Previous this 12 months, I used to be passed an AI-generated content material venture with a deceptively easy objective: adapt e mail messages for world audiences.
This wasn’t my first time navigating world nuance. With an MBA in Global Trade and revel in running on a world consulting venture in Portugal, I’d already observed how messages land another way relying on tradition, tone, and language. However this used to be my first time making use of that lens to AI content material era in my MarTech AI function at HubSpot — and it used to be extra complicated than anticipated.
We already had an AI-generated e mail steered that labored neatly in English—conversational, pleasant, and context-aware. The problem? Making it paintings in Spanish and French with out sounding robot, clumsy, or culturally off-base.
Sounds simple. It wasn’t.
The Hidden Complexity of “Simply Localizing”
What we have been in point of fact doing used to be asking an AI type — educated predominantly in English — to talk different languages as naturally as a local marketer would.
Our first makes an attempt fell flat.
Instance (authentic AI output in Spanish):
Right here’s what we aimed for in English:
“I noticed you have been scoping across the platform and that you simply have been inquisitive about talking with us. Do you want to satisfy on one of the vital following days?”
That is the unique output in Spanish:
“Estuve revisando tus interacciones en nuestra plataforma y quería ofrecerme como tu punto de contacto.”
In English, it interprets to:
“I reviewed your task and sought after to develop into your level of touch.”
Whilst grammatically right kind, this sounded invasive in Spanish — like we have been gazing the consumer too carefully. It didn’t really feel herbal. One reviewer referred to as it “creepy.”
Right here’s every other instance:
- Authentic English intent: “I spotted you’ve been exploring our platform and expressed pastime in connecting with us.”
- Authentic Spanish output: “Me pareció interesante tu interés en nuestros servicios.”
- Translation in English: “I discovered your pastime in our services and products fascinating.”
Once more, it’s technically correct, but it surely’s redundant and robot. It’s the type of phraseology that makes a reader prevent and pass, “Did a bot write this?”
The takeaway: Even if the interpretation is correct, the tone can also be off. And tone is the entirety in advertising and marketing.
The Shift from Translation to Language-aware Instructed Design
At this level, I noticed we wanted greater than AI outputs — we wanted a machine for directing the AI to suppose like a multilingual marketer.
I constructed a language-portable steered framework — a structured steered that would adapt throughout languages whilst respecting every one’s distinctive grammar, tone, and cultural context.
Right here’s What Modified
As a substitute of 1 static steered, I broke the good judgment into variables:
- : Goal language (e.g., Spanish, French, German)
- : Pronoun and tone stage (“tu” vs. “usted”, “vous” vs. “tu”)
- : Inbox-friendly, conversational, skilled
- : Direct vs. suggestive phraseology
- : Enforced the place grammar allowed
We additionally added transparent, language-specific laws.
Instance (Spanish):
- Use tú constantly, by no means usted (too formal for our emblem)
- Keep away from gendered adjectives like interesado/interesada when conceivable
✅ “Mostraste interés en … ”
❌ “Estuviste interesado en … ”
Instance (French):
- At all times use vous, now not tu, in B2B messages
- Keep away from ambiguous endings like intéressé(e)
✅ “Vous avez montré de l’intérêt … ”
❌ “Tu t’étais intéressé(e) … ”
Why This Shift Mattered
In English, a pleasant CTA may seem like:
“Would you be to be had for a short lived dialog on one of the vital following days?”
We attempted immediately translating it into Spanish:
“¿Quieres agendar 15 minutos para hablar sobre lo que estás buscando?”
It used to be grammatically right kind, but it surely sounded too informal and unprofessional in a B2B context. Now not pushy, simply fairly off-tone.
So, we reworded it to be pleasant however formal:
“Si te parece bien, podemos agendar una conversación breve esta semana.”
This interprets to:
“If it really works for you, we will be able to time table a brief chat this week.”
Right here’s every other instance in French:
- Authentic output: “Souhaitez-vous prendre rendez-vous pour en discuter ?”
