THE CEO VIEW | By David Harris, CEO, HAWK Network Defense
What decision could your board make from the cyber report it received last quarter?
I do not mean whether directors understood every technical term. I mean whether they could explain the business exposure, evaluate management's recommendation and identify what would make them reconsider it.
A technically accurate presentation can leave those questions unanswered. It can describe a great deal of work while giving the board little basis for exercising judgment.
I believe management owns that translation. The board should bring curiosity and challenge, but directors should not have to reconstruct an enterprise risk decision from operational statistics. A useful report makes the connection between evidence and consequence visible enough to examine.
Begin with the business decision
Before assembling the slides, management should decide what the board needs to do with the information. Is this an update within an approved strategy? A request to change investment? An exposure that may exceed the organization's tolerance? A choice reserved for the board under its governance arrangements?
Those purposes call for different conversations. An informational update should not conceal a decision management is reluctant to make. Equally, an operational issue should not be escalated merely to distribute accountability.
NIST CSF 2.0 places oversight, roles and risk strategy within its Govern function.[1] My recommendation is to connect that governance work to an explicit decision at the beginning of the report.
The first page should allow a director to answer: what requires attention, why now, what does management recommend, and what happens if we defer?
Build a brief around a decision directors can test
I would organize that first page around a single consequential choice. In the hypothetical customer-account example, the choice might be whether to fund additional controls before allowing the AI service to change records. The brief should name the affected service, describe the present exposure and show the recommendation with its expected business effect.
The supporting evidence should be traceable without overwhelming the initial explanation. A director who asks why management believes a boundary is effective should be able to reach the relevant assessment, its scope and its owner. The presentation should summarize the finding accurately while preserving a path to deeper scrutiny.
The brief also needs a decision deadline and a reason for it. A contract renewal, a planned expansion or an expiring exception can explain timing. An arbitrary date cannot establish urgency. If the issue does not require a board decision, management should say what it is handling within delegated authority and what development would bring the matter back. That distinction protects the board's attention while keeping ownership clear.
Translate the exposure without stripping away the mechanism
Consider a hypothetical AI service that can access customer records and initiate account changes. Reporting that its access review is incomplete describes an activity gap. It does not yet explain the business exposure.
A more useful account would identify which records the service can reach, which changes it can make, what restrictions have been tested, and what customer harm an unauthorized action could create. Management should distinguish a plausible path from a confirmed event.
Directors do not need implementation detail on every permission. They do need enough causal explanation to challenge the proposed response. If the exposure depends on several conditions occurring together, those conditions belong in the explanation.
Plain language should preserve precision. Describing all access as dangerous or every unresolved issue as critical makes prioritization harder. Management must explain what distinguishes the exposure that deserves board attention.
Make uncertainty visible
A single financial estimate can look authoritative while concealing weak assumptions. I would rather see a defensible range, the conditions behind it and the limits of available evidence.
If management cannot quantify an exposure reliably, say so. Explain the consequences that can be described, what remains unknown, and how that uncertainty affects the decision. A transparent qualitative assessment can be more useful than a precise number produced from assumptions nobody can defend.
For AI, the evaluation context also matters. NIST's AI RMF calls for documented performance or assurance measures under conditions resembling deployment.[2] A favorable test in a narrow setting should not become an unqualified statement about the entire production service.
The board should be able to distinguish observed results, modeled outcomes and management judgment. These forms of evidence can all inform a decision, provided the report makes their differences clear.
Present alternatives that directors can compare
An investment request should describe credible choices. For the hypothetical AI service, those might include delaying deployment, limiting it to recommendations, or permitting a narrowly bounded set of actions after additional testing.
Each option has consequences. Delaying may postpone a business benefit. Restricting authority may increase manual work. Proceeding with broader access may expose customers to a consequence management cannot yet contain adequately.
Management should explain those tradeoffs using the same assumptions across options. Do not compare an optimistic benefit estimate for the preferred option with a worst-case cost for every alternative.
The recommendation must remain management's recommendation. Presenting choices without a reasoned position asks directors to do the operating team's work. Presenting only one choice without explaining the alternatives prevents useful challenge.
Show how individual exposures relate to the enterprise
I would also ask whether the exposure shares a cause or dependency with other risks already reported. If several services depend on the same supplier or identity platform, treating each service as a separate decision could conceal the value of a common response. Conversely, similar labels do not prove that two exposures have the same business consequence.
