Leidos launches HeadWay Dispatch to set limits on AI agents from different vendors

US technology company Leidos introduced HeadWay Dispatch on November 4, a platform that helps organisations govern AI agents from different vendors: what they may do on their own, what needs human approval, and what is blocked, with tamper-evident records of their decisions.
US technology company Leidos announced on November 4 the launch of HeadWay Dispatch, a platform that helps organisations govern AI agents from different vendors. As agents take on more responsibility and act with limited human involvement, Leidos said organisations need to set clear operating limits while keeping people in charge of critical decisions.
Three parts of the control
The platform rests on three parts: first, visibility — knowing which agents are running, what they can do, and which systems they can reach; second, control — setting and enforcing boundaries that decide which actions agents can take independently, which require human approval, and which are blocked; and third, accountability — tamper-evident records of agents' decisions so they can be reviewed afterwards. Steve Hull, president of Leidos Digital, said the platform is designed to give organisations oversight without requiring them to start over with their existing AI investments.
Autonomy set by risk
Under the platform's factsheet, autonomy is set by task complexity, consequence severity and operational context, with each action assigned to one of four outcomes: block, confirm before acting, notify and proceed, or execute automatically. Low-risk, reversible work therefore does not carry the same approval burden as a consequential action. Supporting features include Flight Recorder, which provides tamper-evident reasoning provenance and a full chain of custody; a Kill Switch that contains activity at tool boundaries regardless of cloud connectivity; and Circuit Breakers that throttle battery, bandwidth and compute quotas to constrain runaway agents. Password and API-key requests are routed through Dispatch, so agents do not store those credentials directly.
Federal policy context, and what is still to come
Leidos said the platform can assist federal agencies in addressing two US Office of Management and Budget memoranda issued on April 3, 2025: M-25-21 on AI governance and human oversight, and M-25-22 on competition and portability in AI purchasing. Dispatch is delivered as a software licence with implementation services, runs on infrastructure selected by the customer, and is modular enough to cover existing investments, custom-built agents and third-party task agents. Two capabilities remain in development: Experience, which tailors what each operator sees by role and mission need, and Intelligence, which coordinates agents from different vendors under a common policy; Leidos also plans for the controls to keep working when network connections degrade or disappear.
The cover image is an illustration of AI agents, not a photograph of the Leidos product.
N環球 
💬 Comments
No comments yet. Be the first!