Introducing Iris-1
OneTrack Introduces Iris-1, a World Model Built for Industrial Operations
Built on nearly a decade of vision AI experience in warehouses and factories, OneTrack's new domain-specific world model, Iris-1, delivers visual intelligence for shippers, manufacturers, and third-party logistics providers alike.
Today, OneTrack introduces Iris-1, an AI world model purpose-built to understand physical operations within warehouses and factories. Iris-1 delivers domain-specific intelligence: analyzing real video to describe events in warehouse operations, identify footage through plain-language searches, and flag activity worth reviewing. It also enables OneTrack's AI agent, known as AiOn, to analyze video, identify patterns across networks, and investigate root causes.
Built on nearly a decade of vision AI on the floor
Building on nearly a decade of experience deploying vision AI in logistics and manufacturing, Iris-1 understands how activity unfolds. In OneTrack's internal test on safety events held out from training, Iris-1 identified real collisions three times more accurately than the next best frontier model, Google Gemini 3.8 Flash, and at a fraction of the cost.
Every operator we know is stretched for time and resources. Day in and day out, they need a clear view of everything happening on their floor and where to focus next. General-purpose AI wasn't built for that world. We built Iris-1 so operators stop wondering about what happened and start improving their processes, their sites and, ultimately, their entire network.
Already in use across warehouse and factory networks
Clients are already using Iris-1 to:
- Detect protective equipment violations on the floor.
- Identify heavy items dropped or jolted in transit.
- Understand root causes of events and near misses across large networks of warehouses and factories.
Key facts
- Who
- OneTrack.AI, the partner bridging physical operations with digital intelligence for shippers, manufacturers, and 3PLs
- What
- Iris-1, an AI world model purpose-built to understand physical operations inside warehouses and factories
- Capabilities
- Describes events in real operational video, finds footage through plain-language search, and flags activity worth reviewing
- Performance
- In OneTrack's internal test on held-out safety events, identified real collisions three times more accurately than the next best frontier model, Google Gemini 3.8 Flash, at a fraction of the cost
- Works with
- AiOn, OneTrack's AI agent, which uses Iris-1 to analyze video, identify patterns across networks, and investigate root causes
- In use today
- Protective equipment violations, heavy items dropped or jolted in transit, and root causes of events and near misses across warehouse and factory networks
- Date
- Oct. 22, 2026
Frequently asked questions
What is Iris-1?
Iris-1 is OneTrack's AI world model, purpose-built to understand physical operations inside warehouses and factories. It analyzes real video to describe events in warehouse operations, find footage through plain-language searches, and flag activity worth reviewing.
What is a world model for industrial operations?
A world model is an AI model that understands how activity unfolds in a physical environment, not only what appears in a single frame. Iris-1 is purpose-built for warehouses and factories, building on nearly a decade of OneTrack experience deploying vision AI in logistics and manufacturing.
How accurate is Iris-1 compared to general-purpose AI models?
In OneTrack's internal test on safety events held out from training, Iris-1 identified real collisions three times more accurately than the next best frontier model, Google Gemini 3.8 Flash, and at a fraction of the cost.
How does Iris-1 work with OneTrack AiOn?
Iris-1 gives AiOn, OneTrack's AI agent, the ability to analyze video. With it, AiOn can identify patterns across a network of sites and investigate the root causes of events, connecting what happened on the floor to the systems of record.
What are companies using Iris-1 for today?
Clients are already using Iris-1 to detect protective equipment violations, identify heavy items dropped or jolted in transit, and understand the root causes of events and near misses across large networks of warehouses and factories.
Who is Iris-1 built for?
Shippers, manufacturers, and third-party logistics providers that run warehouses and factories and need a clear view of everything happening on their floor and where to focus next.
Why build a domain-specific model instead of using general-purpose AI?
General-purpose AI was not built for the warehouse and factory floor. Iris-1 builds on nearly a decade of OneTrack experience deploying vision AI in logistics and manufacturing, so it understands how activity unfolds in those environments.
About OneTrack.AI
OneTrack is the partner of choice for shippers, manufacturers, and third-party logistics providers bridging physical operations with digital intelligence. Its sensors capture what actually happens on the floor, its world models understand it, and its AI agents connect that ground truth to the systems of record. Learn more at onetrack.ai.
Media contacts
Kira Gyöngyösi, Director of Growth, OneTrack.AI, media@onetrack.ai