Marketing in several markets without multiplying the team

Translating a campaign for another country is the fastest way to prove you don’t know that country.

From Spain we fall into this often. Because we speak Spanish, we assume the same message works in Mexico, Colombia or Argentina. And when someone realises it doesn’t, the usual fix is to translate it into “Mexican Spanish” and use it for the whole of Latin America. Both ideas are completely wrong.

You can grow in several markets without one person per country. What you can’t do is grow without someone who knows each market.

Every market has its own words, and its own topics

Every country has its own vocabulary, expressions and nuances, and they aren’t always the ones you expect. From Spain we assume that in Mexico you address everyone formally, yet on LinkedIn people often talk to each other as equals; email, on the other hand, is more formal. When localising for France we thought the equivalent of “Dear” plus the first name was enough, and there the surname is used more than the first name.

These are details AI doesn’t know from the start. Often neither do we: you learn them by working with the market, and they’re the ones that make the difference.

Localising goes beyond language. You need to know what kind of content each market consumes, which formats work, which cultural references it understands and which topics, political ones included, are worth touching or avoiding depending on what each country is going through. A video that works in one market can clash in the next.

That’s why a campaign gets adapted as a whole: the message, the examples, the offer and, sometimes, the product itself.

AI translates well and localises with clichés

I’ve seen many companies enter a new market leaving everything to AI, because it translates well and that seems enough to them. It isn’t. It lacks exactly those details, and when they’re missing you can tell there’s a machine behind it rather than someone who knows the market.

The reason is that AI still leans on clichés. If you give it nothing else, it writes the average version of a country, the one that comes from what it has read online, and nobody knows whether that’s true.

In Europe there’s also a regulatory reason. The AI Act already requires generative systems to mark their content so it can be detected (systems on the market before August 2026 have until December to do so). Outside Europe the rules are, for now, more permissive, although each country moves at its own pace. My prediction is that in Europe it will get easier and easier to tell which piece came out of an AI without passing through a person, and that costs credibility in a second.

What works is the reverse order. First, a professional who knows each market’s mindset adjusts the content and teaches the AI how it should sound. With that knowledge gathered in a base, with examples of what works and what doesn’t, an agent can then run that market. Always under supervision.

There are two ways to get there: already having that knowledge base, or having someone who builds it. Feeding the agent only with what it finds online isn’t worth it.

Better a mixed team than a big team

I prefer a mixed team, with people who think from different markets. I’ve lived it from a central office working for several countries: we didn’t adapt the content, the product, the offer or the language enough. When we started working with people from those countries, localising and preparing each launch went much better.

This doesn’t take people out of the process. It moves them: instead of someone per country translating, someone who knows the market teaches the agent and reviews what it produces. And that work has to happen before the agent saves anything.

If I had to launch a new market with a small team and agents, I’d follow this order:

1. Research the market. Analytics, mindset and the details that don’t show up in reports.
2. Talk to the people who know it. Specialists in that market and, if there is one, the local team.
3. Prepare the agent. Load all that knowledge into its base so the content comes out accurate from the start.
4. Test before launching. A/B tests on the key pieces and only then, the launch.
5. Measure and improve. Analyse the results and keep adjusting the agent and the content, with someone who knows the market reviewing.

If your campaign sounds the same in Madrid, Mexico City and Buenos Aires, you haven’t entered three markets. You haven’t entered any.