A board can receive every scheduled report and still discuss the most important investment after the realistic alternatives have disappeared. For a Norwegian business considering AI, that might mean approving a supplier after integration has begun, or reviewing staffing assumptions after customers have already been promised a lower price.
The chair’s practical challenge is to get the right decisions onto the agenda while the company still has choices. An annual plan should therefore show when the board needs to decide, what evidence it will need, and when management must begin preparing it.
Start with the next commitments
Ask the CEO which decisions in the coming year could materially change earnings, cash needs, capabilities or dependence on others. Include customer contract renewals, major system changes, financing and critical supplier renewals. Then work backwards from each commitment.
An AI investment may deserve an early discussion about customer value, followed by a limited experiment and only later a funding decision. A cybersecurity discussion may need to precede a busy delivery season so the company can test recovery before an interruption becomes especially costly.
Keep enough room for unexpected matters. An annual plan is a working agreement about preparation, not a reason to postpone an urgent issue until its scheduled month.
Separate the chair’s role from the CEO’s
For Norwegian AS companies, sections 6-20 and 6-21 of the Companies Act place responsibility on the chair to ensure relevant board matters are considered, and on the CEO to prepare matters in consultation with the chair. Papers must provide an adequate basis for consideration. Board members and the CEO can also require specific matters to be considered.
My recommendation is to turn that division into a visible preparation schedule. The chair clarifies the question and protects time for discussion. Management develops the options, tests assumptions and identifies implementation capacity. The board examines the proposal and makes decisions within its remit. The chair should not become a parallel project manager.
The NUES recommendation dated 28 August 2025 is aimed primarily at listed companies. Its emphasis on constructive discussion and evaluation of the board’s work is also useful inspiration for smaller boards, but should not be presented as a universal legal requirement for every AS. These references were checked on 12 October 2026.
Four conversations to place in the year
Use these as connected conversations, timed to the business. They are my proposed structure, not statutory quarterly requirements or a complete calendar of corporate obligations.
1. Where could our earnings change? Ask management to show how customers buy, what they will pay for, and where competitors or suppliers could change the economics. For AI, include Norwegian-language quality and the cost of human checking. The output is a short list of opportunities and exposures worth investigating, with assumptions that could prove wrong.
2. What will we commit, and what will we stop? Review evidence before approving a larger investment. Compare a limited trial, a wider rollout and postponement. Name the management capacity each option requires. Delaying a useful improvement has a cost, but a company cannot run every initiative at once. The output is a decision with spending limits, an accountable executive and conditions for the next stage.
3. Can we deliver when a dependency fails? Select one business-critical service and examine a realistic interruption. Ask what a recovery exercise demonstrated, what remains untested and what customers would experience. Management owns the exercise. The board considers whether the remaining exposure is acceptable and whether priorities must change. Use a cyber-risk report tied to business consequences to support that discussion.
4. What did we learn, and what changes next year? Compare promised outcomes with observed results, including quality, workload and customer effects. Revisit decisions that depended on uncertain assumptions. Ask whether board papers and the board’s own competence were sufficient. The output is a revised agenda and a small number of concrete improvements.
A hypothetical example: the renewal that cannot wait
Imagine a Norwegian wholesale AS whose order-management contract renews in September. Management also wants AI assistance for incoming orders. If the board first discusses the dependency in August, the practical choice may already be renewal on the incumbent’s terms.
A better sequence could be an options paper in February, a limited test in April and a decision in June. Management would examine integration costs, Norwegian order accuracy, supplier alternatives and staff capacity. The board could then choose renewal, negotiation or a staged change with evidence in hand. The months are illustrative, not a recommended timetable for all companies.
The decision paper should identify what would be lost by waiting as well as what could go wrong by proceeding. For the investment mandate itself, see what a Norwegian board should require before approving an AI strategy.
Put these five fields beside each major matter
- The board decision required, expressed as a question.
- The latest useful decision date, before commercial options close.
- The executive responsible for preparing the matter.
- The evidence and alternatives the board expects to receive.
- The event or result that would bring the matter back earlier.
At the next meeting, test the plan against one upcoming commitment. If the board cannot identify when it will see the evidence, the agenda is not yet helping it influence the outcome.