Adaviv captures plant-level data using artificial intelligence and machine vision to improve the production of high-quality crops at scale.
Growing any kind of plant, be it in a house or on a farm, requires patience, managing resources and paying attention to the environment. Part of the success comes from instinct; another part from the experience of trial and error. But using data? It’s not done enough, mostly because there isn’t enough that’s actionable. It means that farmers don’t know how to best spend, so often they overspend on inputs and labor. It means decreased yields and increased costs, says Ian Seiferling, CEO and co-founder of Adaviv. The challenge is that getting consistent performances from every plant requires individual care and the labor to provide it.
We’re making farms more profitable by automating time-intensive tasks
Adaviv is looking to help controlled environment farms make more efficient and informed labor decisions, all to get the most out of crops with the least amount of spending. Through artificial intelligence and machine vision, its platform for precision agriculture can capture plant-level data and give it back to growers, taking out the guesswork, and showing them what’s effective now and what they should be planning for in the future.
“We’re making farms more profitable by automating time-intensive tasks,” he says. “It allows growers to focus their attention where it matters most.”
Getting In and Getting Up Close
Adaviv started in 2019, but the roots go deeper. Seiferling grew up in Saskatchewan, a province where seemingly everyone had a farm, he says. That early connection to nature led him to studying environmental sciences and biology. While doing doctoral and post-doc work at MIT on the scalability of growing food around cities, and working in the Senseable City Lab, he met his eventual co-founders, Mo Vazifeh and Julian Ortiz.
They saw that new technology, like AI and vision technology, gave the ability to collect plant-level data. That information would allow growers to see what was actually working and allow them to streamline operations, allocate tasks better, and improve yields. “Overall, it would raise the bar on the quality of the work being performed,” Seiferling says. And the ideal target would be greenhouses and indoor farms.
Those spaces have certain advantages in being able to control elements like water and energy, but they also work with certain constraints, some self-imposed. The scale is enormous and management is labor-intensive. The subjective nature of many decisions leads to inconsistencies in productivity. It requires people to be in the field, but the managers and cultivators, those with the expertise, can’t be everywhere. “They don’t have the time, don’t have the bandwidth,” he says.
The result is that rather than knowing what specifically works, a grower might do what they think will work, and, for example, use too much of an input, like water or pesticide, and blanket a crop. As much as experience is in play, there’s guesswork as well. Farming relies heavily on the human eye, and without hard data, things get inevitably missed.
Adaviv’s solution is its patented SAAS technology, which includes the crop-scanning Mantis and Ladybug robots. The Mantis can move easily through indoor settings and tight aisles, giving root-level views and reducing the human labor needed for ongoing plant monitoring. The Ladybug offers full autonomy, navigating from above. Both give profiles from the side “where often many problems start” and can detect disease “days before” a person could, Seiferling says.
But collecting the data is just one part. Without context, it doesn’t mean much. Adaviv’s platform takes the information and assists field technicians by giving them a mobile, digital map of the area that automatically highlights pending tasks, completed tasks, and areas of concern. “It’s their AI farm companion,” as Seiferling calls it.
One constant challenge with farming is turnover; it’s 30-50 percent annually, he says. It makes it easy to fall behind on cultivation and never be able to catch up. With the platform, managers receive automated plans that give them complete visibility on the state of the farm. It allows them to allocate people and resources, but, because they’re getting real-time data, they can actually hire fewer workers while still “hitting 100 percent completion.”
Because the platform can uncover what was “more or less invisible,” Seiferling says, the owners and growers also get a big-picture view of their environment and will be able to see patterns, something that was elusive with traditional approaches. They now have the power to reconfigure their space, change the layout, adjust airflow. They can start planning at the optimal time for future crops, and so a process that once was measured in hours “can be managed in minutes and executed with greater accuracy,” he says.
Shaking Up the Industry
As Seiferling says, agriculture is a legacy industry. Farmers and growers have long-standing relationships with farmers, suppliers and vendors, which is a plus. The business comes with tradition, which can be a plus, but it can also mean that it’s slow to move and slow to change, particularly with labor-intensive processes. That doesn’t just apply to growers, but to everyone who has a hand in making things grow.
With data, everything for the growers changes for the better. Information goes from subjective to objective
The data that helps a farm owner cultivate better crops is the same data that can help farm systems like lighting, irrigation or pesticides products be more responsive, targeted and adaptive. It also allows for automation to be introduced, which in turn will allow companies to better direct and use resources and money. By using real-time data, it means that companies can implement labor efficiencies and streamline the lengthy trial-and-error process, things that at one time were never thought possible.
“With data, everything for the growers changes for the better. Information goes from subjective to objective. Decisions about crops and labor are now made with full visibility. Operations become leaner,” Seiferling says. “And the whole process is accelerated for maximum productivity.”