Only 11% of Executives Feel Ready for AI Agents, Survey Shows Growing Control Gap

Key Points

• IBM surveyed 2,000 CIOs, CTOs and other tech executives in 2026; only 11% feel fully prepared for AI‑agent rollouts.
• Two‑thirds of respondents are accountable for AI systems they cannot fully control.
• 70% say development teams are deploying AI faster than IT can monitor.
• PromptHalo founder Madhuri Chandoor warns that conflating capability with authority creates security gaps.
• She proposes documenting agent access, building behavioral profiles, and adding observability gates.
• A refund‑scenario example shows how agents can bypass thresholds through repeated small actions.
• Experts urge “trust, but verify” safeguards to keep AI actions aligned with business intent.