At the beginning of this year I received an AI-generated content project with a deceptively simple goal: adapting email messages for international public.
This was not my first time when I navigated global nuance. With an MBA in international business and experience in working on a global advisory project in Portugal, I have already seen how messages end up depending on culture, sound and language. But I was the first time that I was looking for this lens on the generation of AI content in my Martech -Ai -Ai -Ai -Ai Roll at Drift Kings Media -and it was more complex than expected.
We already had an email entry prompt with AI-generated, which works well in English-converting, friendly and context-related. The challenge? It works in Spanish and French, without robots, awkward or culturally outside the base.
Sounds just. It wasn’t.
The hidden complexity of “just localize”
What we really did was a AI As natural as a local marketer.
Our first attempts fell flat.
Example (original AI output in Spanish):
Here is what we aim for in English:
“I saw that they got around on the platform and that they were interested in talking to us. Would you like to meet one of the following days?”
This is the original edition in Spanish:
“Esttuve revisando tuS interacciones en nuestra plataforma y quería of recmeme como tu punto de contacto.”
In English it translates:
“I checked their activities and wanted to become their contact point.”
While is grammatically correct, this sounded invasive In Spanish – as if we had watched the user too closely. It didn’t feel natural. A reviewer called it “creepy”.
Here is another example:
- Original English intention: “I found that you explored our platform and expressed interest in the connection with us.”
- Original Spanish edition: “I Pareció interesting tu interés en nuestros Servicios.”
- Translation in English: “I found your interest in our services interesting.”
It is also technically precise here, but it is superfluous and robot. It is the type of phrasing that brings a reader to a standstill and: “Has it written a bot?”
Taking it: Even if the translation is exactly that clay can be switched off. And sound is everything in marketing.
The shift from the translation to language consciousness prompt design
At that time it became clear to me a system So that the AI leads to thinking like a multilingual marketer.
I built A Language -portable prompt framework – A structured input request that can adapt to languages and at the same time respect the unique grammar, the sound and the cultural context of a unique one.
The following has changed
Instead of a static prompt, I broke the logic into variables:
- : Target (e.g. Spanish, French, German)
- : Pronoun and sound level (“tu” against “usted”, “vous” against “tu”)
- : Inbox -friendly, conversal, professional
- : Directly against suggestive phrasing
- : Enforced where the grammar allowed
We also added clear, language -specific rules.
Example (Spanish):
- Use Tú Consistently never Tusted (too formal for our brand)
- Avoid gender -specific adjectives like Interestado/interestada if possible
✅ “Mostreste Interés en …”
❌ “Estuviste interresado en …”
Example (French):
- Always use Vousnot tuIn B2B messages
- Avoid ambiguous endings like Intéressé (e)
✅ “Vous Avez Montré de l’intérêt …”
❌ “Tu t’éta intéressé (e) …”
Why this shift was important
A friendly CTA could be in English like:
“Would you be available for a short conversation on one of the following days?”
We tried to translate it directly in Spanish:
“Quieres Agendar 15 Minutos Para Hablar Sobre Lo Que Estás Buscando?”
It was grammatically correct, but it sounded Too casual And unprofessional in a B2B context. Not intrusive, only slightly abolished.
So we have reformulated it in a friendly but formally formulated manner:
“Si Te Parece Bien, Podemos -agendar Una Conversación Breve Esta Semana.”
This means ::
“If it works for you, we can plan a short chat this week.”
Here is another example in French:
- Original edition: “Souhaitez-vous Prendre rendez-vous Guspose?”
((“Would you like to plan a meeting to discuss this?”))) - New version: “Auriez-vous 20 minutes pure voir comment hubspot pourrait concrètement vous aider?” ((“Would you have 20 minutes to see how Drift Kings Media could practically support you?”)))
The second version gives the CTA added value. Not only time – but also purpose.
Secure it with a stakeholder questionnaire
Localization is not just a linguistic problem – it is a problem with the division of business.
To do it right, I created a simple document of the stakeholder recording and shared it with marketing ops, regional marketers and content lines. The aim was to align tone, content limits and regional sensitivities at an early stage.
These are some of the questions I asked:
- What formality is appropriate in your market?
- Should we avoid gender -specific terms?
- Can we refer to the company or user product use?
- How directly should we ask for measures?
- Are there any phrases, cultural references or phrases that we should avoid?
We have some pretty interesting findings.
For example, the stakeholders preferred in some regions not To refer to the recipient’s company type in the copy, although this was common in English (e.g. “I saw that you help startups with HR with HR”).
The localized alternative became more general:
“DENENDO QUE ESTán Buscando Formas de Mejorar SuS Procesos Internos.” ((“I understand that you want to improve the internal processes.”)))
The results of this survey have contributed to creating clarity between content, surgery and regional marketing teams – and drastically reducing our revision cycles.
The end product: people sounding on the scale
With the updated input and recording area, the new outputs were immediately better.
Previously:
- Original edition: “Hola (first name), soy María de Drift Kings Media. He visto que has navegado nuestra plataforma y parece que te interests nuestro producto.”
- English translation: “Hello (first name), I am María from Drift Kings Media. I saw that you have searched our platform and it seems as if you are interested in our product.”
After:
- Original edition: “Soy María de Drift Kings Media. VI que estuvieste Explorando la Plataforma y Querías Saber Más Sobre Cómo Podemos Apoyar Tu Negocio.”
- English translation: “I am María from Drift Kings Media. I saw that you wanted to explore the platform and learn more about how we can support your business.”
And stakeholders reacted positively:
- “It finally sounds like someone had written it from our team.”
- “Perfect sound – natural and locally.”
- “No gender -specific errors or strange formalities. We can actually use this.”
Even better, we did not have to write separate input requests for each campaign. The same core frame now supplies with AI-generated messages in several languages with consistent quality.
Snack for marketers
Regardless of whether you work on AI copies, global ads or multilingual content, here is what I have learned:
1. Do not simply translate – localize yourself for the intention.
In literal translations you will receive “technically correct” content. But only localization will bring it to the country.
2. Use commands such as creative letters.
Enter sound, formality, CTA style, gender neutrality and other language rules than variables. Do not leave the nuance to chance.
3 .. Create language -conscious templates.
Languages behave differently. Plan for things such as verb conjugations, pluralization and phrase rhythm in advance.
4. Get feedback early.
Use a stakeholder input document before the generation, not afterwards. They later avoid rework and misalignment.
5. Aim for a real human tone.
If your AI edition doesn’t feel like something You Would write to a customer, it will not convert. Read it loud. Would you meet?
AI localization is now a marketing competence.
This project taught me something that has stayed with me since then: The future of global marketing It’s not just about scaling content – it’s about scaling the context.
The companies that are successful with AI are not those that generate most of content. You will be the ones who create that most resonant Content because they know how they demand. And that begins to understand the languages that their customers speak – in more than one way.