AI brings transparency to farmland values

Sometimes the answer to a question is right in front of you, but you can’t see it. Like where is the box of tissues in the storage closet? My husband insists we are out of tissues. I insist there is a box in there, he just needs to keep looking. We bicker about it some more, shouting from opposite sides of the house, until I finally join him at the closet and find it myself…in seconds.  

Artificial intelligence (AI) is on its way to helping humans efficiently sort through the “storage closet” of farmland data to answer complex questions, such as how much a field is worth and who is likely to buy it. Yet, some say it has a long way to go. 

Building on the data 

In some states, farmland sales data can be hard to come by. For others, it is public information that can be requested at county offices, but even then, the information has not been aggregated for easy consumption. Today databases can help farmers and ag real estate professionals investigate not only sales prices but also specifications such as soil type and crop history. Such databases are in their infancy but are the foundation AI needs before it can provide meaningful insights. 

“This is why I’ve been collecting every bit of data I can,” says David Whitaker, owner of Whitaker Marketing Group. “Eventually, it’s going to be very important.” 

Whitaker is an ag real estate agent in Iowa, where sales data is reported in every county. Existing AI technology can efficiently sort through and analyze the sales data he has curated over the years. He can prompt the AI platform to find comparable properties within his dataset that can help him estimate for clients what a piece of land may be worth. 

“If I can give them a more accurate picture with data…then it’s not just my opinion,” he says. “It’s my opinion based on data…. It gives them security to know they’re not overpaying, but they’re not underpaying for it either.” 

He can also use AI to help him find potential buyers by prompting the platform to search the data for individuals who have bought in a certain county in the past. 

Whitaker may be ahead of the curve, however. Steve Bruere, president of Peoples Co., says the ag real estate market as a whole has a lot of undigitized records.  

“We’re doing our part to digitize our transactions so that you can, in fact, leverage some of these tools down the road,” he says. “And if you don’t go to that effort to digitize your internal workflows and your internal records, AI’s cool, but you can only do so much with it.” 

Today he says Peoples Co. has “dabbled” with AI for marketing listed parcels.  

He says when an AI platform such as ChatGPT is given details about a listing, it can generate language for a marketing brochure. His firm has also used it to generate boilerplate legal documents that an in-house lawyer can work from for new or unique situations. 

“Sometimes it’s really good; sometimes it’s not,” he says. “Sometimes it needs some help; sometimes it doesn’t.” 

‘Not even close’ 

While AI may be able to help professionals like Whitaker draw conclusions concerning land values when given enough data, Carter Malloy, founder and CEO of Acres, says it will be a while before an AI platform can accurately provide its own land value estimates. 

“We and plenty of others have thrown advanced statistical models, machine learning algorithms, at the problems of land valuation and appraisals, and it’s just not even close,” he says. “It’s important to enable the true professionals in the industry, the brokers and managers, the appraisers, and farmers out there with the tools to be able to draw their own inferences.” 

His company’s database, Acres.com, has information on 150 million U.S. land parcels. The site has three levels of access: free, premium, and enterprise. Each tier has more insights for users to sift through, such as sale prices, cash rental rates, county yields, crop history, historical satellite imagery, and more. 

“I do think that AI will play a much broader role in the land industry, but it will first require a platform that pulls together sufficient high-quality data and users,” Malloy says. “So that’s where we are really intensely focused, on equipping the industry with data and with transparency.” 

But even when the day comes that AI can provide fair estimates based on quantitative data, Malloy and Whitaker agree accurately assessing the many qualitative variables that impact land values, such as the makeup of a local buyer pool, will still require local expertise.

Just the beginning 

However, Farmers Business Network may beg to differ as it is close to releasing its first AI farmland valuation tool.

Earlier this year, Farmers Business Network (FBN) launched AcreVision, a database similar to Acres.com but exclusively for FBN members. 

As of this writing, Daniel English, general manager of FBN Finance, says FBN has rolled out an update to the database in six states that includes a simple farmland valuation based on recent adjacent sales and manual staff adjustments. He says the next generation of the tool should be out this fall and will consider more variables and use AI to create valuations. 

English says FBN hopes someday to offer an interactive AI chat feature where customers could not only receive an AI-generated estimation of value for a parcel but also be preapproved for a loan based on their interest in the property. 

“We used to farm with horses and plows,” Whitaker says. “Today, we have self-driving tractors. If that’s the case, what does the next 20 years look like? So, stay open to new technology.”