Excerpt based on the original article published in Fortune en español magazine in its April 2020 edition. Written by Zacarías Ramírez. Illustration by Alfredo Quintana Garay “Tintorera”.
Elizabeth Hernandez is more than Sauza’s Agave Planning Manager: she is in charge of the project with which the tequila company is moving towards smart agriculture, the new concept of crop management that relies on data, genetics, artificial intelligence, and drones.
Using drones on agave plants has allowed Hernandez and her team to respond quickly to contingencies and prepare their crops to meet future demand for tequila.
The first problem that drones solved for Sauza was not technical but labor-related. The control of their 30 million agave plants was done by two teams of 20 workers each who counted the plants twice a year; however, there was a shortage of people willing to be hired. “It is a tough job; you have to imagine the 40 degrees Celsius heat in Nayarit and go counting agave by agave, (verifying) that the plants are still there, that they are not damaged,” Hernandez acknowledges.
But it is one thing to have a drone that flies over and photographs crops, and another is to have one that can differentiate between an agave plant and weeds or rocks.
Elizabeth realized this early on. She recalls that no equipment could identify agave; it had to be developed and made to do the plant counting that used to be done by the two teams of workers.

And this is where Manglar*, a Colombian company that developed an algorithm with which the drone could recognize tequila agave, came into play. “We already had experience with crops in Colombia, so what we did (for Sauza) was to build an artificial intelligence model for agave cultivation,” adds Germán Medina, CEO of Manglar. “It was the first client with whom we worked on agave crops.”
Soon, however, counting agaves was not enough. Sauza thought it would be good to identify areas of the crops invaded by weeds, as this affects the weight of the plants at each stage of their six-year development. And why not go for the whole menu?: to detect sick or infested plants from the air, record their location on maps, and plan herbicide irrigation.
Drones produce agrochemical savings because, unlike when applied by people with portable pumps or small planes, the drone is guided by a map to make a focused application. Drones become crucial for Sauza as some of their agave fields are up to 3.5 hours away by car.
With the help of their drones, Sauza counts agave plants every two months in 60% of their crops, but they aim to increase this to 90% (they do not aim for 100% due to the presence of organized crime in some areas where they have crops).

This higher frequency in the counts allows their engineers to read the information the week after it is collected, which is almost real-time knowledge of what is happening with the plants. Data and photos are collected from each crop and then pasted together in an orthophoto that allows engineers to see the condition of the farm through a digital platform, make decisions, and plan. “They can say: ‘Now I know that this crop has weeds, diseases, wilted agave… so in my next weekly plan I’m going to include the solution to this problem,'” exemplifies Hernandez.
There is no need to wait until their agaves are six years old to see the results. A plant must have a certain yield every year, so with the help of drones, they monitor according to pre-established indicators and can, from now on, calculate how they will respond to the projections of tequila demand in six years. “I won’t see the results at the end, but year by year, and that allows us to have a plan,” concludes the interviewee.
What began four years ago as an alternative due to the difficulty of finding people to work in the field is now a cornerstone in the company’s planning.
*Manglar (CNX in the original article) was born as a business unit of CNX, specializing in advanced analytics, big data and artificial intelligence, and focused on agriculture, mainly in aerial imagery analysis. Currently, all efforts go into Manglar.