For many years, precision agriculture has largely been compartmentalized with the promise of true integration being just out of reach. Now, through the use of super-computer Watson and artificial intelligence (AI), IBM says it will draw all the disparate bits together into a whole that will usher in a more efficient, easier manner of growing crops and doing business through the agricultural chain.
To explain the company’s vision and what it expectsin the near future, Mark Gildersleeve, vice president, Head of Business Solutions Watson Media and Weather, spoke with Delta Farm Press in late November. Among his comments:
On what problems the company is trying to fix…
“As a personal story, I did a startup in precision agriculture in the late 1990s that ended up getting sold. Coming back to agriculture now, I was surprised at how little progress had been made over the last 25 years. Frankly, it’s disappointing.
“The problems that farmers have are still here. It’s an awful lot of work, an awful lot of risk and an awful lot of capital for a farmer to eke out relatively small margins compared to other industries. So, we need to help growers get to a higher level of profitability given the high amount of risk at stake.
“This industry has been made up of ‘silos.’ If you look at the ecosystem between lenders, bankers, input providers, growers and the government, it’s still very ‘siloed.’ Very little progress has been made in connecting up the siloes.
“When you look at the need to improve food quality and how to address sustainability, that silo nature makes it very hard. So, we’re looking at these problems anew and saying: ‘these problems are really big, are global and take a massive effort to fix them, and we believe we have a number of things we can execute to help begin to improve the situation.’”
On what is being done…
“There are things that are driving us to say, ‘why now?’
“The first is that IBM has many enterprise clients — whether food companies, banks, insurers, or income providers — who are beginning to understand that we need to connect the silos up. We need to connect the growers up with the other enterprises to improve (effectiveness) overall. So, we’re getting a push from some of the enterprise customers to help them connect up with the ag sector.
“Second, there have been continued improvements in technology. That includes everything from remote imagery to sensors on farm equipment to how … to get better decision recommendations for a grower. Those are all advancing and things we can take advantage of.
“Part of that is also artificial intelligence. We have done a lot of things on the research side of IBM all involving Watson, or some version of AI, that we believe can be brought together to make the sector work better. So, there have been a series of technical innovations and a series of revelations inside the sector. There are people now understanding we must take a different approach to solve the problems.
“IBM has been doing many things in the sector for many years. But now is when we’re trying to bring all the pieces together.”
On the first things a farmer who adopts the IBM approach will notice…
“We’re working with business partners who are typically supporting growers themselves. There are lots of enterprises that are serving growers directly today.
“We believe we can give growers earlier insight into what’s going on in their fields. If they make decisions earlier that will turn into money.
“Let me give a couple of examples. First, we can give them an early indicator of a pest or disease outbreak before that pest or disease is even in their field. So, it will give them a likelihood of disease outbreak based on remote imagery, based on the weather and the time of the year so we can predict when the pests will emerge. Many of them are based on weather.
“Second, we give them early and more precise indications on a lot of weather when it’s a good or bad time to plant, a good time to apply fertilizer, or a bad time to spray. For example, you wouldn’t want to make an application and have it rain three hours later and waste the benefit of the application.
“Third, we have a scouting application that enables a grower to take a picture of a crop that looks to have some kind of stress. (The app) is able to have Watson help identify the most likely relief for that crop stress.
“Fourth, we have a number of yield models that can help growers understand what the likely yield is throughout the growing season. That helps them better understand what they’re likely to have at harvest so they can make decisions about a marketing plan.
“As for an enterprise, let’s say you’re a food company and have growers under contract. Many food companies have growers under contract but don’t necessarily have enough insight into what the yield is coming off a grower’s field.
“I know of a food company this year that was really surprised their growers ended up being 50 percent short of the supply the company expected. So, one place we’re helping such companies is they can work with the growers and know, in real time at any time of the season, how much supply they can expect. That provides the company a much better indication of risk and how to mediate it.”
On what might be possible in five or 10 years…
“Artificial intelligence is still relatively in its early stages and I think it’ll take a number of years to mature in agriculture. That’s true for lot of applications of artificial intelligence.
“We’re seeing eight or nine areas of potential uses for AI in agriculture. Everything from using Watson to identify what disease you have in a field — what we’re doing today — to crop yields to price models. But I also think we’ll be applying a lot more to navigation of equipment in the field and minimizing the amount of land you’re driving on to reduce (soil) compaction. There will be predictive maintenance on farm equipment and utilization of labor in the field. All those things will be in play the next few years.
“Virtually every aspect of agriculture will be impacted by artificial intelligence over the next 10 years. It will become more automated. But even in that picture of the future, there will be a need for the computer to give you the best first guess it can. Then, an agronomist or grower will still need to apply their own judgment to that guidance.”