Designing the Finance Function of the Future: Hybrid, Outsourced, AI-Enabled
Most SME finance functions weren’t built — they accumulated. Here’s how to move from a function that happened to one that’s designed for where you’re actually heading.
Nobody designs their finance function. Not really.
What actually happens is this: the business starts small, a bookkeeper handles the invoices and the VAT, an accountant files the year-end, and the founder fills in the gaps. They reconcile the bank on a Sunday evening. They approve every purchase order because nobody else can. They build the forecast in a spreadsheet they haven’t shared, because sharing it would mean explaining it, and explaining it would mean admitting how many of the numbers are instinct dressed up as analysis. It works. For a while, it works well enough. But at some point, usually somewhere between £3m and £5m, the weight of those gaps becomes difficult to ignore. The monthly board pack arrives late and tells the room too little. A pricing decision gets made on gut feel because nobody can produce margin data by client. The MD stays late to chase numbers that should have been on their desk before the first coffee. If that pattern feels familiar, we’ve explored why the founder so often becomes the bottleneck in an earlier piece.That’s not a failure of effort. It’s the natural consequence of a finance function that evolved by accident rather than by design.
This article is about what happens when you stop inheriting your finance function and start designing it. Not in theory, but in practice, for UK SMEs navigating a landscape where the old models are breaking and new capabilities, particularly AI, are creating options that didn’t exist two years ago. The question isn’t whether you need better finance. The question is: what shape should that finance function actually take?
What you will learn
- Why most SME finance functions were never designed, and why that starts to matter at £3–5m.
- Where the traditional bookkeeper-accountant-MD model breaks down, and what the warning signs look like in practice.
- The three finance function models — in-house, outsourced, and hybrid — and an honest assessment of when each genuinely fits.
- What AI actually changes in the finance function today, and what remains irreducibly human.
- A practical, stage-based approach to designing the right finance architecture for where your business is heading, not just where it’s been.
Why the Old Model Breaks at Scale
Let’s be honest about the model most SMEs are running. A bookkeeper who knows the systems inside out. An accountant who delivers solid year-end numbers. And an MD who’s become the de facto Finance Director by default, not because they wanted the role, but because nobody else could fill the gap.
At £1–2m, that arrangement is often perfectly adequate. The transactions are manageable, the decisions are relatively contained, and the founder’s instinct covers most of the strategic ground. But businesses don’t stay at £2m. They hire. They take on bigger contracts. They open a second site or enter a new market. And every step up in complexity places a new demand on the finance function that it was never built to handle.
The failure points are specific, not abstract. Decisions get delayed because nobody owns the financial narrative — the bookkeeper produces the numbers, the accountant interprets them once a year, and the MD tries to fill the gap in between. Cashflow surprises appear because forecasting is backward-looking: you know what happened last month, but you can’t see what’s coming next quarter. Pricing defaults to instinct because margin analysis doesn’t exist at product or client level. And here’s the part that catches people out: these gaps don’t stay isolated. Poor forecasting leads to poor hiring decisions. Poor hiring decisions lead to margin erosion. Margin erosion leads to cashflow stress. One structural gap cascades into three or four decision failures, and by the time the damage is visible, the root cause is buried. We’ve written in more detail about the hidden financial risks that growing SMEs tend to overlook, and the pattern is remarkably consistent.
At what point does the person running the business stop being the leader and start being the finance department?
That question lands differently depending on who you ask. But for most MDs we sit across from, the honest answer is: it happened a while ago, and they’re not sure how to reverse it. They know something isn’t right. They can feel it in the late nights, the unanswered board questions, the growing distance between what the numbers say and what their gut tells them. But the gap between instinct and clarity is precisely the space where the wrong decisions get made, or worse, where no decision gets made at all.
Three Models, and When Each Genuinely Fits
There are, broadly, three ways to structure a finance function. None of them is universally right, and all of them can fail if they’re chosen for the wrong reasons. What matters is not which model you pick, but whether you picked it deliberately, based on where the business is heading, or simply inherited it because that’s how things have always worked.
The fully in-house model means an employed FD or CFO, an internal finance team, and all capability under one roof. It gives you proximity, continuity, and complete control. For businesses with genuine complexity — multi-entity structures, international operations, significant regulatory obligations — having a full-time senior finance leader embedded in the business can be exactly right. But it costs £120–180k fully loaded at the FD level alone, before you count the team underneath. For many SMEs in the £3–10m range, that’s paying for capacity the business can’t fully use. And there’s a subtler risk: an in-house FD can become absorbed into the operational rhythm, losing the external perspective that makes strategic finance leadership valuable in the first place.
The fully outsourced model sits at the other end. A fractional or outsourced FD provides strategic leadership without the overhead. Bookkeeping and accounts are outsourced too, keeping the internal footprint lean. This works exceptionally well for businesses that need high-quality financial thinking but don’t generate the volume or complexity to justify a full-time hire. The risk is disconnection: if the outsourced FD isn’t embedded enough in the rhythm of the business, they can become a voice on a quarterly call rather than a partner in the decisions that shape the month ahead.
The hybrid model is where most growing SMEs eventually land, whether they plan to or not. In-house operational finance — a bookkeeper, perhaps a management accountant — handles the day-to-day: transactions, reconciliation, payroll, the mechanics that keep the engine running. Strategic leadership comes from an outsourced or fractional FD who owns the financial narrative, connects management information to decisions, and brings cross-sector experience the business couldn’t access any other way. Specialist capability — tax structuring, funding advisory, systems architecture — is brought in as needed rather than carried permanently on the payroll. It offers the best of both worlds, but only if the roles, handoffs, and reporting lines are properly designed. A hybrid model without clear accountability is just a collection of people doing finance-adjacent work without anyone steering the ship.
