"Agentic AI" — systems that take multi-step actions, not just answer — is the dominant theme of the moment. Gartner's forecasts capture both the hype and the hangover.
The growth case
- 40% of enterprise apps will feature task-specific AI agents by the end of 2026 — up from under 5% in 2025.
- By 2028, 33% of enterprise software will include agentic AI (from <1% in 2024), and 15% of day-to-day work decisions will be made autonomously.
- Agentic AI could drive ~$450 billion in software revenue by 2035.
The reality check
- Over 40% of agentic AI projects will be cancelled by the end of 2027, Gartner predicts — due to escalating costs, unclear value and weak risk controls.
- Only 17% of organisations have actually deployed agents today, though 60%+ expect to within two years.
Agents amplify everything — including your gaps. An unreliable workflow doesn't get better when you let it act on its own; it gets faster at being wrong.
The teams that succeed start narrow, instrument heavily, and keep a human approving anything irreversible. The ones that cancel tried to make an agent autonomous before it was even reliable.
Sources
Written by ivector
Start a project →