Case Study – Liquats Vegetals

Liquats Vegetals leverages artificial intelligence with VOLTA Forecast to improve forecast reliability and streamline its S&OP process

Liquats Vegetals is a leading player in the European plant-based beverage market. Based in Catalonia, the company produces and markets the brands Yosoy, Monsoy and Natrue, distributed in over 40 countries. Since 2017, Liquats has relied on the BEVOLTA platform to manage its production planning and the replenishment of components and raw materials. This long-standing partnership has helped establish reliable and well-controlled flow management processes.

As volumes increased and product ranges expanded, demand planning became more complex. In 2021, Liquats decided to take things further: integrating artificial intelligence at the heart of its forecasting process. Upon the official release of VOLTA Forecast, BEVOLTA's machine learning-based forecasting module, the company agreed to become one of its first pilot clients.

Anticipating better, without overloading processes

At that stage, Liquats already had a solid planning foundation. Forecasts were built on historical data, manually enriched by planners and commercial teams, following well-established business logic. But the process consumed time and resources. And it became increasingly difficult to adapt to a more unstable market context: last-minute promotions, unusual weather conditions, emerging nutritional trends, trade-offs between channels…

The aim wasn’t to disrupt existing expertise, but to rely on a technology capable of processing large numbers of variables, automatically generating reliable forecasts, and offering teams a more robust, dynamic and usable baseline.

A collaborative and progressive project approach

The project was co-led by the supply chain teams, IT leads and BEVOLTA consultants. It began with a functional scoping phase to define the target scopes (product families, geographic areas, update frequencies), involved stakeholders, priority use cases and data interfaces.

The first step was to connect existing data from the JDE ERP system: sales history, stock levels, product master data, and lifecycle information. Based on this foundation, BEVOLTA helped Liquats identify and structure explanatory variables relevant to their operations.

Explanatory variables tailored to Liquats' business

Beyond standard calendar effects and promotions, several elements proved especially useful in the Liquats context:

- Local weather data, particularly to anticipate demand for fresh or plant-based milk alternatives in warmer regions.

- Client location: some export customers (especially in Asia or Northern Europe) show consumption patterns that are highly dependent on cultural or logistical factors.

- Specific sales campaigns, with volume-based incentives that significantly impact demand curves.

- Cannibalization across product lines, for instance between almond-, soy-, or oat-based ranges, which sometimes fulfill similar needs but evolve differently depending on the market.

- Packaging changes, which can trigger anticipation or delays in order intake.

- Logistical variations, including production or transport constraints that indirectly affect expressed demand.

These variables were integrated into the VOLTA Forecast engine. Thanks to the AutoML approach, the platform automatically selects the best-performing models based on available data and the specific characteristics of each product/channel/country combination.

A solid foundation for planning and collaboration

In less than three months, the first models were trained, tested, validated and deployed in production. Users now access three levels of probabilistic forecasts (P10, P50, P90), directly integrated into the BEVOLTA platform. These forecasts feed into Liquats' planning, simulation and decision-making modules already used daily by the supply teams.

Beyond improved forecast accuracy, the main shift occurred in the way the S&OP process is managed. Forecasts are no longer manually assembled by each team—they are automatically generated and shared as a common base. This leads to:

- Shorter, better prepared meetings

- Quicker and more informed decisions

- Stronger alignment across supply, commercial and executive teams

More alignment, saved time, and better decisions

Implementing VOLTA Forecast enabled Liquats to gain consistency, agility, and coordination across its planning processes. Time saved by automating forecasting is now invested in analysis, coordination, and building more relevant scenarios. The tool doesn’t eliminate uncertainty, but allows it to be handled more methodically—and aligns key stakeholders more quickly around realistic action plans.

Planners remain central, with an expanded role: they’re no longer just producing numbers, but driving conversations between data, industrial constraints, and commercial dynamics.

Moving toward broader intelligent forecasting

Liquats is already exploring further use of VOLTA Forecast: integrating e-commerce trends, forecasting raw material needs based on harvest cycles, and dynamically adjusting procurement in response to market signals. The collaboration with BEVOLTA continues as a long-term partnership to build a more agile, resilient and market-connected supply chain.