Silos are more than just a place to store grain on Paulman Farms. They also represent isolated pockets of yield, moisture, machine, and nutrient data that could help the Sutherland, Nebraska, operation get the most out of the nearly 10,000 acres it covers.

Roric Paulman relies on 40 different apps to control and monitor the 14 dryland and irrigated crops grown on the land. The software generates one terabyte of site-specific data every month, which equates to 75 million pages of information.

That information is key to helping him make sound management decisions. The problem is, no one has created a system that effortlessly connects the dots to give Paulman better insights. Also, if he can’t access the information from his smartphone, he’s not interested. “The apps and the information being collected have stopped being useful,” says Paulman, who farms with wife Deb and son Zachary.


Companies with a long history in agriculture have been working to build a user-friendly system that seamlessly processes and integrates data from myriad apps. Yet, data sharing and interoperability are still not easy or seamless.

These companies also face a problem of their own: How do they collaborate with others to develop a solution yet remain competitive? As agriculture’s existing players work on a resolution, outsiders like IBM are emerging with their own approach.

Launched in 2018, the Watson Decision Platform for Agriculture leverages the power of artificial intelligence (AI) to analyze silos of data and then generate evidence-based insights. Watson begins by creating a digital representation of a field. This electronic field record (EFR) includes soil, equipment, farm practice and workflow, and imagery data. It can also accept weather data from The Weather Company.

Applying AI, machine learning, and advanced analytics to the EFR, the platform highlights key factors that might affect crop yields like soil temperature, moisture levels, crop stress, pests, and diseases. Ultimately, each EFR becomes a digital twin of everything that happens on Paulman’s 113 fields. A unified dashboard lets him easily see and monitor data as well as receive alerts when critical elements like weather could affect a crop.

The difficulty with many of the decisions Paulman tries to make is that they are biologically based. “They are almost always influenced by weather we don’t yet know. Having the ability to forecast conditions has to be an integral part of any decision platform,” says Kenneth Sudduth, research agricultural engineer at USDA-ARS.

In addition, the process has to be automated from start to finish. Technologies like automatic guidance, shutoffs, and boom height control – systems that had little or no direct human control – saw fairly swift adoption because they improved the workflow without requiring operator interaction.

Today, too many applications require farmers to input information over and over again. “Every time farmers make an entry, there is a chance they’ll get it right, but there is also a chance they’ll get it wrong,” says Michael Gomes, VP business development IoT, Topcon Agriculture.

More often than not, the most common variety planted is labeled “one” because the window to get that seed in the ground is continually shrinking.

It’s a painful process, and farmers are tired of it.

If farmers can select from a pick list, Gomes says, their risk of getting it wrong is a whole lot lower than having to punch it in letter by letter or ensuring they call it the same exact thing every time.

“Only about 8% of the data being collected is actually usable,” says John Fulton, associate professor at Ohio State University.


To make the analytics better, a much cleaner data set is needed, and many believe AI can take producers there. Applying it to data provides Paulman with myriad new abilities.

From the air, he can deploy a drone to capture a field of corn and use AI visual recognition to identify crop disease or a pest infestation. From the ground, plants can be photographed up close, so Paulman can react in real time.

“Simplifying the process also enables agronomists – who currently spend 80% of their time trying to collect and analyze a farmer’s data – to make decisions with greater confidence,” says Kristen Lauria, general manager of Watson Media and Weather Solutions.

By collating and curating the data, Paulman can also identify the best practices for his irrigated acres. With an annual allotment of 13 inches of water for a corn crop that requires about 22 inches of water, he has to ensure every drop is used wisely. That means relying on technology that understands he has some soils that will take 2 inches of water per hour and others that will take ¼ inch per hour.

Because prices fluctuate constantly, Watson also offers a tool that marshals huge amounts of pricing data – from the local grain elevator to the futures markets – and recommends the best time to sell to maximize profit. It’s the type of data gathering and analysis that would be impossible without AI and analytics.


As more data flows in, the decision platform becomes a more robust solution. That’s the caveat. In order for AI to be effective, it requires a large database to draw from. Farmers not only are going to have to allow others access to their information, but also will need to share data to take advantage of digital tools.

“Although we talk about having so much data, in many cases, it’s very localized. It’s almost as though we have too much data, yet not enough data at the same time,” Sudduth says.

The key, Gomes says, is to get the right data that farmers accept, so they can then take action with confidence.

So how do you get farmers comfortable with sharing their data? Billy Tiller maintains that it has to be a producer-led initiative.

Founded in 2012, the Grower Information Services Cooperative (GiSC) is a farmer-owned data cooperative providing secure cloud storage for its farmer members. Headquartered in Lubbock, Texas, the company’s platform collects and manages multiple layers of agronomy and yield data across a variety of crops including corn, soybeans, wheat, and sorghum.

“It is time for farmers to have options that are built on objective motives, not on a reason to buy another product,” says Tiller, who is the founder and CEO of GiSC.

IBM is also a firm believer in data cooperatives. By building thousands of farmer experiences into a data set, Paulman could understand, for example, what is common among all corn growers in Nebraska that is driving yields 20% above the average compared with those who had yields 20% below average. Because he is viewing his operation from a different perspective, he can evaluate which practices are truly driving better yields and which ones are not contributing.

“Instead of relying solely on data from their own farms year after year, farmers can learn from each other as well,” Lauria says.

Access and sharing are key components to the infrastructure, because value from analytics will come from different companies, Fulton says.

Skeptical about companies with a vested interest in his data, Watson also offers the independence Paulman is looking for. “IBM is not trying to sell me more fertilizer or machines,” he says. “It’s a trust thing.”


Infrastructure is the biggest component in making digital ag a success story. According to Ag Gateway, 84% of farmers and their trusted business partners say they find it moderately or very difficult to compile and analyze the data coming from farm fields.

Established in 2005, Ag Gateway has been chipping away at the interoperability friction. Its Standardized Precision Ag Data Exchange (SPADE) project has produced the Ag Data Application Toolkit (ADAPT), which enables different software applications and hardware systems to seamlessly exchange information – with broad adoption as the end goal. To date, 26 companies have committed to ADAPT by either developing a plug-in for their file format or integrating ADAPT support into their software systems.

“We use technology wherever and however we can, because we have to get better at what we’re doing for future generations,” Paulman says. “Insight from data helps us do that.”

Until there is a single system in place that standardizes and connects the entire ecosystem, the silos will remain, and the value of data will continue to be limited for Paulman Farms.


Before farmers can gain value from their data, they have to create a foundation. John Fulton, Ohio State University, suggests farmers consider the seven points below when developing a digital strategy.

1. Identify the technologies you use as well as the data generated from those technologies.

2. Organize your stored data (e.g., year, crop, farm, field).

3. Store an original copy of your data both
on and off the farm so there is a backup.

4. Ensure data can be accessed from
any location and that offline information is updated once a connection is reestablished.

5. Collect complete and quality data so you can execute desired analyses.

6. Protect data with secure passwords.

7. Define a strategy for sharing files, which includes an easy-to-copy format both on and off the farm. Don’t share information without permission.

Source: Laurie Bedord,