7 min read
Most agencies can now point to at least one place AI has sped things up. Fewer can say with any confidence whether that speed is actually paying for itself. That gap, between “this feels faster” and “this is making us more money,” is exactly where the payback formula becomes useful. It answers the question agencies are increasingly being asked and rarely able to answer: is this AI spend actually paying back, or does it just feel like progress?
Why “it feels productive” isn’t the same as payback
Emily Hatton, founder of AI in Agencies and a guest on The Fractional CFO Show, has watched this play out across dozens of agency clients. Individual productivity gains are real and easy to spot. Whether they translate into anything an agency’s finance function can actually measure is a separate question entirely.
That gap shows up at scale too. Recent McKinsey research covered by The Register found that 80% of respondents using AI in their roles report improved individual productivity, yet only 37% of organisations attribute any earnings impact to AI at all, and just 6% qualify as genuine high performers attributing 5%+ of EBIT to it. Feeling more productive and moving the numbers are two different things, and most agencies are only measuring the first one.
The payback formula: How to tell if AI is actually paying back
This is one of the simplest tools in finance, and one of the most useful for a decision like this. Where the financial benefit is reasonably consistent, the basic payback period formula divides the initial cost of an investment by the measurable financial benefit generated each period, showing how long it takes to earn back what it cost. As CFI explains, it is often called a back of envelope calculation precisely because it does not need technical modelling to be useful. Applied to an AI tool, the question becomes: how long before the time saved, or the additional output delivered, is worth more than what the subscription and the time spent learning it cost the business?
A broader measure like return on invested capital has its place too, comparing profit against the total capital invested across an entire business, but that is a whole of business question for a CFO to track over time. At the level of deciding whether to keep paying for one AI tool, the payback formula gives a founder a faster, more practical answer.
Working through the formula for one tool
Take a research task that used to take a junior team member a full day and now takes an hour with AI assistance. If that task represents £150 of staff time each time it is completed, that is £150 of capacity released. Freed capacity is not automatically the same as money in the business though. It only becomes real value if that time is redeployed into billable work, extra client capacity, or work that would otherwise have needed a new hire. If the time saved simply gets absorbed into a quieter day, the payback formula still balances on paper but the agency has not actually banked anything. Tracking what happens to freed up time, not just measuring that it exists, is what turns the calculation into an honest one. A tool paid for over six months without that check having been made is itself worth flagging.
Start with process, not the tool
Hatton’s starting point with clients is rarely the AI tool itself. “You need to look at your processes first and then you look at the tools that we have available now to address those processes.” Agencies that skip this step end up bolting AI onto a broken process and wondering why the payback never appears. An audit of where time is genuinely being lost, or where output is being under-delivered against what clients now expect, has to come before any tool decision.
Set a goal, and a realistic timeframe, before you roll anything out
The other habit separating agencies who can run this calculation with confidence from those who can’t is goal-setting before rollout, not after. Hatton is blunt about what happens without it: “you should set goals on your use of it because otherwise it’s just very unstructured use and processes on top of processes, which means actually you’re likely to be eating into your profit margin rather than improving it through using AI.” A goal tied to a specific number, hours saved per week, output increased, needs a realistic timeframe attached to it too. Some AI efficiencies show up within weeks, though this varies by tool and by how quickly a team actually adapts, so checking in monthly rather than waiting for an annual review catches problems while they’re still fixable.
Change management
Rolling out AI well is a change management exercise like any other, not a special category of its own. Some team members will be enthusiastic from day one and others will be resistant, in the same way any new system meets resistance regardless of how much better it makes things. Training, structured time to experiment, and identifying the people on a team who are genuinely enthusiastic about a new tool all matter more than the tool’s feature list, because those early adopters are usually the ones who bring the rest of the team along. Skipping that groundwork is one of the quieter reasons the payback formula never balances: the tool gets bought, a handful of people use it properly, and the rest quietly revert to the old way of working.
The same discipline ACC has written about for testing marketing spend against a control group applies here: assume nothing works until it has been measured.
What this means for agency profit margins
The bigger point, in Hatton’s view, is that AI should be a lever on margin, not just a productivity anecdote. “It should be increasing every agency’s profit margin.” That only happens with deliberate tracking.
In the same way ACC has argued that disciplined pricing strategy protects margin as agencies scale, AI spend needs the same scrutiny: tracked against a goal, measured with a proper payback formula, and judged on what it has actually freed up.
One book worth your time
Emily recommends Traction by Gino Wickman as a practical framework for agencies that need to move fast without losing structure, particularly its approach to running focused weekly meetings and setting clear goals tied to outcomes.
ACC’s CFO Perspective
Over the last year or so, we’ve seen agencies investing steadily in AI tools without ever tracking whether the spend is paying back. Smart agency owners understand that the same discipline must be applied to AI as to any other capital decision. Set a goal, set a timeframe, and check the return on invested capital before scaling spend further. A fractional CFO brings that discipline as standard, not as an afterthought.
AI is not going to make an agency more profitable by default. It becomes a genuine driver of profit margin only when its payback formula is checked with the same rigour as any other investment decision.
Frequently asked questions
What is the payback formula for an AI tool?
Divide the cost of the tool, including time spent learning it, by the value it returns each period, whether that’s time saved, output increased, or extra billable capacity. The result shows how many weeks or months it takes for the investment to earn back what it cost.
How long should an agency wait before judging whether AI is paying back?
It varies by tool and by how quickly a team adapts, but many AI efficiencies show up sooner than a typical capital investment, sometimes within weeks. Checking in monthly rather than annually makes it easier to catch problems, such as low adoption or unclear processes, before they compound.
Should AI adoption sit with the operations team or the finance function?
Both. Operations teams identify where AI genuinely saves time or increases output, but connecting that back to profit margin and a proper payback calculation needs financial oversight, in the same way marketing spend needs a finance lens to prove real ROI.
What’s the biggest reason agencies fail to see a return from AI?
Skipping process review beforehand and rolling out tools without a specific, measurable goal. AI applied to a broken process produces faster busywork, not a better payback.
Author
ACC Finance are a team of experienced fractional CFOs and management accountants who combine executive financial leadership with practical commercial judgement to work closely with founders and leadership teams to strengthen margins, improve cash flow, and guide critical financial decisions.
Apply for a Financial Health Check
Gain independent clarity on profitability, cash flow, and financial controls before your next stage of growth.
Applications are reviewed to ensure a strong fit.
