Getting value from AI in Agriculture was the theme of a presentation by Matthew Smith, who has just joined Agrimetrics to head up product development.
The event was organised by the Institute for Agricultural Management and hosted at Lincoln Institute for Agri-Food Technology (LIAT) as part of Agri-Tech Week 2019.
AI in Agriculture
To implement AI in the field needs more than connectivity, it also requires a common framework for relating one type of data, eg crop data, with another such as soil data. Matthew explained how work by Agrimetrics on digitising field boundaries is helping to provide this framework and this will be super helpful for ensuring data is meaningful.
It will allow, for example, a high accuracy soil moisture map to be created with a few soil sensors on the ground combined with aerial imagery. AI would be able to find and link the data and provide meaningful insights.
Examples are already emerging, such as the work with BASF to develop a planning tool that uses information about weather forecasts and soil moisture to determine when to spray to avoid runoff.
IoT platform FarmBeats uses white space
Matthew, who until recently was director of business development at Microsoft, described how Microsoft’s Azure platform is being used in agriculture to facilitate the application of AI. For example, FarmBeats is an Internet of Things platform which uses unlicensed TV white spaces — the radio frequencies allocated to broadcasting services — to establish a high-bandwidth link from a farmer’s home internet connection to a base station. Sensors and drones then connect to the base station, which draws power from a battery-backed solar panel pack.
Other applications include smart tools to improve precision in time (eg crop emergence), in detail (eg identifying an intruder in a cow shed), or within forecasting (eg weather or harvest date).
Matthew also touched on robotics, for example robotic milkers reduce stress and also provide information that can be used by the farmer to improve decision making.



AI for insights
Another area where there are tangible benefits is in traceability – collation of data at all points in the value chain can provide evidence, for example of the use of medication in livestock rearing or tracing a product back to its source. Matthew predicted that traceability is going to be as regular a part of business as accounting is now.
Matthew also predicted that there will be more organisations basing advice on data gained from generators. In South America financial people are also the agronomists and provide guidance on the data they collect from farms for accounts.
However, he did raise a number of issues:
- Need more investment in validation of tech to ensure that its useful on farm
- Data ownership is a big challenge – there is a need to provide value to the data generator. There are different models for doing this, for example the Dutch producers work together and pool their data to make the country more competitive at a global level
- To gain value from your own data you really need someone to do the analysis for you and this can be a good investment in the longer term
- EU has rules on data sharing to protect data for generators in agriculture – just like approving cookies on the internet we may have to give permission for data sharing in future.