We should be honest here. Sapien is an outsourced FD partner, and the hybrid model is our natural territory. But we’ve worked with businesses at every stage, and some of them genuinely need a full-time CFO. That’s not a failure — it’s a signal that the business has grown into a level of complexity where permanent, in-house strategic leadership is the right architecture. We’ve helped businesses make that transition, and it’s one of the outcomes we’re proudest of. The point was never to be permanent. The point was to be right for the stage.
The model matters less than whether it was chosen or inherited. A deliberately designed finance function at £4m will outperform an accidental one at £10m, every time.
What AI Actually Changes, and What It Doesn’t
This is the section where it would be easy to get carried away. So let’s not.
AI is already changing parts of the finance function, and those changes are real. Automated transaction coding, anomaly detection in spend patterns, cashflow forecasting that learns from historical variance, even elements of management reporting that can be assembled without a human touching a spreadsheet. For the operational layer of finance — the transactions, the reconciliation, the routine reporting — these tools are genuinely useful, and they’re getting better quickly. Any SME ignoring them is leaving efficiency on the table.
But there’s a line, and it matters.
AI is very good at spotting that your debtor days have crept up by four days over three months. What it cannot do is sit down with your sales director and work out whether that’s a client relationship problem, a contract terms problem, or a sign that you’re chasing the wrong kind of work entirely. It can flag that your gross margin has shifted. It cannot tell you whether that shift is a pricing failure, a delivery inefficiency, or a strategic pivot that hasn’t landed yet. The interpretation — the commercial judgement applied in context, under pressure, with incomplete information and real consequences — remains irreducibly human. That’s not nostalgia. It’s just honest.
The practical implication for SMEs is this: AI doesn’t eliminate the need for financial leadership. It changes where that leadership spends its time. The operational layer gets leaner and faster, which is welcome. But the strategic layer — the interpretation, the narrative, the decisions that move the business forward — becomes more important, not less. If anything, AI raises the bar for what an FD needs to deliver, because the basic reporting is no longer the bottleneck. The bottleneck is meaning.
And here’s where it connects back to the design question. The businesses that will benefit most from AI in finance are those with a clear architecture — where it’s obvious which tasks are operational and automatable, and which are strategic and need a human mind. Without that clarity, AI just makes a messy function faster at being messy. Automating a broken process doesn’t fix it. It accelerates it.
Designing — or Redesigning — What You Actually Need
So how do you get from here to there? Not in one leap. And not by starting with the org chart.
The approach we take, and the one we’d recommend regardless of who delivers it, starts with the decisions, not the roles. What does the leadership team need to decide each month, each quarter, and each year? What financial information supports those decisions? Who is responsible for producing it, interpreting it, and, critically, challenging it? Once you can answer those questions clearly, the structure of the finance function starts to reveal itself. You’re no longer guessing at roles. You’re designing around outcomes.
As a rough framework — and it is rough, because context always matters more than thresholds — we see a fairly consistent pattern by stage. At £1–3m, most businesses need a strong bookkeeper, a reliable accountant, and periodic FD input for planning, forecasting, and the decisions that carry real weight: pricing, hiring, investment. At £3–8m, a fractional FD becomes essential. Someone needs to own the financial narrative, connect the reporting to the decisions, and bring the external perspective that the internal team can’t provide. This is the stage where most businesses are under-served — too complex for the old model, not yet large enough to justify a full-time CFO. If you’re navigating this transition, we’ve written about building financial controls without a full finance department as a practical companion to that journey. At £8m and beyond, or where the business is approaching a funding round, an acquisition, or a significant structural change, the question becomes whether hybrid or fully in-house is the right architecture. The answer depends on complexity, sector, and ambition. There’s no formula, but experience narrows the odds.
These stages aren’t rigid. A £4m business preparing for Series A funding may need more strategic finance leadership than a stable £12m business in a mature sector with predictable cashflows. A professional services firm at £6m with thirty staff, variable utilisation, and complex project accounting has fundamentally different needs from a £6m e-commerce business with high transaction volume and thin margins. The stage matters, but so does the shape of the business underneath.
Three questions worth asking before you redesign your finance function
How quickly can your finance team give you a number you’d stake a decision on? Not a rough estimate from memory — an actual, defensible figure.
Who in your business currently owns the financial narrative — the story that connects the numbers to the strategy — and is that the right person?
And if you were building this business from scratch today, would you build the same finance function you have now? The gap between that answer and your current reality is where the design work begins.
From Accident to Architecture
The goal here isn’t to find the perfect model. Perfection is a distraction. The goal is to move from a finance function that happened to one that was designed — deliberately, around the decisions that matter, with the right blend of in-house capability, external expertise, and technology.
That shift, from accident to architecture, is quieter than it sounds. It doesn’t require a transformation programme or a six-month project. It usually starts with a conversation. Someone sits down, looks clearly at what exists, and asks: is this what we’d build if we were starting today? More often than not, the answer is no. And the relief in the room when someone finally says it out loud is real.
If that feels familiar, it might be worth a conversation. A short one can often clarify what months of internal debate cannot.