8. How to Build a Business Case for AI at Your Charity

Charities face a business case problem that no other sector has. When a commercial business wants to invest in AI, it has one audience: the board, who want to see return on investment. A charity has three.

Your board wants to know the risk is managed and the governance is sound. Your funders want to know you're spending wisely and delivering impact. And your staff want to know their jobs are safe and their workload is about to get better, not worse.

These three audiences want different things, worry about different things, and need to hear different messages. A business case that satisfies one can actively alarm another. Tell your board AI will save money and they'll ask about risk. Tell your staff AI will transform workflows and they'll hear "redundancies." Tell your funder AI is cutting costs and they'll wonder whether you're still delivering quality.

This post shows you how to build a business case that speaks to all three — with a worked example you can adapt.

Audience 1: Your board — governance and risk

Trustees aren't opposed to AI. But their primary concern is duty of care: is this being done responsibly, and could it harm the charity? If you walk into a board meeting leading with "AI could save us thousands of pounds," you'll spend the next forty minutes answering questions about data protection and reputational risk.

Lead with governance instead.

The business case for your board should open with three facts: the charity already uses AI (if staff have been using ChatGPT informally, say so — this reframes the decision from "should we start?" to "should we get organised about what's already happening?"); the Charity Governance Code now expects boards to consider technology and AI risk; and you have a plan to manage that risk through a policy, data boundaries, and human review.

Then — and only then — present the opportunity. Frame it as capacity recovery, not cost cutting. "AI tools can recover the equivalent of X hours per week in admin time, allowing staff to focus on [delivery/fundraising/beneficiary work]." Trustees are much more comfortable with "do more with what we have" than "spend less on people."

The specific data point to use: if one staff member saves five hours per week on administrative tasks — and the evidence from organisations that have adopted AI suggests this is achievable for admin-heavy roles — that's 260 hours per year. At a fully loaded cost of £15 per hour, that's £3,900 in recovered capacity per person. For a two-person charity, that's nearly £8,000 — the equivalent of a month's full-time salary redirected from admin to mission.

But be honest about the "workslop" factor. A 2025 Workday survey found that more than a third of time saved through AI is spent correcting, rewriting, or contextualising subpar AI outputs. So your net saving is closer to 3.5 hours per week, not five. Present the conservative figure. Trustees trust honesty more than optimism.

End with a clear ask: approve the AI policy, agree to a three-month trial, and schedule a review at the next board meeting. That's a decision trustees can make in ten minutes.

Audience 2: Your funder — value for money and impact

Funders care about two things: are you spending their money well, and are you delivering outcomes? AI speaks to both, but only if you frame it correctly.

The value-for-money argument is straightforward. If AI tools save staff time, that time can be redirected to direct delivery. A funder who gave you £50,000 to run a community programme would rather see your project coordinator spending 30 hours a week with beneficiaries and 7 hours on admin than 20 hours with beneficiaries and 17 hours writing reports. AI doesn't reduce the cost of the programme. It increases the proportion of funded time spent on the thing the funder is paying for.

This is important because funders are increasingly aware that charities use AI, and many are forming views about it. The Charity Excellence Grant Making Survey 2026 found that 64% of charities have used AI for grant applications. IVAR has flagged concerns about AI-generated proposals obscuring organisational voice. The Fundraising Regulator published AI guidance in December 2025. Funders are paying attention.

The strongest funder-facing message is: "We use AI to reduce the administrative burden on funded staff, allowing more time for delivery. All AI-assisted outputs are reviewed by a staff member. We have a policy in place. We're happy to share it."

Some funders will ask whether AI was used in your application or report. Be transparent. "We used AI to structure the first draft and check for clarity. The content, data, and analysis are ours." That answer builds trust. Evasion or silence erodes it.

For impact reporting, AI can genuinely help you tell a better story. If your monitoring data shows you supported 150 people, AI can help you structure that data into a narrative that demonstrates outcomes rather than just outputs. A funder who receives a clear, evidence-rich report is more likely to fund you again — and AI makes producing that report faster and more consistent.

