A budget request describes what a team wants to buy. A production inventory describes what it has put to work. A new security survey makes the distance between those records visible.

In “The New Security Agenda,” published October 28, Steph Zhang and Owen Colby report a survey of 100 chief information security officers at companies with more than 1,000 employees. Among respondents, 87% expect cybersecurity spending to increase in 2027. Only 18% report production use of AI security-operations tools. AI offensive security and AI exposure-and-remediation each register 13%. Another 45% plan a purchase or upgrade for internal AI agent security within twelve months.

Those percentages answer different questions across different categories. They cannot be subtracted to calculate a failed-deployment rate. They do show why a planned purchase should retain its own status until the organization has evidence of a working capability.

The human-work forecast is also more complicated than a smaller payroll. Only 4% of these respondents expect security headcount to fall; nearly half expect it to grow. That is a reported expectation about staffing, not a measured effect of deploying an AI product.

The authors write from an investor’s perspective and discuss companies in which a16z has invested. Their survey is useful evidence of this sample’s intentions and reported deployments. The article does not establish a representative adoption rate for all enterprises, causal productivity gains or independent proof that the featured products work. Keep those limits beside the numbers when using them in a budget discussion.

For an internal review, put each proposed capability beside a specific operating gap: an investigation waiting for evidence, a repair waiting for an owner or a completed change waiting for verification. Name the person responsible for taking the result into service, and measure how much work still returns to that person.

Then follow one representative case from alert to verified outcome. A growing findings queue may require more review capacity. A completed repair with evidence gives the spending request something firmer to stand on.