Why the 2026 Cycle Is Different
The H1 board pack is a mid-year artefact for boards on a December financial year, prepared in May or June for a June or July meeting. The template most companies use was built between 2018 and 2022. It is now unfit for the 2026 board environment in three specific ways.
First, non-executive directors report that AI spend, both in the run-cost line and in capitalised software, is now the single most opaque item in most packs. A board that was comfortable with a single "IT and software" line in 2022 is no longer comfortable with the same line when 30 to 40 per cent of it is model inference, API charges and vendor subscriptions to LLM providers that did not exist eighteen months ago.
Second, the runway conversation has changed. In 2022 to 2023 the standard runway page showed a single point estimate against a Series A or B ambition. In 2026, boards want a bracketed range against three scenarios (base, downside, and a specifically-named upside) and an explicit statement of what triggers a fundraising decision. The point estimate has become misleading in an environment where MRR growth, churn, and gross margin can shift by 3 to 5 points across a quarter.
Third, geopolitical exposure was a non-issue in 2018. In 2026 it is a first-order risk for any company with US customer concentration, EU regulatory exposure, or a supply chain touching Asia. A board pack that does not name this exposure is not doing its job.
Add: The Runway Scenario Page
Replace the single-runway chart with a three-scenario view. The scenarios should not be "conservative, base, aggressive" — that framing invites the board to average them and land back at the base case. Use scenarios that force the board to engage with the underlying assumptions.
The scenario page should include one graph (runway across the three scenarios, months on x-axis, cash on y-axis) and a short table showing the assumptions that differ. Total page length: one page. The board does not want ten scenarios; it wants three that stress the model in ways they recognise.
The "fundraise trigger" line is the piece that non-executive directors most consistently ask for and most rarely see. Naming the trigger before you need to raise moves the conversation from "should we raise?" to "the trigger is in view; here is the plan" — which is a materially better place to be with a board.
Add: The AI Spend Disclosure Line
The board wants to know three things about your AI spend: how much you are spending, on what, and what governance is in place around it. Most 2024 board packs answered none of these questions.
How Much
Break AI spend out of the "software and IT" line into its own line. Split into three sub-categories: model inference costs (per-token API charges to model providers), agent and workflow platform subscriptions (Anthropic Claude, OpenAI, Vercel AI SDK, LangChain platform, etc.), and internal AI tooling (Cursor, Copilot, Windsurf and equivalents used by engineering and support teams).
On What
For each sub-category, state the primary use case in one line. "Model inference: customer support triage bot handling 60 per cent of tier-1 tickets." The board cares less about the total than about whether you can articulate what the spend is doing. If you cannot articulate it in one line per category, you have a governance gap.
What Governance
State whether you have a documented AI usage policy, whether new AI vendors go through a security review, and whether any AI-produced output is being incorporated into customer-facing product without human review. Three yes/no questions on the page. If two of the three answers are no, the board will (rightly) push you to address them.
Add: The Geopolitical Exposure Map
A single page listing the exposures that could materially affect the business, with an assessment of the current risk level. This is not a Chief Risk Officer paper; it is a CFO framing of business risk. Include only exposures where the answer to "would this materially move revenue, cost or continuity within 12 months?" is yes.
For a typical UK-headquartered fintech in 2026, the exposures worth naming are:
- US customer concentration: Percentage of ARR from US customers, and any known regulatory or tariff exposure. Currency risk if not hedged.
- EU regulatory divergence: Where UK and EU rules are drifting (MiCA vs UK crypto regime, DORA extraterritoriality, GDPR-UK GDPR divergence). Any operational or compliance cost this is creating.
- Cloud and vendor concentration: AWS, GCP or Azure concentration, and the specific region hosting your data. Vendor lock-in for AI models used in production.
- Supply chain and payments rails: Any dependency on non-UK/EU payment rails, custody providers, or hardware supply. Sanctions exposure through customers or investors.
Format: a one-page table with columns for exposure, current status, direction of travel, and specific mitigation. Directors want to see that management is tracking these; they do not want a five-page essay.
Delete: The Commentary Narrative Section
Non-executive directors in the IoD Board Effectiveness Survey 2026 reported that the narrative section (typically pages three to six of most board packs, containing extended commentary on the previous quarter's performance) is the section they read least. The reasons given were consistent: it repeats data already visible in the KPI dashboard, it is written to justify performance rather than to inform, and by the time it reaches the board the operational context has moved on.
The replacement is a single-page "management commentary" that answers three questions: what surprised us in H1, what are we changing as a result, and what is the single ask of the board. Everything else — the "we saw strong momentum in the SME segment" language, the multi-paragraph vendor updates, the marketing funnel deep-dives — goes into an appendix that any director can read if they want to, but that does not consume the meeting.
"Directors are not asking for less information. They are asking for information organised around decisions they need to make. The runway scenario page, the AI spend disclosure, and the geopolitical map exist because there are decisions attached to each of them. The narrative section has no decision attached to it, and that is why it fails."
Restructure: The KPI Page
The KPI dashboard was invented for a period when the primary board question was "are we hitting the plan?". In 2026, the primary board question is closer to "is the plan still the right plan?". The dashboard needs to reflect that shift.
Two structural changes work. First, remove the "traffic light" column. The traffic light forces a binary judgement on data that is often ambiguous, and consistently over-signals problems in green (metrics that are on plan but structurally weakening) and under-signals problems in amber (metrics that are marginally below plan but on a trajectory that is fine). Replace with a direction-of-travel arrow.
Second, add a "cohort commentary" row under each headline metric. One sentence stating what has moved within the metric since the last board meeting. "MRR up 4 per cent, but the movement is heavily weighted to two new large customers; net new logos flat month on month." The board reads the headline number, sees the direction, and reads the commentary in ten seconds. That is a superior signal to a traffic light that has already been rounded.
Length, Format, and What to Cut
The H1 2026 pack should target 22 to 28 pages, down from a typical 35 to 45 in most companies' 2024 versions. The compression is achieved by moving material out of the pack proper into a well-organised appendix, not by omitting anything.
The functional updates section is where most companies over-deliver. Four pages is enough — one for product, one for go-to-market, half for people, half for operations. If any function feels the need for more space, that is a signal to have a dedicated deep-dive in a subsequent board meeting rather than to expand the pack.
Key Takeaways
- The H1 2026 board pack cycle begins in May for December year-ends. Redesign now rather than after the meeting.
- Add: a runway scenario page with three named scenarios and an explicit fundraise trigger, not a single point estimate.
- Add: an AI spend disclosure line broken into inference, platform, and internal tooling, with a three-question governance status.
- Add: a geopolitical exposure page listing the four or five exposures that could move revenue, cost or continuity within twelve months.
- Delete: the narrative commentary section. Replace with a single-page "management commentary" answering what surprised us, what we are changing, and what we ask the board.
- Restructure: remove the KPI traffic light column, replace with direction arrows and a one-sentence cohort commentary under each metric.
- Target 22 to 28 total pages, with cut material moving to a well-organised opt-in appendix rather than being omitted.
- Apply the "chair test" before sending: can your chair, on 48 hours of reading, articulate the three decisions the meeting should reach?