Greg Stewart is the founder and CEO of Geco Predictive Weed Control. He has a bachelor’s in physics, a master’s degree in applied mathematics, and a PhD in control engineering. Before launching Geco, he cut his teeth making products to enhance quality and performance across a wide range of industries like pulp and paper, oil and gas, and manufacturing. Today, Stewart has shifted his focus to digital agriculture where he helps farmers identify where and when weeds will grow.

What first piqued your interest in agriculture and agrifood?

My career has involved making new products using sensors, data, and automation, often for new markets. I was lucky enough to get exposed to a wide range of industries – paper making, diesel powertrains, oil and gas, manufacturing, and optimization of data centers. Every industry is on its own journey in digital adoption. Some have processes that are completely automated and they are looking at “lights-out automation” where everything is automatically controlled, and other processes are manually operated. Often it is partly a function of the potential available value versus the technical feasibility of the problem that determines the growth of digital approaches. This gave me a broad range of exposure and the opportunity to learn about technology, and how people do or don’t adopt digital technologies into their daily work.

I really wanted to leverage these experiences for human good, and I was seeing more and more agricultural measurements and data becoming available, especially in agronomy. Sensing can be satellite, drone, onboard field equipment, or sensors poked into the ground – which are providing increasingly measured insights into a wide variety of aspects of cropping.

What’s something that surprised you when you started working in digital agriculture?

I think it would be the willingness of farms to perform experiments that could expose their farm to a new upside opportunity. Digital or non-digital experiments are a common part of farming. Sometimes we will hear agtech or tech people say that the A/B testing of software engineering has no place on a farm. However, strip trials are pretty much the same thing but implemented in a field – a side by side comparison of an input or technique on a limited portion of your business to help you decide whether to more broadly deploy a change in practice.

Why do you think digital agriculture is important for farming?

I will talk about the specific challenge of controlling weeds, as that is the main focus of Geco. It is well-understood that weeds must be controlled, or they will significantly rob crop yields. However, a key challenge is that a population of weeds exists over hundreds of acres, and some of their properties, such as their emergence within a season or the evolution of herbicide resistance, occur at time scales measured in months and even years. This scale of space and time makes them difficult for humans to monitor and control.

At Geco we became interested in the possibilities of using data to quantify and optimize weed control at the population level. This led us to the development of predictive control of weeds. We use a combination of AI and agronomic modeling to get an in-depth understanding of weed behavior in a given field. We use this understanding to better fight weeds by providing predictions of where weed hotspots will emerge in the field, enabling proactive and targeted action – whether using residual chemical herbicides or using nonchemical techniques such as increased seeding density. We can also detect when and where resistance may be evolving and enable directed scouting.

What do you think would help digital agriculture develop and advance more quickly?

In the earlier days of agtech, there was a lot of excitement around obtaining visibility into crops and operations. Tools like satellite imagery and NDVI maps promised the potential for a new and more detailed view on what was happening in the fields. There was a period of curiosity about these technologies, but only limited real engagement and integration into farm practices. It seems like more people are now deciding that digital agriculture – like any other innovation – is much more likely to be adopted if it solves a real problem and creates value on the farm, while not requiring an unreasonable amount of effort or expense to be implemented by busy people.

Geco Predictive Weed Control is one of EMILI’s Emergence Grant recipients. Since 2023 EMILI has worked with Geco at Innovation Farms to assess the performance of their technology for predicting weed locations and detecting emerging herbicide resistance.

This profile is part of EMILI’s This is Agriculture series, highlighting talented and diverse individuals across the digital agriculture sector. While individuals working in agriculture come from a variety of backgrounds, they share a common interest in growing and strengthening Canadian agriculture to ensure an environmentally and economically sustainable future for generations to come.