AI tools, models, data and services aren't the same thing.
Learn the difference between what you can use, what you can build with and where you may need expert help.
Do you want to use something, build with something or get help?
Products & platforms
Products and platforms are generally the most approachable resources in the directory. They are designed for an end user to accomplish a task without building the underlying AI system.
They might help users:
Analyze imagery
Answer agronomic questions
Automate reporting
Monitor operations
Is there already a tool that can help me do this?
Models
A model is an AI system trained to recognize patterns, generate outputs, make predictions or perform a specific kind of analysis.
Some models are broad and general purpose. Others are trained for highly specific tasks. Agriculture and food applications may include:
Plant or disease identification
Image segmentation
Yield prediction
Weather forecasting
Document analysis
Classification
A model is usually one piece of a larger solution.
+ software / integration
= Something useful in a workflow
Is there already a model we could use instead of building one from scratch?
A model is usually not a complete end-user product. It often needs data, software, infrastructure or integration around it before it becomes useful.
Datasets
Datasets are organized collections of information that can be used to train, evaluate or operate AI systems.
They might contain:
Images
Text
Sensor readings
Weather records
Crop or livestock observations
Genomic information
Market or geospatial data
Is there data available for this problem?
What could we use to train or evaluate a model?
APIs & developer tools
An API allows one software system to request information or functionality from another.
Instead of asking someone to manually use a separate AI application, a technical team can use an API or developer tool to bring that capability into an existing system or workflow.
For example, an API might allow software to:
Send an image for analysis
Request a forecast
Retrieve market information
Generate or summarize text
Access a model
Can we integrate this capability into something we already use?
APIs and developer tools generally require technical implementation.
Agents & automation
AI agents and automation tools can do more than produce a single answer. Depending on the system, they may interpret a request, gather information, use other tools and complete multiple steps.
For example, an agent might:
Monitor incoming information
Identify what changed
Summarize the change
Update another system
Prepare a report or next action
Could AI handle more of this workflow instead of helping with only one step?
"Agent" is used broadly. Some agents are sophisticated systems with significant autonomy; others are simpler automated workflows with AI embedded in them.
Simulation & digital twins
Simulation tools create digital representations of environments, processes or systems so users can explore what may happen under different conditions.
A digital twin generally represents a specific real-world asset, operation or environment and may update that representation using real or near-real-time information. Agriculture and food applications might include:
Crop or field simulation
Equipment simulation
Environmental modeling
Production systems
Supply chains
Can we model this before changing it in the real world?
Services & solution providers
Sometimes the problem is not finding a tool. It is figuring out what to do, what to choose or how to make it work.
Service providers may help organizations:
Identify AI opportunities
Assess readiness
Select technology
Prepare data
Build or integrate systems
Redesign workflows
Train teams and support adoption
Service providers help when the challenge is not just finding technology, but deciding what to do, what to choose and how to implement it.
We know what we want to accomplish. Who can help us figure out how to do it?
Some providers specialize in AI technology. Others combine AI with agriculture, food, business or domain expertise.
One resource may fit more than one category
AI resource types often overlap. A platform may expose an API. A model may sit inside a ready-to-use application. A service provider may offer proprietary technology. A dataset may be part of a commercial platform.
That is why a single directory listing may appear in more than one category.
The goal is not to force every resource into one box. It is to help you understand what you are looking at and how you might use it.
Start with the problem, not the technology.
Not sure whether you need a dataset, model, API or service? That’s fine. Start with the work you’re trying to get done and search the directory to see what exists.
Need help figuring out what fits?
9 North helps agriculture and food teams identify where AI can create business value, evaluate the right approaches and turn promising opportunities into practical systems.