Agentic AI

Working knowledgeAgentic Protocols and Infrastructure

Also called: Agentic

Agentic AI is the design approach in which software plans a sequence of steps, uses external software tools and takes autonomous action to execute multi-step goals, rather than generating a single response to a single prompt. It describes how a system behaves, not a product category: the same underlying model can be used agentically or not, depending on what it is permitted to plan and do. The working unit built this way is an AI agent.

In practice

"Agentic" on a vendor slide names an architecture, not a level of capability. The two questions that separate the label from the substance are what the system is allowed to do without asking, and what happens when one step in its plan fails. See AI Agent for the unit itself.

Not sure where your organisation stands?

Take the free AI-readiness diagnostic.

Start the diagnostic