How to Build an AI ROI Business Case That Gets Sign-Off in UK SMEs

Building a credible AI ROI business case for a UK SME means translating vague promises about efficiency into numbers a board or senior stakeholder can test, challenge, and ultimately approve. Most AI proposals fail at the sign-off stage not because the idea is wrong, but because the person presenting it has not connected the technology to a specific financial outcome the decision-maker actually cares about. This post gives you a repeatable framework to change that.

Why most AI proposals get rejected before the meeting ends

The usual pattern goes like this. Someone in the business has spotted an AI tool that looks promising. They put together a slide deck, talk about time savings, mention that competitors are already doing it, and then face a wall of silence from the people holding the budget.

That silence is not scepticism about AI. It is scepticism about the numbers.

Stakeholders in UK SMEs, whether that is a board of directors, a managing partner, or a founder who still signs every invoice personally, need to see three things before they approve any meaningful spend. They need to know what problem you are solving, what it costs the business right now, and what a realistic return looks like over a specific time horizon. Most AI proposals answer none of these questions with enough precision to justify a yes.

The result is a defer. Which in most organisations means a no that takes six months to officially become official.

The cost of doing nothing is the foundation of your case

Before you put a single number on an AI investment, you need to quantify what the current situation is costing the business. This is the step most people skip, and it is the one that matters most.

Start with time. Find the process you want to automate or improve, then work out how many hours per week your team spends on it. Be specific. Do not estimate. Pull the data from your project management tool, your CRM, or simply ask the person doing the task to track their time for five working days. You will almost always find the number is higher than anyone expected.

Then apply a cost. Use fully loaded hourly rates, not just salary. A team member earning £35,000 per year costs the business closer to £45,000 to £50,000 when you add employer National Insurance contributions, pension, and any overhead allocation. The HMRC employer NIC guidance gives you the current rates to work with.

So if a manual process takes 10 hours per week across two people at an average fully loaded cost of £22 per hour, that is £220 per week, £11,440 per year. That is your baseline cost of inaction. Now you have something to measure against.

Building the return side of the equation

This is where AI ROI cases fall apart. People claim AI will save 80% of the time on a task and present that as certain. Boards do not believe it, and frankly, neither should you until you have tested it.

The framework I use, and that Luke Stillwell has applied across multiple business turnarounds and client engagements over 19 years of running businesses, separates the return into three tiers.

Tier one | hard savings

These are direct, measurable reductions in cost or time that you can commit to with reasonable confidence. For example, an AI-assisted first-draft tool that cuts the time to produce a client report from four hours to forty-five minutes. You can measure this before and after. You can put a number on it. This is your anchor.

Tier two | soft savings

These are real but harder to prove precisely. Faster response times to customer enquiries, fewer errors in data entry, reduced rework on proposals. You should include these, but present them honestly as indicative rather than guaranteed. Stakeholders who spot inflated soft savings will discount your entire case.

Tier three | growth opportunity

This is where you point to what becomes possible when you free up capacity. If your sales team spends twelve hours a week on admin that AI handles, those twelve hours can go back into relationship building or new business development. Do not model this as certain revenue. Frame it as unlocked potential with a conservative estimate attached.

Present all three tiers transparently. Label them. Stakeholders respect honesty about uncertainty far more than they respect optimistic projections that do not survive contact with reality.

A real example from a professional services firm

We worked with a financial planning firm in the East Midlands who wanted to automate their client onboarding process. The existing process involved manually collecting documents, chasing clients by email, cross-referencing data across three systems, and manually populating compliance records. It took an average of four and a half hours per new client, and they were bringing on roughly fifteen new clients per month.

That was 67.5 hours of staff time per month, at a fully loaded cost of around £24 per hour. So approximately £1,620 per month, or £19,440 per year, spent on a process that added no value to the client experience.

We built an automated onboarding workflow using n8n, connected to their CRM and document management system, with AI-assisted data extraction handling the compliance cross-referencing. The new process took forty minutes per client, including the human review step we kept deliberately. Total implementation cost including our time was £6,800.

Payback period was four months. Annual saving year one, after implementation cost, was just over £14,000. Year two and beyond, the full £19,440 drops straight to the bottom line.

That is the kind of case a board signs off on before the meeting ends.

