Gen AI Used In NZ Agri Research

Managing a website which acts as a library for 30 years’ worth of research is a constant challenge for the Foundation for Arable Research’s (FAR) communication manager Anna Heslop.

“FAR is a research, development and extension organisation funded by New Zealand arable farmers. Our website contains thousands of research reports and extension documents relevant to growing and harvesting of over 40 different crops. The crops themselves range from cereals like wheat and oats to vegetable seeds like clover or radish; while topics of interest include everything from crop establishment, nutrition and agrichemical use, to soil quality and greenhouse gas emissions.“

The site is a goldmine of information, but a labyrinth to negotiate, especially for growers who are often looking for very specific data, and don’t have time to find and read through wordy reports. No matter how we tweak search functions or tags, website searches still bring up a list of 50 or more relevant documents rather than seasonal information for a specific crop and problem.”

Anna says that when Custom D, FAR’s web solution partners, suggested their Gen AI product ‘Caitlyn’ as a possible solution, she was cautious, but interested.

Caitlyn harnesses the power of AWS Bedrock and advanced Large Language Models (LLMs) to deliver a natural, conversational experience while keeping your sensitive data secure and private.

“Apart from a few play sessions on Chat GPT, I knew very little about Gen AI or how it could be harnessed. However, we’re an R&D organisation so why wouldn’t we embrace a new technology? After the first demo, we were sold on the concept and that’s how we got involved in the Caitlyn Project.”

“AWS Bedrock provides a game-changing foundation for the Caitlyn platform which gives organisations the flexibility to swap out different LLMs.” says Custom D director Josh Smith “This means FAR and other Caitlyn users can remain flexible and not tied to any one AI model. This is important in today’s landscape where AI models are rapidly evolving and leapfrogging each other on a monthly basis. Bedrock essentially allows FAR to stay at the cutting edge of AI with minimal effort.”

“We’re investing in Caitlyn to help our growers access the information they want quickly and easily. We anticipate that this will lead to increased engagement and improved uptake of the information in our research and extension documents. That in turn will increase the return on research investment and make life easier for our levy payers.

“Our goal is to transform growers’ perception of the website from having a body of documents that were difficult to navigate” to “the go-to site for all things arable”.

© Scoop Media