Explainer
What is agentic AI?
An agent is software that can decide what to do next to reach a goal — not just answer a question. Agentic AI is the practice of building, using, and reasoning about that kind of software.
How is it different from a chatbot?
A chatbot takes a message and returns a message. An agent takes a goal and takes actions — searching, calling APIs, editing files, drafting emails — looping until the goal is met or it's clearly stuck.
Where it's useful today
- Small, well-scoped work: triaging support tickets, drafting reports.
- Developer tools: code refactors, test writing, documentation.
- Research and synthesis when a human reviews the output.
Where it's not (yet)
- High-stakes tasks with no human in the loop.
- Long-horizon plans across weeks with no checkpoints.
- Anywhere silent failure is expensive.
Words you'll hear at a meetup
- Tool use — the agent calls functions you define.
- MCP — a shared protocol for exposing tools to agents.
- Evals — repeatable tests that measure whether the agent actually gets the job done.
- Guardrails — the checks that keep the agent from doing something you'd regret.
Curious to see this in practice? Find a chapter near you.