2025 | AI-assisted farm identity and regulatory information
Government authorities maintain extensive records describing agricultural land, water use, irrigation permits, resource consents and other licensed activities.
These records contain important evidence about farming enterprises and their activities, but are often held in separate databases, documents and regulatory systems.
This project investigated whether artificial intelligence could accurately read, interpret and extract information from these existing records, reducing the time and cost required to establish trusted digital farm identities; recreate farm permits and consents accurately; resolve farm permits and consents to a farm digital identity; and enable the automation of farm consent and permit renewals.
Establishing a reliable digital identity for a farming enterprise requires more than identifying its name or geographical location.
A farm may occupy multiple land parcels, undertake several agricultural activities and hold licences or consents issued by different authorities.
These regulatory records may describe irrigation rights, water allocations, operating conditions and other authorised activities. However, the records are not necessarily organised around a unique farming enterprise.
Manually identifying, interpreting and linking these records to individual farms is time-consuming and expensive, creating a significant barrier to establishing digital farm identities at scale.
The project examined whether AI-assisted document processing could extract reliable information from irrigation permits, resource consents and other licences issued by local government authorities.
The research focused on determining whether AI could:
The approach combined AI-assisted information extraction with authoritative data sources and connected information models.
A central research question was whether AI could extract information from regulatory records with sufficient accuracy and consistency to support trusted agricultural information systems.
The project also examined how extracted information could be linked to farm identities and represented as relationships between farming enterprises, licensed activities, land parcels and other attributes.
This required distinguishing between information explicitly recorded by an issuing authority and information inferred through analysis, with appropriate verification of the resulting relationships.
The research explored whether these processes could provide a more efficient and scalable alternative to manually reviewing and connecting regulatory records for each farm.
The current rate of issuing a farm permit in the province is governed by available human labour. Eact permit or consent is taking approximately one month of labour per person to investigate, assemble and issue.
By investigating AI-assisted extraction of authoritative regulatory information, the work explored a pathway for significantly reducing the cost of establishing digital farm identities.
It also extended the concept of farm identity beyond a name and location to include relationships with licensed activities, regulatory obligations and other attributes of the farming enterprise.
This provided an important foundation for the development of connected agricultural intelligence, where information about farms can be interpreted in its wider operational, geographical and regulatory context.
The research contributed to the progression from digital farm identity and verified credentials towards knowledge graphs, contextual information and AI-assisted agricultural analysis.
These capabilities provided a foundation for subsequent work connecting agricultural enterprises, land, water and environmental information across freshwater catchments.
TANZ does not retain ownership of intellectual property developed through this project. Project intellectual property is either made available under open-source arrangements or remains the property of participating TANZ members, subject to the relevant project agreements.
← Previous project: Digital Farm Wallet
Return to all projects & case studies
