How AI is Reshaping Agricultural Innovation

AI and data science support selected Bayer research, development and agronomic decision-support activities. We use these tools to evaluate seed and trait candidates, assess potential crop protection solutions and develop agronomic recommendations. Where available, recommendations may use farm-specific data to support decisions such as planting and irrigation; results depend on factors including local conditions, data quality and implementation.

AI is already supporting specific applications across product development, seed production, supply-chain planning and on-farm decision support. The examples below describe how Bayer is applying these tools and distinguish current uses, observed results and future objectives.

Long before the recent expansion of generative AI, Bayer had developed a large precision breeding platform. Combining extensive breeding data with genomics and artificial intelligence is helping researchers evaluate potential genetic combinations and advance selected breeding programs more efficiently.

Our breeders use machine learning and digital twin technology to evaluate potential genetic combinations and model plant performance across a range of climatic and soil conditions. These models help inform the development of crops intended to address identified grower needs; field performance varies by product, environment and management practices.

In selected applications, steps that historically took five or six years can now be completed in approximately four months using Bayer’s AI-enabled precision breeding workflow. Bayer’s stated objective is to use these capabilities to increase the rate of genetic innovation by 2030; progress should be assessed against the program’s defined metrics and reporting. 

The E.L.Y. expert GenAI system is designed to support agronomists in selected aspects of their daily work. Bayer has trained the underlying large language model using internal data, insights from trials in its testing network and aggregated expertise from Bayer agronomists.

The system is designed to respond to questions about agronomy, farm management and Bayer agricultural products using natural-language prompts. In a United States pilot, farmer-facing employees evaluated its outputs and used the tool to support selected tasks. Internal testing indicated faster access to relevant information and stronger performance on defined agricultural use cases than the general-purpose models included in the comparison; results depend on the question, available data and validation by qualified personnel.

Bayer internal teams are exploring this technology and testing multiple scenarios where large language model capabilities could add value and increase productivity through the ability to interact with farm data in a natural way.

We also use AI to evaluate crop protection candidates against defined performance and safety criteria, including environmental considerations relevant to product development.

Our scientists have developed a process for identifying and evaluating crop protection molecules called CropKey. AI-supported virtual screening allows researchers to assess more potential candidates than conventional screening alone. Predictive models evaluate defined characteristics, including expected behavior in soil and potential effects on non-target organisms, before selected candidates advance to further laboratory, field, safety and regulatory assessment. Bayer is applying this approach to pipeline candidates, including a novel fungicide for fruits and vegetables and icafolin, a broadacre herbicide molecule; any future market introduction is subject to successful development and applicable regulatory approvals.

Beyond the farm, Bayer is exploring how AI and intermediate oilseed crops could support renewable-fuel value chains. Bayer’s stated objective is to develop crop and agronomic solutions for renewable diesel and sustainable aviation fuel feedstocks while evaluating effects on land use, food production, farm viability and greenhouse gas emissions within the relevant value chain.

AI algorithms can support variety selection, crop-rotation planning, logistics and processing for intermediate oilseed crops. Any greenhouse gas benefit must be assessed using defined lifecycle boundaries, baselines and methodologies, and results may vary by crop, geography, farming system and end use. Bayer is also evaluating how precision breeding, crop protection and data tools can support performance across this value chain. 

Bayer has announced plans to launch newgold® seed in the United States and Canada in 2026. It is Bayer’s first multi-crop seed brand developed for intermediate oilseed crops, including  camelina and winter canola, intended for use as renewable-fuel feedstocks. Any carbon-intensity claim should be supported by the applicable lifecycle methodology, geography and production data.

Bayer uses advanced data science to support seed-production decisions for new hybrids and varieties. The models inform choices such as field placement, planting density, planting timing, irrigation timing and harvest timing. In the production fields included in Bayer’s assessment, use of these models was associated with a productivity increase of more than 30%. The result applies to the assessed operations and period; resource-use outcomes depend on local conditions and how recommendations are implemented.

Given the volatility and complexity of our crop protection supply chain, we use machine learning to identify potential vulnerabilities, model their possible effects and evaluate response options under different demand scenarios. We are also assessing how these models could support planning for the active ingredient supply chain.

Now let’s look at how AI can be deployed by farmers.

FieldView is Bayer's digital farming platform and manages more than 220 million acres in over 20 countries. In supported markets, its AI-enabled features provide decision-support information intended to help growers plan and manage selected field activities. Outcomes vary with available data, local conditions and farmer implementation.

Here’s how it works. First, in applicable regions, the app offers seed scripts, an advanced feature that uses machine learning with satellite images or a farmer's crop yield data to divide a field into different management zones. Next, predictive models suggest how densely to plant the crop in certain areas based on the grower’s specific yield or profit goals.

Through these advancements on and off the farm, Bayer is improving efficiency and productivity while paving the way for a more sustainable and responsive agriculture industry.