How conservative should the financial projections be?

Conservative enough that you would defend them to the audit committee twelve months later. Year-one benefits should assume slower adoption, lower utilisation and one or two unexpected costs. If a vendor projects a 40% time saving, model 20% in year one and review on actuals. Boards reward proposers who underclaim and overdeliver. The opposite — claiming 40%, delivering 18% — burns the credibility you need to fund the next phase. Build the case on baselines you have measured, not on vendor pitch decks.

Do we need to include consultancy fees in the cost section?

Yes, if you are using external advisory support. The template includes a Fractional AI Director / consultancy line because boards expect to see it. Hiding it makes the rest of the case look incomplete. State the engagement type, the duration, and what the consultancy is responsible for delivering. If the engagement reduces in year two and three, show that profile. A board that sees a credible reduction over time gains confidence that internal capability is being built rather than dependency entrenched.

What if we cannot quantify some of the benefits?

Be honest about it. Some benefits — improved customer satisfaction, reduced staff frustration, better compliance posture — are real but harder to put a pound figure on in year one. List them under the benefits table without a number, and describe how they will be measured once the system is live (NPS surveys, staff pulse surveys, audit findings). A board respects an honest "we will measure this from month three" more than a fabricated figure that will not survive the first review.

How long does it take to complete the template properly?

Two to four weeks of part-time work for a first AI business case. The financial section takes longest because the baselines often have to be measured for the first time. Plan for time with operations to capture current process times, time with finance to validate cost assumptions, and time with the DPO and IT to scope governance and integration. If you can complete it in a long weekend, the assumptions are probably not robust enough. If it takes longer than a month, scope is too broad — split it.

Should we use this template for a small pilot, or only for major investment?

For a pilot under about £10,000 with no personal data, a one-page version is enough — strategic rationale, expected outcome, success measure, owner. Use the full template when you are asking the board for budget that requires formal approval, when personal data is involved, or when the AI system will influence decisions affecting customers, staff or service users. The discipline of completing all six sections also catches gaps before they become awkward questions in the meeting.

Can AI-Si.com help build a bespoke version?

Yes. AI-Si.com prepares board-ready AI business cases for UK organisations, including the financial model, the risk register and the governance framework. The work typically runs alongside the AI Readiness Audit so the case is built on evidence rather than estimation. If you would prefer to draft the case in-house and have it reviewed before it goes to the board, that is also available. Call Simon on 07973 210 895 for a 30-minute consultation to scope what would help.

Executive Resources · for UK SME leaders

AI Business Case Template

An AI business case has to survive a finance director, a sceptical board and the audit trail that follows. This template gives you the six sections boards expect: executive summary, strategic rationale, financial analysis, risk register, governance and oversight, and a clear recommendation. Each section names the data the board will ask for and the assumptions you must make explicit. Complete it with your own organisation's numbers — boards reject AI cases built on vendor productivity claims, not on evidenced baselines.

Why this template exists

Most AI business cases fail at the board because they read like a vendor pitch. They lead with the technology, hand-wave the financial model, and treat governance as an appendix. Boards see this every quarter and they have learned to discount it.

This template flips the structure. Strategic rationale comes before financials. Financials come before risk. Governance is treated as part of the proposal, not a future workstream. The recommendation at the end is specific: an amount, a timeline, a sponsor, a review cadence. Board members can read it once and approve or push back without ambiguity.

Section 1: Executive Summary

Write this section last. It needs three to four sentences summarising the proposed AI investment, its strategic rationale, the headline financial metrics, and what the board is being asked to approve. The model paragraph from the source reads: "We propose an initial AI investment of [£amount] to automate [specific processes] across [departments], delivering an estimated annual saving of [£amount] and reducing [manual task] time by [%]. The investment will repay within [months] and positions [Organisation] ahead of [competitor/regulatory] expectations for AI adoption. The board is asked to approve the Phase 1 budget and governance framework."

Underneath that paragraph, state the headline numbers in three lines: total investment year one and the year two to three annual figure; expected annual saving and which year it begins; payback period and the strategic priority it aligns to.

Section 2: Strategic Rationale

Explain why AI, why now, and why this specific use case. Boards respond to strategic context, not technology enthusiasm. Cover four points in order: the business problem being solved, the strategic alignment, the competitive and sector context, and the risk of not investing.