Audience 3: Your staff — workload and job security

This is the audience most business cases ignore, and it's the one that determines whether AI actually gets adopted. One-third of nonprofit managers list employee resistance as a barrier to AI adoption, according to the 2024 Nonprofit Standards Benchmarking Survey. In the wider UK economy, 17% of employers expected to reduce headcount due to AI in 2025.

Your staff have read the headlines about AI use and job cuts. Even if those cuts weren't driven by AI, the association between "technology" and "fewer jobs" is strong.

The business case for staff needs to address this directly. Not with vague reassurance — "don't worry, AI won't replace you" — but with a specific commitment: AI is being adopted to reduce administrative burden, not to reduce headcount. The time recovered will be redirected to [specific activities — beneficiary work, service delivery, fundraising]. No roles are being reviewed as a result of this decision.

If you can't make that commitment honestly, don't make it. But for most small charities, AI adoption genuinely is about capacity rather than cuts — you don't have enough people as it is.

The staff-facing message should also emphasise agency. Staff shouldn't feel like AI is being done to them. Involve them in choosing which tasks to prioritise. Ask them where they lose the most time. Let them shape the prompt library. When people have input into how AI is used, resistance drops dramatically — because they can see that it's making their specific work easier, not threatening their specific role.

One powerful approach: ask staff to track their time for a week before and after AI adoption on the tasks you've identified. When they can see their own data — "I spent 12 hours on report formatting last month and 5 hours this month" — the case makes itself. And that data feeds directly into your board and funder business cases too.

A worked example

Let's make this concrete. Imagine a charity with three staff members and an annual income of £180,000. The CEO/manager spends roughly 10 hours per week on admin (board papers, funder reports, correspondence). The project coordinator spends about 8 hours per week on reporting and communications. The admin officer spends about 12 hours per week on data entry, meeting notes, and routine correspondence.

That's 30 hours of admin time per week across three people. At a blended cost of £15 per hour, that's £23,400 per year spent on tasks that AI could partially automate.

Conservative estimate (accounting for the "workslop" factor): AI reduces admin time by 40%, saving 12 hours per week across the team. Net value: £9,360 per year in recovered capacity.

That's not revenue. You can't bank it. But you can demonstrate to a funder that 12 hours per week have been redirected from administration to delivery. You can show your board that staff are spending more time on mission and less on paperwork. And you can show staff that their working week has meaningfully improved.

The cost of achieving this: zero, if using free AI tools. £400 if using the TSAI toolkit for structured implementation. £20–30 per user per month if upgrading to paid AI tool tiers with nonprofit discounts.

Against £9,360 in recovered capacity, any of those costs pays back within the first month.

Building the one-page business case

Keep it short. Trustees, funders, and staff all respond better to a single page than to a strategy document. Structure it like this:

Current state: "Staff spend approximately [X] hours per week on administrative tasks including [list]. This represents [£Y] in staff time annually."

Proposal: "We propose adopting AI tools for [specific tasks] on a three-month trial basis. All use will be governed by our AI policy, which requires [data boundaries, human review, named oversight]."

Expected benefit: "Conservative estimate: [X] hours per week recovered, equivalent to [£Y] annually, redirected to [delivery/fundraising/beneficiary work]."

Cost: "[Zero for free tools / £X for toolkit / £X per month for paid plans]."

Risk management: "AI policy in place. No personal data entered. All outputs reviewed by staff. Three-month review scheduled."

Decision requested: "Approve the trial. Review at [date]."

That's your business case. One page. Three audiences addressed. Evidence-based. Honest about both benefits and limitations.

Module 5 of the Third Sector AI Toolkit includes an ROI calculator that generates these numbers from your organisation's specific data — staff costs, hours spent on admin, team size — and produces a board-ready one-page business case. Or you can build your own using the framework above. Either way, the argument is the same: AI recovers capacity, not just time. And for a sector stretched thinner than ever, capacity is the most valuable thing there is.