NIST IR 8286C Revision 1 describes integrating cybersecurity risk-register information into an enterprise risk portfolio and profile.[3] That supports bringing cyber exposure into the enterprise conversation. It does not justify adding unrelated worst-case estimates together and presenting the result as an expected loss.
Management should explain the relationships it has actually examined. Which investment addresses several exposures? Which assumption could make multiple services fail together? Which proposed response creates a different dependency? I would show these connections selectively, where they change the recommendation. The objective is a more coherent allocation of resources, with enough detail for directors to understand why a shared investment may be preferable to several isolated projects.
Keep ownership distinct from approval
Board approval should not make ownership disappear into the minutes. The business executive accountable for the outcome remains accountable, supported by the technical and control owners responsible for execution.
The report should identify who can implement the decision, who evaluates the evidence and who may accept residual exposure under the organization's arrangements. Those roles may differ. Naming them helps the board assess whether the proposed action is executable.
Directors should also understand the limits of the requested approval. Approving a funding envelope does not necessarily approve every deployment or exception within it. Management needs to explain what stays within delegated authority and what would return for further consideration.
This is a recommended management practice. Specific board duties and reserved matters depend on the organization and its applicable requirements.
The executive presenting the recommendation should be prepared for challenge from people outside the delivery team. I would ask the control assessor, business owner and finance partner to identify any material disagreement before the meeting. A polished consensus statement is unhelpful if the underlying evidence supports different interpretations.
For the hypothetical AI service, operations might believe a restricted release is manageable while the assessor finds that the restriction has not been demonstrated. The report should preserve that distinction. Management can explain whether it recommends postponement, additional evaluation or a narrower service, and who has the authority to accept the remaining exposure.
I would not ask the board to resolve an unexplained technical disagreement. Management should bring the decision consequence: what the conflicting assessment prevents us from knowing, which option remains defensible, and what evidence would resolve the issue. This approach allows directors to challenge management's judgment without requiring them to become the service's designers or release managers.
Close the loop after the meeting
A decision has a useful reporting life beyond the approval date. Management should state the expected change, the evidence that will demonstrate it and the conditions that would trigger another decision.
Suppose the board funds a restriction intended to reduce the hypothetical service's ability to make unauthorized account changes. The next update should address whether the restriction was implemented and tested, what it actually prevents, and which exposure remains.
A completed project is relevant evidence of delivery. It is not, by itself, proof of the risk outcome promised in the investment case.
When assumptions change, the report should show the change explicitly. Perhaps expected transaction volume increased or a supplier altered the service. The question becomes whether the original decision remains appropriate under the new conditions.
Keep the meaning of the measures stable
For follow-through, I would keep a decision record linking the approved investment to its intended result. The record should retain the original assumptions, the accountable executive, the expected evidence and the next review point. If the scope changes, management should show the change rather than silently reset the baseline.
Measures also need consistent definitions. A reduction in the number of exposed services is difficult to interpret if the service inventory changed or the definition of exposure narrowed. The update should explain those changes before attributing the result to the investment.
For the hypothetical account service, a meaningful update could state which actions were tested, which unauthorized actions were prevented and which operating conditions remain outside the evaluation. I would pair that with the effect on customer service and operating effort. A control that reduces one exposure but creates an unmanageable approval queue requires another management decision. Reporting both effects gives the board a more credible account of the investment than a favorable security metric alone.
I would judge a board report by what directors can do after reading it. Can they explain the consequence, challenge the assumptions, understand the owner's recommendation and recognize when the decision needs revisiting?
If they can, management has created a basis for governance. If they cannot, another appendix is unlikely to solve the problem. The next task is to make the decision interpretable.
Sources and references
[1] NIST Cybersecurity Framework 2.0
February 26, 2024; Govern function, GV.OC-04 and GV.OC-05, GV.RR and GV.OV. Supports dependency understanding and governance/oversight categories. Does not mandate the article's particular board report or resilience review.
[2] NIST AI Risk Management Framework 1.0 Core
2023 framework; MEASURE 2.3 and 2.7; GOVERN 2.3; MANAGE 2.4. Supports deployment-like evaluation, documented security/resilience and executive AI risk responsibility. Voluntary guidance; the site notes an update is in progress. No later draft is treated here as a final standard.
Final December 18, 2025; abstract. Verifies integration of cybersecurity risk registers into an enterprise risk portfolio/profile. The decision brief, aggregation cautions and decision record are management analysis, not a prescribed board template.

