How Pernod Ricard built AI apps that influence $2B in sales

Company overview

IIndustry: Wine & Spirits

Size: 19,000+ employees

Global reach: 160+ markets, with direct sales in 73

Portfolio: 200+ premium brands, including Absolut, Jameson, Malibu, Chivas Regal, Beefeater, and Kahlúa

Retool products: Web Apps, Workflows, Retool Database, Retool Storage, Retool Email

Use cases: AI sales support (D-Star), AI forecasting (Quanto), market analysis (Maestria), VIP client management, event systems

“There’s no way you can go live with a vibe-coded solution. It might work for demos, but we build enterprise-grade technology that has to scale across 30 countries. Retool lets us prototype and go live in the same environment.”

—Pierre Yves Calloc'h

When legacy enterprise tools limit how quickly you can build

Pernod Ricard, the global leader in premium spirits, began exploring AI more than five years ago. But like many large, multi-national enterprises, scaling AI was an uphill battle. Every market operated differently: sales teams used their own systems, local data models, and manual processes, making it hard to build AI solutions that worked for everyone.

And, in the five years since Pernod Ricard’s journey began, AI has evolved from an experimental tool to an essential tactic. Builders like Dr. Romain Delgrange, Gen AI Program Lead, needed to move quickly, but couldn’t without the right solutions. David Lepicier, the Global AI Director in charge of establishing and managing infrastructure, needed production apps his team could integrate with legacy systems securely. And leaders like Pierre Yves, Chief Digital Officer, needed to know the tools they were building weren’t part of another hype cycle.

“There is the fear of missing out on a huge opportunity, but at the same time there is the fear of investing in things that will not work or provide value,” Pierre Yves says. “The way we address that is by selecting the right programs, testing quickly, and proving adoption.”

Pernod Ricard’s existing software stack—Salesforce automations and Streamlit dashboards—wasn’t made for complex, production-ready tools. The custom GPTs the team was building also weren’t going to be enough.

Streamlit worked well for data scientists, not for 2,500 field sales reps managing 260 routes across the globe. Salesforce was enterprise-grade, but it took too long to customize. The company needed something that was both production-ready and internally scalable, and there was no obvious path forward.

“Before Retool, we had a lot of challenges developing scalable interfaces that worked for all our teams,” says David. “We needed something secure, customizable, and fast enough to serve different markets.”

Three apps, one platform, immediate impact

Pernod Ricard’s team needed to accelerate its AI adoption beyond basic GPT demos to earn buy-in from senior leaders. Their goal was putting AI-powered tools in the hands of global sales and marketing teams. To do that, they needed a way to build apps that were secure, fast, and easy to use across regions. That’s when they turned to Retool.

Within months, Pernod Ricard’s team built and deployed several apps, including three AI-powered solutions for critical business challenges:

• D-Star: AI-driven sales optimization for 2,500+ field reps across India, the US, China, and Europe. D-Star analyzes customer performance and route efficiency to recommend daily itineraries. Influences $2B in annual sales.
• Quanto: Forecasting engine predicting 20,000 SKU-by-state combinations with +4-point accuracy gain. Quanto replaced manual spreadsheet forecasts with AI-powered precision for inventory and budget allocation.
• Maestria: Real-time market analytics across 18 markets. It replaced slow quarterly reports with instant insights from digital channels and distributors.

When Delgrange needed to prove generative AI could deliver similar value, his team of junior developers built a working data mapping prototype in just one week. It solved a longstanding issue for Pernod Ricard’s global workforce—data from 160 markets arriving in different formats and units.

“700 milliliters from Germany becomes 70 centiliters in Spain,” Delgrange explains. “The goal of the application was for people to map the row market data to the actual reference table that we have inside. It’s replacing very tedious manual work.”

With Retool, they automated what data analysts had been doing manually for years.

“It wasn’t just a wireframe—it was really something that could work,” says Delgrange. “It took us only one week, and it was working fine enough to perform AB testing.”

The accessibility was key. “Even I, someone who is not skilled in development, was able to create something that worked,” says Romain. Because Retool handled authentication, security, and data management automatically, the team was free to focus on business problems.

The team built two more gen AI apps in quick succession: a meeting-to-action translator and a document Q&A tool for 100+ page reports. Each went from concept to working prototype in days.

Going from pilots to production on a platform built for global scale

What used to take a year to prototype now ships in months—and that speed has become Pernod Ricard’s competitive advantage.

“I do believe that we are at least five times faster in terms of development,” David says. “I don’t even talk about maintenance because maintenance costs are so much lower with Retool than traditional custom development.”

Pernod Ricard’s teams now deliver apps five times faster and at lower maintenance cost than traditional development approaches. That speed, paired with Retool’s flexibility, has driven a 20% higher adoption rate among end users.

“When you move fast, people engage,” says Pierre Yves. “That usage increased the adoption rate by about 20%, which means 20% more impact on the business—more sales or marketing efficiency.”

Secure deployment for thousands of users

Pernod Ricard now has the advantage of security and scale. Employees are building quickly, and doing it safely in the Pernod Ricard environment.

“When it comes to deploying something secure for 3,000+ employees, you need a robust platform integrated into all our legacy systems,” says David. “Retool connects all the dots between our AI agents, our data, and our people.”

Retool integrates with Pernod Ricard’s complex technical ecosystem—Snowflake, PostgreSQL, MySQL, Jira, Microsoft Teams, and Google Sheets—allowing the company to standardize AI app development while meeting strict compliance and data privacy standards. LLM-agnostic integration means teams can build and experiment with any model without locking in.

It also opened the door for non-developers to participate. “Retool has freed a lot of time for our AI team to develop new things,” says David. “It’s enabled business teams to drive their own strategies because they can manage parameters themselves.”

A cultural shift in how Pernod Ricard builds

Pernod Ricard’s transformation was as much about culture as code.

Retool compressed the feedback loop between engineers and business teams: developers, designers, and field reps co-created apps in real time, implementing feedback during onboarding sessions.

“In one training session, someone asked to change part of the UI and we pushed the update in an hour,” ways Pierre Yves. “People were amazed. They weren’t used to enterprise apps that move that fast.”

That agility created trust, and adoption followed. Today, Pernod Ricard’s sales, marketing, and analytics teams use production-ready AI apps.

What’s next

Pernod Ricard now treats Retool as its unified environment for AI innovation—a place to prototype, test, and scale internal AI tools across its 200-brand portfolio.

Beyond D-Star, Quanto, and Maestria, teams have used Retool to build VIP client management systems, event check-in platforms, and even AI-powered marketing analytics dashboards.

“Retool is a superpower,” said Dr. Romain Delgrange. “It lets us go from idea to something that actually works. It’s our flashlight—we can see much further than before.”