(“Do you want to time table a gathering to talk about this?”) - New model: “Auriez-vous 20 mins pour voir remark HubSpot pourrait concrètement vous aider?” (“Would you’ve gotten 20 mins to peer how HubSpot may just nearly toughen you?”)
The second one model provides price to the CTA. Now not simply time — however function.
Backing It Up With a Stakeholder Questionnaire
Localization isn’t only a linguistic factor — it’s a trade alignment factor.
To get it proper, I created a easy stakeholder consumption document and shared it with advertising and marketing ops, regional entrepreneurs, and content material leads. The objective used to be to align early on tone, content material barriers, and regional sensitivities.
Those are one of the questions I requested:
- What stage of ritual is acceptable to your marketplace?
- Must we steer clear of gendered phrases?
- Are we able to reference the consumer’s corporate or product utilization?
- How direct must we be in inquiring for motion?
- Are there idioms, cultural references, or phrasings we must steer clear of?
We were given some beautiful fascinating insights.
For instance, in some areas, stakeholders most popular now not to reference the recipient’s corporate sort within the reproduction, even supposing that used to be commonplace in English (e.g., “I noticed that you simply lend a hand startups with HR”).
The localized selection changed into extra normal:
“Entiendo que están buscando formas de mejorar sus procesos internos.” (“I perceive you’re taking a look to give a boost to inside processes.”)
The result of this survey helped create readability between content material, ops, and regional advertising and marketing groups — and dramatically lowered our revision cycles.
The Ultimate Product: Human-sounding Emails at Scale
With the up to date steered and consumption framework, the brand new outputs have been right away higher.
Ahead of:
- Authentic output: “Hola [FirstName], soy María de HubSpot. He visto que has navegado nuestra plataforma y parece que te interesa nuestro producto.”
- English translation: “Hello [FirstName], I’m María from HubSpot. I noticed you’ve browsed our platform and it kind of feels you’re inquisitive about our product.”
After:
- Authentic output: “Soy María de HubSpot. Vi que estuviste explorando los angeles plataforma y que querías saber más sobre cómo podemos apoyar tu negocio.”
- English translation: “I’m María from HubSpot. I noticed you have been exploring the platform and sought after to be told extra about how we will be able to toughen what you are promoting.”
And stakeholders spoke back definitely:
- “This after all appears like any person from our workforce wrote it.”
- “Absolute best tone — herbal and native.”
- “No gender mistakes or bizarre formalities. We will in reality use this.”
Even higher, we didn’t want to write separate activates for each and every marketing campaign. The similar core framework now powers AI-generated messages in a couple of languages — with constant high quality.
Takeaways for Entrepreneurs
Whether or not you’re running on AI reproduction, world commercials, or multilingual content material, right here’s what I realized:
1. Don’t simply translate — localize for intent.
Literal translations gets you “technically right kind” content material. However most effective localization will make it land.
2. Use activates like ingenious briefs.
Come with tone, formality, CTA taste, gender neutrality, and different language laws as variables. Don’t go away nuance to likelihood.
3. Construct language-aware templates.
Languages behave another way. Plan for such things as verb conjugations, pluralization, and sentence rhythm prematurely.
4. Get comments early.
Use a stakeholder consumption document ahead of era, now not after. You’ll steer clear of transform and misalignment afterward.
5. Goal for an actual, human tone.
In case your AI output doesn’t really feel like one thing you would write to a buyer, it received’t convert. Learn it aloud. Would you hit ship?
AI localization is a advertising and marketing ability now.
This venture taught me one thing that has caught with me since: The way forward for world advertising and marketing isn’t almost about scaling content material — it’s about scaling context.
The firms that prevail with AI received’t be those who generate essentially the most content material. They’ll be those who generate the maximum resonant content material as a result of they know the way to steered for it. And that begins with figuring out the languages your consumers talk — in additional techniques than one.
