Phasis World API equips production agents with ontologies and reality layers for precise orientation
The startup’s platform offers production agents structured orientation data, including current states and uncertainties, to improve decision-making accuracy.
Key points
- Phasis launches a World API that provides AI agents with evidence-backed context and reality layers for better orientation.
- The platform uses graph-based ontologies to map culture, values, and subcultures for agent comprehension.
- A primary /orient endpoint returns relevant World frames, including material changes and supporting evidence.
Phasis Inc. has launched a new platform designed to orient artificial intelligence agents using World APIs and reality layers. According to a report by SiliconANGLE, the company aims to help agents comprehend complex domains by providing structured, evidence-backed context. This allows automated systems to make decisions based on fast-changing real-world conditions rather than static training data.
The Phasis World API is built for production environments, offering a maintained state of specific domains. Brandon Farley, co-founder and CEO of Phasis, explained that while reasoning capabilities in AI are scaling rapidly, expertise and world understanding are not advancing at the same pace. He stated that the World API represents reality across many domains, allowing agents to orient themselves against accurate, current information.
Context and Background
The development of this platform originated from the founders’ work in measuring culture through graph structures. According to SiliconANGLE, Phasis initially sought to understand relations between people, values, beliefs, and subcultures. The goal was to map these cultural elements in ways that agents could comprehend and utilize for decision-making.
Farley noted that the processes used for cleaning, verification, and creating ontologies in cultural mapping applied well to other domains. He told John Furrier on theCUBE + NYSE Wired: AI Luminaries that this realization led to the creation of the broader World API. The system provides primitives required to build reality layers and expertise for enterprise agents.
Who Is Affected
The primary users of the Phasis World API are enterprises deploying production agents. The platform is designed to integrate into existing systems, providing agents with necessary orientation before work begins. According to Farley, companies that use these durable World models will stay current as circumstances change, rather than piecing together context every time.
The primary endpoint, called /orient, allows users to send tasks to the system. Phasis returns relevant World frames that include the current state, material changes, supporting evidence, and uncertainties. Farley suggested that the compounding effect of agents oriented against reality will set apart companies that adopt this approach from those that do not.
What should you do?
Organizations deploying AI agents should evaluate whether their current systems provide sufficient real-world context for decision-making. Security and technology professionals should review how agents handle uncertainty and changing circumstances in production environments. Teams should verify that any orientation data used by agents is evidence-backed and regularly maintained.
What to do and how to stay safe: Phasis
- Audit current AI agent workflows to identify gaps in real-time context or world knowledge.
- Verify that orientation data sources are evidence-backed and subject to regular verification processes.
- Monitor agent decisions for reliance on outdated or static information in dynamic environments.
- Ensure that any integrated reality layers provide clear indicators of uncertainty and material changes.
Step-by-step guide: Prompt Injection Warning Signs: How to Spot AI Manipulation
General security guidance from the Threat Times newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
What is the primary function of the Phasis World API?
The World API provides AI agents with evidence-backed context and reality layers to orient them for production tasks.
How does Phasis map cultural data for AI agents?
The company uses graph structures to map relations between people, values, beliefs, and subcultures for agent comprehension.
What information does the /orient endpoint return?
The /orient endpoint returns relevant World frames including current state, material changes, supporting evidence, and uncertainties.
