Mekki Maalej, founder of Agri Safe: "AI does not replace farmers, it transforms data into water savings"

Agriculture absorbs about 70% of global freshwater withdrawals (according to FAO), and climate change is straining this resource: rising temperatures, longer droughts, and more irregular rainfall reduce aquifer recharge. In France, irrigation draws from both rivers and groundwater – the latter providing nearly two-thirds of our drinking water. However, these aquifers are dwindling: as of June 1, 2026, the Bureau of Geological and Mining Research reported declining levels at 77% of its monitoring points, a situation worse than a year earlier as summer approached. In this context, agriculture faces an increasingly tight equation: produce more while consuming less water.

Water has also become a political battleground: water storage for farmers, at the heart of the French emergency agricultural bill, crystallizes tensions between the agricultural world, environmental advocates, and public authorities. But another response is unfolding on the ground, in the way irrigation is managed.

For Mekki Maalej, founder of Agri Safe SAS, the answer lies in data-driven irrigation. The son of a Tunisian farmer, he developed his first solution when his father, ill, could no longer visit his fields daily. From this field experience, not from a laboratory, emerged a platform capable of connecting existing irrigation equipment, collecting agronomic data, and helping farmers make more precise decisions using artificial intelligence. His ambition is not to replace human expertise but to empower it to manage a resource that has become more precious than ever. Three questions for Mekki Maalej.

Can we still produce more with less water?

Yes, provided we rethink our irrigation methods. Many farms still irrigate based on habits or field observations. This experience remains valuable, but it is no longer sufficient in the face of increasingly variable climatic conditions.

I started in Tunisia, a country on the front lines: after several years of drought, reservoirs fell to critical levels, around 20% of their capacity by the end of 2024. Water had to be rationed in the middle of summer. That’s when we equipped over 500 hectares, with very concrete results: on olive trees, 35% savings in water and fertilizers, 20% reduction in pumping energy costs, and up to 12% increase in productivity; on almond trees, about 25% less water for an additional 10% yield; on vineyards, 20 to 25% water savings, 8 to 10% productivity increase, and better grape quality.

The goal is not just to save water: it is to deliver the right amount at the right time, according to the plant's actual needs. This also brings considerable time savings: manually opening a valve and waiting several hours, block by block, is a task that automation eliminates. With climate change, this precision becomes a factor of competitiveness as well as a challenge for resilience.

How does artificial intelligence change the way we irrigate?

It transforms thousands of data points into irrigation decisions that no farmer could make manually. Our system combines temperature, humidity, and pressure sensors with weather data, soil type, crop variety, planting age, and development stage. It deduces the most suitable irrigation scenario and, if necessary, fertigation: for example, more water and nitrogen during a root growth phase, less water during full bloom so the tree can concentrate its resources on the fruit.

Pressure sensors serve another purpose: they detect leaks and blockages in real-time and immediately alert the farmer.

We have also developed an explainable AI. It does not just make recommendations: it explains why, indicating the weight of each parameter in the decision – weather, temperature, or soil moisture. This transparency is essential, as many farmers already have sensors but struggle to turn this data into concrete actions. This is precisely where the added value lies.

What does smart irrigation look like in practice? Can it be deployed everywhere?

One of our priorities has been to adapt to existing farms: not to replace an entire installation, but to connect the equipment already in place – pumping motors, solenoid valves, filtration or fertigation stations – remotely controllable from a smartphone or computer, without heavy construction work.

The system relies on a communication box powered by small photovoltaic panels. It covers a radius of about seven kilometers, or nearly 15,000 hectares, and can connect up to 32,000 devices. It can be shared among several farms to reduce costs: at a vineyard where we are starting on 20 hectares, about thirty devices are sufficient.

The solution is aimed at all sizes of farms, from a few hectares to several hundred, in fruit growing, viticulture, market gardening, or large-scale crops — we equip both vineyards and fields of corn or wheat.

Beyond technology, the main challenge remains human. Many farmers are discovering these tools and must learn to trust them. However, in the face of dwindling water resources, irrigation can no longer rely solely on human experience: it must be based on data and artificial intelligence to support decisions, while allowing the farmer to remain in control of their choices.