, that message put him at the World Cup, playing for a country he'd never represented before it arrived.

It's a story about being found by someone looking for something very specific. It also points at a live problem for founders: when an AI search tool goes looking for evidence of what your business actually does, most LinkedIn profiles have nothing on them for it to find.

The message he almost deleted

In 2018, a message landed in Roberto Lopes' LinkedIn inbox. It was written in Portuguese, from someone he didn't recognise. He assumed it was spam and ignored it.

Nine months later, the same person tried again, this time in English. Lopes ran it through Google Translate. It was from Rui Águas, Cape Verde's national team coach, asking whether a Dublin bank worker with an Irish mother and a Cape Verdean father wanted to declare for the national side.

"I thought it was spam," he later told BBC Sport.

Three weeks after finally replying, Lopes was on a plane to make his international debut against Togo. Seven years on, he's a first-team defender helping Cape Verde through its first ever World Cup, one of the smallest nations ever to qualify for the tournament.

It's a good story on its own terms. Strip away the football, though, and it's really a story about findability: about someone going looking for a very specific thing, finding evidence of it in an unlikely place, and acting on it.

That's worth sitting with, because it's starting to happen to your business too, except the one doing the looking isn't a national football coach. It's ChatGPT, Perplexity, or Google's AI Overview, being asked by a prospective client who you are and whether you're worth calling.

Why relevance beat volume

Cape Verde didn't send the same message to every footballer with a vaguely Portuguese surname. Águas was working through eligibility rules and family backgrounds looking for a small number of specific, qualified players. Lopes fit the criteria exactly. The method was precision, not volume.

Most LinkedIn outreach works the opposite way round. Founders, or the agencies they've hired, send an identical templated pitch to a thousand people and hope for a two percent reply rate. It rarely produces anything worth having.

The Cape Verde approach is a useful reminder that specificity beats scale, in outreach and in the content you publish. A message, or an article, written for one exact kind of reader consistently outperforms one written to please everybody.

The bigger problem: most founders have nothing on LinkedIn for AI to find

Here's the part that actually matters for your business right now. When someone asks an AI search tool a question like "who's a good [service] company for a business our size", the model isn't guessing. It's assembling evidence: things it has read that it can attribute to a real source and that back up a specific claim.

A 2024 study by researchers from Princeton, Georgia Tech, the Allen Institute for AI and IIT Delhi tested nine content techniques against 10,000 real queries to see what actually changed whether AI tools cited a piece of content. Adding specific statistics improved citation likelihood by up to 41 percent, the strongest single lever they tested. Adding real, attributable quotes and citing credible sources produced similar gains, in the 30 to 40 percent range (Aggarwal, Murahari, Rajpurohit, Kalyan, Narasimhan and Deshpande, KDD 2024). Keyword stuffing, the old SEO standby, didn't help at all.

Most SME founders' LinkedIn presence gives an AI tool nothing to work with. A headline, a job title, maybe a handful of posts from three years ago. No evidence of expertise, no specific claims, nobody vouching for you. When the AI goes looking, there's nothing there worth citing, so it recommends someone else, usually a competitor who happens to post occasionally and has a few genuine recommendations sitting on their profile.

This isn't really about posting more often. It's about posting things that count as evidence: specific results, named clients where you have permission, direct quotes from people who've actually worked with you, clear statements of what you do and for whom.

That's the same list the GEO researchers found worked for AI citation generally. It applies just as much to a LinkedIn profile as to a web page.

What this actually means for your business

If a prospect asked an AI tool "who should I speak to about [your service] in [your area]", would there be anything for it to find? For most of the founders we work with, the honest answer is no. Not because they lack the expertise, but because none of it is written down anywhere a model can read it.

The fix isn't complicated, but it does take discipline. Specific case studies with real numbers, published on your site and echoed on LinkedIn. Client quotes you're actually allowed to use, attributed properly rather than left vague. Posts under your own name with a clear point of view, not a generic company update.

None of this requires the volume tactics that used to work for SEO. It requires being specific about a small number of things you can prove.

Frequently asked questions

Does posting more often on LinkedIn help with AI search visibility?

Not by itself. Volume without specificity doesn't move the needle much. The GEO research found that what mattered was the presence of real statistics, quotes and sourced claims, not the amount of content published.

Is this the same as normal SEO?

No. Traditional SEO rewards keyword density and backlinks. The GEO study found that keyword stuffing didn't help with AI citation, and in some cases hurt it. AI tools are looking for evidence and clarity, not repetition.

Can a small business realistically compete with bigger companies for AI visibility?

Yes, arguably more easily than with traditional Google rankings. AI tools assemble the most specific, well-evidenced answer to a narrow question, not just pull from the biggest domain. A small firm with three detailed, honest case studies can outcite a large one running on generic marketing copy.

How long does it take to see a difference?

There's no fixed timeline, and anyone promising one is guessing. It depends on how quickly the content gets indexed and how often the underlying AI tools refresh their sources. Treat it as a medium-term investment, not a quick win.

Does a well-written LinkedIn profile guarantee AI visibility on its own?

No. Technical indexing, genuine backlinks and real third-party mentions all matter too. A strong profile and strong content are necessary, but not sufficient, on their own.