By Saviona Pereira
Most boards are not short of security data. They are short of seeing it in time, and the Cyber Security and Resilience Bill is about to make that gap a governance failure.
Picture two board meetings in the same week. In the first, the cyber security item is a quarterly slide: a risk rating, a list of audit actions in green and amber, and the directors move on. In the second, the chair asks a single question: if a significant incident began tonight, who would decide within the hour whether we must report it, and what could we tell the regulator about our security posture yesterday?
Both organisations may own the same security tools. Only one board can see what those tools are telling it.
That distinction now matters. The Cyber Security and Resilience Bill, in its final stages in the House of Lords, widens the UK’s NIS Regulations, tightens incident reporting and links penalties to turnover. Peers sought personal liability for senior executives; the government declined, saying the requirements to follow would mandate board-level governance in line with the NCSC’s Cyber Assessment Framework. Either way, accountability has moved upstairs.
The problem: governing at meeting speed
Cyber security is often called too technical for directors. The real issue is that boards see cyber risk too late and too rarely, and three habits keep it that way.
Firstly, assurance is periodic while exposure is continuous. Committees meet quarterly, but cloud environments change daily. Attackers are speeding up too: in 2024, the NCSC assessed that AI will almost certainly increase the volume and impact of cyber attacks.
Secondly, findings stay in the language of the security team. Posture tools flag a misconfiguration or an over-privileged account, but nobody translates that into which service and which obligation are now at risk.
Lastly, the hardest decisions are left until the incident. The Bill requires a significant incident to be notified to the regulator and the NCSC within 24 hours, with a fuller report within 72. That is no time to work out who decides.
Four steps to continuous assurance
My career has given me three vantage points on this problem. I began running enterprise email and collaboration environments. As a technical lead, I supported enterprise cloud tenants, for example, with just-in-time access for privileged roles. Today, as a Technical Security Architect, I design cloud security architectures, run posture management and security monitoring, and map controls against cyber assessment frameworks.
Drawing on all three, I argue that boards should replace periodic assurance with a continuous loop: Monitor, Map, Report, Decide. The order is deliberate: exposure is captured, tied to obligations, brought to directors and owned.
Monitor. Posture management and security monitoring already assess configuration, identity, and workload exposure around the clock. The data exists. It simply needs to be used beyond the security team.
Report. Directors should see material change by exception, not a dashboard of everything: which critical services have drifted from their obligations, what is overdue, and what has been accepted.
Decide. The board funds, accepts, or escalates, and records the decision. It also settles the incident decisions in advance: who judges whether an incident is significant, who notifies the regulator, and who briefs the board.
Mapping: where AI earns its place
Of the four steps, mapping is the one I have fixed first. In my work, the gap is rarely a lack of findings; it is the distance between a finding and the obligation it affects. That is where AI can help, by correlating alerts, drafting control mappings, and summarising exposure in plain language. The Bill itself is silent on AI. The threat is not.
Three rules make the step work. AI produces the first draft, and a named person validates it, because an unchecked mapping is not a control but a new risk. Every mapping points to a specific obligation, such as a Cyber Assessment Framework outcome or a duty under the Bill, with the evidence attached. And where a finding cannot be resolved in time, it becomes an explicit risk decision by a named director, never something quietly left with the security team.
What this does not solve
None of this makes an organisation immune. Tools do not create accountability; people do. AI can produce a confident but wrong answer, which is why validation sits inside the loop. And the detail of the regime will depend on the Bill’s final text and secondary legislation.
The government’s Cyber Governance Code of Practice sets out what directors should oversee, and the loop gives them the information to do it.
The board’s problem now
Directors do not need to become security engineers. The task is narrower: see exposure as it changes, connect it to the obligations it threatens, and decide before the clock starts. Monitor continuously. Map to obligations. Report by exception. Decide in advance.
The chair in that second boardroom was not difficult. She was asking the question the Cyber Security and Resilience Bill expects every board to be able to answer.
About the author
Saviona Pereira is a cloud security specialist and leader in AI-assisted governance, continuous assurance and regulatory compliance. She uses cloud-native monitoring and AI-enabled methodologies to bridge the gap between security engineering and the boardroom, linking live exposure to regulatory obligations. Her work builds scalable, accountable oversight processes that help organisations meet incident reporting duties and adopt cloud and AI technology safely.
Feature image by Kampus Production from Pexels