E-commerce case study
How Millesima deployed generative AI in its e-commerce and brought the whole team on board
Industry: Wine & spirits e-commerce | Company: Millesima (Bernard Group of Bordeaux)
DEFINITION
Generative AI in e-commerce: what are we really talking about?
THE CONTEXT
When generative AI becomes essential for a wine e-commerce business
A European leader in online wine and spirits sales, Millésima is a Bordeaux-based company affiliated with the Bernard Group of Bordeaux. Its content team manages the brand’s multilingual communication: product pages, newsletters, blog articles, and documentary databases. A colossal volume. An uncompromising standard of quality.
Brice, Content Manager and the company’s AI lead, was the first to see the potential of generative AI to transform his department. And the first to take action, as early as March 2023.
THE CHALLENGE
Deploy generative AI in e-commerce without fragmenting tools.
In March 2023, Brice began experimenting with generative AI tools. The results were very promising. Soon, it became clear: if generative AI could impact his department, it could impact the entire company.
But deploying generative AI across an e-commerce organization is not so simple. The market for language models is booming. OpenAI, Anthropic, Google… Each provider offers its own interface, accounts, and privacy rules.
Without a unified solution and a dominant model, the risk is real: scattered data, incompatible methods, compromised security. It was necessary to constantly juggle between different models and providers depending on the projects and model releases. For a French company operating internationally, this was unacceptable.
The solution was not yet known, but the need for it was perfectly clear.
"We couldn’t allow each user to create their own account with OpenAI, Anthropic, and so on."
THE DECISION
Swiftask to centralize generative AI in e-commerce
Swiftask meets all these requirements: early market maturity, data hosting in France, and an interface designed for non-technical users.
The decision is made. Swiftask becomes the official platform to provide Millésima teams with access to generative AI.
Brice set out in search of a solution capable of meeting several non-negotiable requirements:
Centralize access to all major generative AI models from a single interface.
Manage AI agents independently, without advanced technical skills.
Be accessible to business experts, not just developers.
Host data in France, in compliance with digital sovereignty requirements.
"Swiftask addressed the challenges of a French company, with data hosted in France. It was an undeniable advantage."
THE DEPLOYMENT
A gradual adoption of generative AI by the teams.
Brice chose a gradual approach to bring the teams on board with this digital transformation.
The three step method:
Short training
One hour is enough to get started. No jargon, no overload. The goal is simple, every employee should understand what generative AI can do for them.
Demonstration workshops
The company organizes sessions to showcase what has already been achieved. Inspiration through example.
Voluntary adoption
No forced deployment. Each employee engages with generative AI at their own pace.
"The pitch is fairly quick to make. And it’s quite easy to be convincing."
USE CASES
Generative AI use cases in e-commerce: where Millesima proves its value
Thanks to a dedicated AI agent, the workflow is now smooth and accessible to any business expert:
.png){kind=small}
.png){kind=small}
.png){kind=small}
.png){kind=small}
.png){kind=small}
.png){kind=small}
THE RESULTS
Generative AI is concretely transforming Millésima’s e-commerce.
Comparison table: before and after adopting generative AI
| Indicator | Before generative AI | After generative AI | Gain |
|---|---|---|---|
| Time to translate a product sheet | 20 minutes | Less than a minute | 95% time saved |
| AI tool adoption rate within the team | 0% | 50% (used 2-3x/week) | 50 active users |
| Number of deployed AI agents | 0 | One agent per business use case (translation, newsletter, matching...) | Automated processes |
| Time to translate a product sheet | 20 minutes | Less than a minute | 95% time saved |
"It’s truly an indispensable tool."
A digital transformation accelerated since September 2025
Since September 2025, the acceleration has been even more pronounced. The new generative AI models (Claude Sonnet/Opus 4.6, Gemini 3) deliver high-quality results without complex prompts further lowering the entry barrier for all employees.
BRICE’S ADVICE
Don’t wait ,equip yourself properly for generative AI.
For Brice, the question is no longer whether generative AI is useful in e-commerce. It’s essential.
The real challenge? Getting employees on board with this digital transformation. And for that, you need a solution that unifies, secures, and harmonizes access without adding complexity.
And the most compelling proof? Millésima had identified the need for such a solution even before knowing it existed.
"If we want to bring all employees on board, support them, and give them access to the main models, we obviously need an aggregation solution a solution that manages security, harmonizes access, and standardizes workflows."