AI in Agriculture Hindered by Inadequate Data

AI in Agriculture: A Conditional Promise

Artificial intelligence (AI) is redefining possibilities in the agricultural sector. However, leaders must exercise caution before making massive investments in AI without first establishing the necessary data infrastructures.

The potential applications of AI in agriculture are numerous and promising. In a sector facing challenges such as volatile fertilizer costs, unpredictable weather conditions, and narrow profit margins, AI offers attractive solutions. Studies indicate that the use of AI-based predictive models can increase crop yields by 26%, reduce water consumption by 41%, and decrease chemical usage by 33%.

However, AI providers do not always highlight a crucial aspect: these solutions are only effective if they are built on a clean and robust data foundation. At Reltio, a data management company, we have seen how essential a solid technology strategy is to leverage AI, particularly through our experience with a large agricultural distributor.

The Promises of AI Providers: Beware of Data

Discussions with AI providers in the agricultural sector often follow a well-established pattern. The promises are enticing: real-time monitoring of crop health, optimization of irrigation, and maximization of yield per hectare through AI.

However, the question of the quality and completeness of the underlying databases is rarely addressed. An inaccurate or incomplete database can lead to misleading AI results, which, while appearing authoritative, may lead to counterproductive actions.

For example, a yield forecasting model based on inconsistent historical data will produce erroneous predictions. Similarly, a precision irrigation system using fragmented sensor data risks making ineffective watering decisions, thus wasting valuable resources.

In these scenarios, AI fails because the data it relies on is insufficient to ensure reliable outcomes. In agriculture, every AI error can have serious consequences, and the risk of error is high.

Agriculture: A Complex Challenge for AI

The data landscape in a modern farm or at a large distributor serving thousands of producers is extremely complex.

Modern farms widely integrate IoT devices and automated machinery. Irrigation systems are often automated, tractors navigate autonomously, and drones capture aerial images of fields on a large scale.

However, the data generated by these machines is often disparate. When external sources such as weather forecasts, data from the U.S. Department of Agriculture, and third-party market information are added, the challenge of integrating them into a coherent whole becomes immense.

Agricultural AI must also understand complex aspects such as GPS coordinates, farm boundaries, field blocks, and soil variation within the same property. Knowing where to apply fertilizers, at what rate, and in which specific area of the farm is crucial. An AI system that treats all parts of a field as identical will produce inaccurate, even harmful, recommendations.

There is also a compliance dimension to consider, due to the chemicals and liabilities involved. AI in agriculture requires much stricter controls and governance than in other sectors. An erroneous recommendation implemented in the field can have serious consequences.

Data Preparation: An Imperative

Data preparation is what distinguishes effective AI from a "garbage in, garbage out" scenario. Essentially, being ready for AI means having a data model that accurately reflects the workings of the business.

For a company like Wilbur-Ellis, a 104-year-old family-owned agricultural distributor, this involves understanding who their customers are, which fields they cultivate, what inputs they require, where those inputs come from, what they paid last season, and how all of this connects to margins. This information must be up-to-date, consistent, and accessible across the organization, rather than siloed in separate systems.

Similarly, for the farms themselves, data preparation means having a reliable and connected picture of what is happening in each field: soil health records, input application histories, yield data from previous seasons, equipment performance, and real-time readings from irrigation system sensors.

Data governance is just as crucial as data structure. Prices change, relationships evolve, and suppliers come and go. An AI system based on data that was accurate six months ago but has not been updated will make recommendations based on a version of the business that no longer exists.

Building a Reliable Data Foundation

The good news is that data preparation is an achievable goal. It starts with a solid data model: a single, governed source of truth that connects customers, suppliers, products, prices, orders, and margins in a way that reflects the organization's operations.

Next, it is essential to establish data pipelines that are fast enough to provide insights at the moment decisions need to be made, governance frameworks that maintain the reliability of this data over time, and security controls that ensure sensitive business information is accessible to the right people under the right conditions.

This is precisely the challenge that Reltio, a SAP company, was created to solve. Reltio enables businesses to unify their fragmented data so that agents and AI systems can operate from a complete picture of the organization. Reltio builds a reliable context system, known as the contextual intelligence layer, that brings together all entities, relationships, and rules under one roof, making business data easy to access and interpret.

For Wilbur-Ellis, building this reliable data foundation has meant being able to ask more complex questions and trust the answers, which is a prerequisite for an AI system to be truly useful.

Maximizing the Value of AI in Agriculture

Before engaging in a new conversation about AI, it is crucial to ask whether the underlying database is solid enough to ensure reliable outcomes.

Agriculture has always required its leaders to make high-stakes decisions under uncertainty, and AI offers the prospect of making these decisions more quickly and with better information. This prospect is only achievable for organizations that have first done the foundational work, and the companies that will get the most out of AI are those that invest in this foundation now.