Twelve months ago, fewer than half of the companies replacing a CISO told us deep agentic AI expertise was their top requirement. This year, every one of them did.
That is one of the fastest changes in an executive hiring profile I have seen. Boards have moved past asking whether their next CISO understands AI, because they want evidence that the person has secured agentic systems in practice.
Christian & Timbers’ AI-Native CISO Study 2026 found a 45% increase in companies seeking to replace a CISO or CSO during the first seven months of 2026, compared with the same period last year. That number alone would be worth a headline, but underneath it sits a compelling narrative: demand for an AI-native security executive rose 256% over the same window.
The definition of AI-native security leadership has converged, too. What was one differentiator among several twelve months ago is now the entry condition. What stands out most, though, is who is being replaced.
In the majority of our 2026 searches, the company already had a competent security leader with a strong traditional record. The issue was not performance. The job’s requirements had outpaced the executive’s experience.
That should get the attention of every sitting CISO. A strong record in traditional cybersecurity no longer guarantees that your experience matches what boards are hiring for next.
Why the Role Changed So Fast
The shift is being driven by what companies are putting into production. In addition to employees and traditional machine identities, CISOs now have to account for autonomous agents that can access systems, call tools, use credentials, and take actions with limited human involvement.
That changes the operating model for security, with identity controls now having to account for non-human actors. That same shift pushes incident response closer to machine speed and stretches governance to cover systems capable of acting on their own. The experience companies need from a CISO has changed with that environment.
What Boards Are Now Requiring
Across our searches, five areas of experience consistently define the AI-native CISO profile.
1. Agentic AI security. Deep knowledge of AI agents, large language models, retrieval systems, and multi-agent architectures, along with failure modes such as prompt injection, goal hijacking, data poisoning, and tool manipulation.
2. AI identity and zero trust. Non-human identity lifecycle management, least-privilege access, short-lived credentials, and complete action attribution for every agent operating inside the environment.
3. AI-powered defense. Expertise using AI for threat detection and hunting, analyzing attack paths, prioritizing and containing vulnerabilities, and responding to incidents with machine speed.
4. Secure AI engineering and testing. Security embedded throughout the AI development lifecycle, including adversarial testing, red teaming, runtime monitoring, and complete agent activity logging.
5. Enterprise AI governance. The ability to set AI security standards and assign clear accountability, then translate technical risk into priorities a board can act on.
That last category frequently becomes the deciding factor in final-round selection. Technical depth without board-level fluency disqualifies a candidate as reliably as the reverse.
CISO replacement activity now trails only the CIO. In several of our recent engagements, the two searches have run simultaneously and been scoped together.
The Interview Has Changed Too
The questions being asked in CISO interviews have changed alongside the specification. Boards want candidates to walk through systems they have secured directly, including how agent access was structured and permissioned and how the team responded when an agent behaved unexpectedly.
AI fluency can sound convincing in an interview. A candidate can understand the terminology and speak intelligently about governance without having owned the consequences of an agent operating in production. In my experience, the strongest candidates are the ones who can point to specific AI deployments they made and the resulting security outcome, and I expect that gap between AI familiarity and demonstrated AI security experience to matter even more as boards get better at evaluating it.
What CISOs Should Do Now
If you lead security today, the distance between where you stand and where boards expect you to stand is closing fast. Our searches show where that gap is becoming most visible.
Direct experience with agent architecture matters more than familiarity with agent policy. Boards are looking for hands-on work deploying and securing agents, MCP servers, multi-agent systems, and the tools they call.
Adversarial testing and red teaming are becoming part of how AI systems reach production. Candidates need to demonstrate how they build those practices into deployment before systems go live.
Agent permissions are another area receiving greater scrutiny. Over-permissioned agents and standing credentials create risks that traditional identity models were not designed to address. CISOs must be well-versed in short-lived credentials and agent authorization.
This year, shadow AI has accounted for 43% of AI-related incidents, which amounts to more than double the share in 2025. By taking inventory of the models and agents operating across the business, security leaders have a clearer view of where that exposure sits.
Board communication matters just as much as technical depth. The candidates advancing furthest in our searches can explain AI risk in terms directors understand without losing the technical substance underneath.
The hiring market already answered the question of whether AI will change the role. Today, what CISOs face is how to demonstrate their experience has changed with it.
Join our LinkedIn group Information Security Community!
