520,000 conversations. 9 brands. 12 languages
Four years designing the conversational strategy of a multinational food group: chatbots for nine brands, in thirty-three countries and twelve languages.
520,000 conversations. 9 brands. 12 languages.
The challenge:
A multinational food group wanted to scale its conversational strategy from a single brand to the entire portfolio. The problem was not technological — the chatbot platform existed. The problem was one of design: how do you design a conversational experience that works equally well in Spain, Mexico, Germany and Turkey? How do you maintain the personality of nine different brands within the same system? How do you define when the bot responds and when it refers to a person?
These questions are not answered in an Excel FAQ. They are answered in a rigorous conversational design process.
The approach:
We started with analysis of actual conversation patterns — what users asked, when they abandoned, what bot responses generated frustration. This analysis was the basis for redesigning the main flows and defining the derivation rules.
For each brand, we develop a specific conversational personality: tone, vocabulary, level of formality, how the bot presents itself, how it handles error situations. It wasn't the same bot with different colors — it was a different experience for each brand, built on the same architecture.
The cultural and linguistic adaptation work was one of the most demanding of the project. The same answer that works in Spain can be cold in Mexico or excessively informal in Germany. We iterate with real users in each market before each launch.
Over four years, the project evolved with technology: from rule-based chatbots to systems with natural language processing, with new flows for new brands and new markets every year.
The result:
9 brands of the group with an active conversational strategy. 33 countries. 12 languages. More than 520,000 conversations managed. 3% bounce rate — in a sector where the industry benchmark is between 8 and 12%.
The project demonstrated that conversational design on a multinational scale is not a technology problem. It is a problem of judgment about the user, about the brand and about when AI helps and when it hinders.
Applied services: Conversational design · Brand personality and tone · Flow architecture · Cultural and linguistic adaptation · Continuous optimization
Sector: Mass consumption
[→ See more projects in Intelligent Experiences (iX)]