How to present the case without losing the room

The format of your presentation matters as much as the numbers. Stakeholders in SMEs are not reading forty-page business cases. They want clarity, not comprehensiveness.

Structure your presentation around four questions in this exact order.

First, what problem are we solving and how much is it currently costing us. Second, what are we proposing to do about it. Third, what does a realistic return look like and over what timeframe. Fourth, what do we need to approve today to get started.

Keep it to one page or five slides. Every additional slide is an opportunity for a stakeholder to find something to object to rather than focusing on the decision you need them to make.

One rule that never fails. Present your worst-case scenario first. Show the board what happens if the AI saves only 50% of the projected time rather than 75%. If the business case still holds at 50%, you have a resilient proposal. Stakeholders who see you have stress-tested your own numbers are far more likely to trust the numbers you are showing them.

The mistake of starting with the tool

‘We should implement Claude’ or ‘we need HubSpot AI’ is not a business case. It is a product preference dressed up as a strategy.

Start with the pain. Find the process that is costing the most time or producing the most errors. Quantify it. Then, and only then, identify which AI approach addresses that specific problem. The tool is the answer to a question, not the starting point for a conversation.

At 07 Heaven Systems, the first thing we do with any client considering AI is map the processes that are already breaking. Because AI applied to a broken process just makes the breakage faster. Fix the process first, then automate it. The ROI case writes itself when you approach it that way.

‘AI applied to a broken process just makes the breakage faster. Fix the process first, then automate it.’ That is the line I say in every workshop and every client kickoff. It changes the conversation immediately.

What sign-off actually requires

UK SME boards and senior stakeholders generally need five things to say yes to an AI investment.

They need a clear problem statement tied to a specific cost. They need a realistic return projection with a defined payback period. They need to understand the implementation risk and what the fallback plan is. They need to know who owns the project internally once it is approved. And they need a defined point at which the business reviews whether the investment delivered what was promised.

That last one is often missed. Build a review milestone into your proposal. Three months after go-live, we will measure X, Y, and Z against the baseline. This signals that you are treating the investment as a business decision, not a technology experiment. Stakeholders respond very differently to that framing.

The BSI AI standards guidance also gives UK organisations a useful framework for responsible AI governance, which can strengthen the credibility of your proposal if you are presenting to a risk-conscious board.

A note on government support you might be leaving on the table

UK SMEs investing in AI and automation may be eligible for support through schemes like Innovate UK Smart Grants or the Made Smarter programme in certain regions. These do not make a weak business case strong, but they can meaningfully reduce the net cost of a strong one. The Innovate UK Smart Grants guidance on GOV.UK is the right place to check current availability. Worth twenty minutes of your time before you finalise the investment figure you are presenting.

Common questions

How do I calculate ROI for an AI project in a small business

Start by calculating the fully loaded cost of the process you are targeting, including staff time, error correction, and any downstream delays the process causes. Then estimate the reduction in that cost once AI is handling part of the workflow, using a conservative multiplier rather than the best-case scenario. Subtract the implementation and ongoing licence costs from the annual saving to get your net return. Divide the implementation cost by the monthly saving to get your payback period. If that payback period is under twelve months, most boards will approve it. If it is under six months, the conversation is usually very short.

What does a business case for AI automation actually need to include

A credible business case for AI automation in a UK SME needs four core components. A baseline cost calculation for the current process, a realistic projection of savings broken into hard savings and indicative soft savings, a clear implementation plan with a named owner and a defined timeline, and a review milestone that specifies how you will measure success. Anything beyond that risks burying the decision in detail. Stakeholders do not need to understand how the technology works. They need to believe the numbers are honest and that someone is accountable for delivering them.

How long does it take to see ROI from AI in a small business

For process automation projects targeting a specific, well-defined workflow, payback periods of three to six months are realistic and common. We built an onboarding automation for a financial planning firm that paid back in four months. For broader AI implementations touching multiple departments or requiring significant integration work, twelve to eighteen months is a more honest expectation. The key variable is how clearly you have defined the problem before you start. Vague implementations take longer and deliver less. Specific, well-scoped projects deliver returns faster because there is less room for the project to drift.

If you want to build your own AI ROI business case with guidance from people who have done this across multiple industries, join us at our next AI workshop. Join our next AI workshop.

Rather have it built for you? See what an AI automation agency actually does and how we build and run these systems for UK businesses.