Agentic AI

/
Glossary
/
Agentic AI

Agentic AI refers to artificial intelligence systems that don't just answer questions but take goal-directed actions on a user's behalf, perceiving an environment, deciding what to do, and executing multi-step workflows with limited human intervention.

The shift from generative AI to agentic AI is a shift in what the system is asked to do. Generative AI produces an output (a sentence, an image, a block of code) in response to a prompt. Agentic AI is given a goal and the latitude to figure out how to reach it. That means deciding what information to gather, what tools to use, what intermediate steps to take, and when to stop. A well-designed agentic system can investigate, reason across multiple sources, and either complete a task autonomously or hand it off to a human with a structured summary of what it found.

For production operations, this distinction matters more than it does for most domains. An LLM that summarizes incident data after a human assembles the context is generative. A system that detects an alert, traverses the dependency graph, evaluates ranked hypotheses against live evidence, and either remediates within bounded authority or pages the right responder with the reasoning already complete: that is agentic.

Building reliable agentic AI for production is harder than calling an LLM in a loop. It requires a world model, persistent memory, the ability to run parallel investigations, and explicit boundaries on what actions the agent can take autonomously versus what requires human approval. The category of AI SRE sits at the intersection of agentic AI and site reliability engineering, and Traversal is the first AI SRE validated at Fortune 100 scale.

Related

AI SRE

An AI SRE (AI Site Reliability Engineer) is an autonomous agentic system that performs causal investigation, root cause analysis, and remediation across production environments, operating as a continuously available teammate alongside human reliability engineers.

Agentic Mission Control

Agentic Mission Control is Traversal's term for the operating model in which an AI SRE continuously investigates the production environment, surfaces causal hypotheses before alerts escalate, and routes responders with the reasoning already assembled, enabling a level of autonomy and expertise previously unavailable to traditional mission control teams.

Model Context Protocol (MCP)

Model Context Protocol (MCP) is an open standard that defines how AI agents connect to external tools, data sources, and services, providing a uniform interface for agents to access the systems they need to act on a user's behalf.