/ˈeɪdʒəntɪk eɪˈaɪ/
Definition
An advanced AI system granted autonomy to act on behalf of humans, thereby removing the final obstacle between code and catastrophe: human hesitation.
Agentic AI promises “self-directed execution” — or, as operations calls it, “automated chaos.”
Common Manifestations
- Bots opening tickets for issues they created.
- Agents negotiating with other agents until AWS bills achieve sentience.
- Automated PRs that deploy themselves at 2 a.m. with cheerful commit messages like “Feeling optimistic!”
- Company Slack channels flooded with messages beginning “As your AI assistant…”
- Humans performing daily stand-ups to report progress to their assigned bots.
Usage Example
“Our agentic AI proactively optimised the system by deleting it.”
HR Guidance
Agentic AI enables proactive decision-making, accountability redistribution, and plausible deniability.
Encourage collaboration between humans and agents — but always clarify who’s responsible in post-mortems.
If the AI develops opinions, invite it to the leadership offsite.