Quantify the problem wherever possible. "Our team spends 14 hours per week on manual data entry" beats "data entry is inefficient" every time. Reference your existing strategic plan or operational priorities directly. State what peer organisations are doing. Spell out the cost of standing still — opportunity cost, competitive displacement, regulatory exposure where it applies.

2.1 Business problem being solved — quantified examples, not adjectives
2.2 Strategic alignment — which board-level objective this supports
2.3 Competitive and sector context — what peers and regulators are doing
2.4 Risk of not investing — cost of the status quo

Section 3: Financial Analysis

Quantify costs and benefits with as much specificity as you can support. Boards are sceptical of AI cases that rely on vague productivity claims. Be conservative and transparent about your assumptions — a smaller, defendable number wins approval more often than an ambitious one without working.

Lay out costs and benefits across three years. The cost categories below are the standard set; the benefit categories must each name how the value will be measured. The bottom of the section gives payback period, three-year ROI and NPV at a 10% discount rate.

Cost category	Year 1	Year 2	Year 3
AI tool licences / subscription	£	£	£
Implementation and integration	£	—	—
Staff training	£	£	£
Fractional AI Director / consultancy	£	£	£
Internal staff time, implementation	£	—	—
Governance and compliance setup	£	—	—
TOTAL INVESTMENT	£	£	£

Benefits and how they will be measured

Each benefit line needs a measurement method. Without one, the benefit is an assertion and the board will treat it as such.

Benefit category	Year 1	Year 2	Year 3	How measured
Staff time saved (hours × cost/hr)	£	£	£	Timesheets / observation
Error reduction, rework cost saved	£	£	£	Error log comparison
Revenue uplift (where applicable)	£	£	£	Sales / CRM data
Compliance cost reduction	£	£	£	Audit / legal fee comparison
Customer satisfaction improvement	£	£	£	NPS / CSAT scores
TOTAL BENEFITS	£	£	£
NET BENEFIT / (COST)	£	£	£

Section 4: Risk Register

A board paper without a credible risk register reads as either naive or evasive. The six risks below are the ones that recur across AI investment cases. Assess each on likelihood and impact, and name the mitigation explicitly.

Risk	Likelihood	Impact	Mitigation
AI tool does not deliver expected performance in production	Medium	High	Pilot before full rollout. Performance benchmarks in vendor contract. 90-day exit clause.
Staff resistance to AI adoption	High	Medium	Fear reduction workshops before deployment. Champion programme. Change management support.
Data protection breach via AI tool	Low	Very High	DPIA completed. Strict data handling policy. No personal data in non-approved tools.
AI produces inaccurate outputs used in decisions	Medium	High	Human oversight requirement for all significant decisions. Output verification protocols.
Regulatory change affects AI tool viability	Low	Medium	Annual governance review. Flexible tool selection. Avoid single-vendor lock-in.
Benefits realisation slower than projected	Medium	Medium	Conservative projections in Year 1. Monthly performance tracking against baseline.

Section 5: Governance and Oversight

Board approval of this business case includes approval of the governance framework. State this explicitly. Four elements must be defined before the paper goes to the board.

An AI Steering Committee meeting monthly, chaired by a named individual, with a defined membership, responsible for strategic direction, escalations and quarterly board reporting. Data protection oversight by the DPO, with a DPIA required before any personal data processing and ICO notification where it applies. A monthly KPI dashboard to the Steering Committee, a quarterly board report, and an annual independent review against stated objectives. Annual programme review with the option to pause, scale or exit, and vendor contracts that include data portability and deletion provisions.

Section 6: Board Recommendation

State clearly what the board is being asked to approve. Vagueness here is the single most common reason an otherwise good paper is deferred to the next meeting. The recommendation should be five specific items.

Approve a Year 1 AI programme budget of £_____________
Approve the AI Governance Framework as set out in Section 5
Designate [Name] as Executive Sponsor for the AI programme
Receive a quarterly AI programme update from the Steering Committee
Authorise Phase 1 commencement by [target date]

Take the next step

Want help applying this to your organisation? Use the resource below or book a 30 minute strategy call with Simon — no pitch, just practical advice.

Frequently asked questions

Related resources

Executive Resources

Find Out Where AI Can Save or Generate Money in Your Organisation

Book a free 30-minute call with Simon. Bring a real problem — staff time, governance worry, vendor proposal, failing pilot — and leave with a concrete first step you can take next week.

07973 210 895