Further complicating the situation, the enormous variety of growing scenarios precludes a ‘one size fits all’ approach to technological solutions. A high-tech greenhouse in The Netherlands for example, faces substantially different challenges than open field farming in Sub-Saharan Africa. From climate variation, to Internet connectivity issues in remote rural areas, to access to capital to invest in often costly new tech, solution providers find themselves in a tremendously difficult position with respect to addressing the market as a whole.
That being said, while adoption of cutting-edge technologies has been somewhat slower within the sector than say, finance or manufacturing, over the past few years growers have been urgently searching for solutions to problems threatening the viability of agribusiness — particularly within Controlled Environment Agriculture (CEA). Soaring energy prices and the declining availability of skilled labour have created situations in which shrinking profit margins and operational challenges have resulted in some growers electing to shutter their farms rather than swim against the tide of unfavourable market forces.
While exorbitant energy prices — particularly in Europe — are a feature of current geopolitical instability, the lack of skilled labour marks a societal shift away from careers in agriculture, and toward, for example, vocations in the information economy. As a result, this worker deficit has become the primary challenge faced by a majority of the world’s growers today.
To address this problem, technology solutions are being created by companies ranging from scrappy startups, to established industry players. Robotics have been employed to assist in a variety of activities: from picking, to packing, to transport between the two. While showing great promise for the future however, these systems are often costly, and many have yet to match the performance of their human rivals at certain tasks.
Well-established Agtech firms on the other hand, have devoted significant time and resources to leveraging environmental control system infrastructure in support of creating fully autonomous growing platforms — the holy grail of horticulture. However, again this technology is still in the early stages of development and adoption. Semi-autonomous systems such as Priva’s Plantonomy for example, focus instead on the dramatic reduction in the number of control parameters a grower needs to concern themselves with: from hundreds to just six. This eliminates the steep learning curve associated with climate computers, which can take new growers literally years to master. The system steers crop development by automatically reacting to transpiration in the plants — allowing them to influence their own growing environment. This both reduces labour and produces more consistent yields.
Over the past decade or so, fledgeling AI companies have entered the sector with products focused on precision agriculture: helping farmers optimize resource utilization by providing precise recommendations for seeding, irrigation, and fertilization. Others focus on crop health monitoring, integrated pest management, weed identification, and even supply chain optimization.
At the outset of their entry into the sector, a lack of internal horticultural expertise, and immature methods for data collection, validation, and analysis led to skepticism of this technology based on its early failure to deliver concrete, repeatable results. Slapping an AI label on a software product was often more marketing than reality, and the resulting failures in its application to problems associated with agriculture led to hesitancy in its adoption. As data science has come of age however, and growers have witnessed the capabilities of generative AI products such as ChatGPT and Midjourney, the market has come around to take a second look at what it actually has to offer.
One such firm now enjoying success in the sector is Fermata, a data science company focused on the early detection of pests and disease through computer vision. Founded by Forbes 30 Under 30 inductee Valeria Kogan PhD, Fermata aims to tackle both labour and profit margin issues by minimizing the need for human scouting, dramatically reducing crop losses, and improving the quality and quantity of yields. Further, detecting pests and disease at their earliest stages and alerting growers to their exact locations, the Croptimus™ platform enables workers to spot-apply remedies instead of having to blanket entire growing areas. This targeted mitigation strategy not only saves on expensive crop inputs like beneficial insects and pesticides, but also helps growers operate in a more sustainable manner — a plus for the bottom line and the environment.
Among the primary issues hindering the success of early competitors in this field were access to quality data to train machine learning models on, and a lack of data labellers with domain expertise — generally outsourced — to validate and correct the output of their algorithms. Fermata has addressed these problems by setting up its own sterile research facility, where plants are purposely infested or infected and tracked over time. This approach provides consistent, quality data which provides a solid base for the detection algorithms. As every grow is different, the models continue to be trained and improve still further in each unique production environment.
The company has also hired a large team of in-house data labellers who undergo extensive training to ensure the first tiny signs of pests or disease — including the damage they do from underneath leaves — are being properly identified and catalogued by Croptimus™. This early detection can save growers up to 50% on scouting labour, 30% on crop loss, and 25% on crop inputs.
So, while the adoption of cutting-edge technologies in agriculture has had a bit of a rocky start, now that they have become substantially more mainstream, with demonstrated success across multiple growing scenarios, we can expect a bright future for both growers and tech entrepreneurs seeking their fortunes within this vital market.
Article in Hebrew