Finance’s AI Opportunity Is Action, Not Just Insight

Beyond the Dashboard

Over the past few years, I have watched AI take hold of the finance function, first in transactional accounting, then in reporting, and now in forecasting, right up to board reporting and the analyst packs we prepare for regulators.

The progress is real. But one area I keep coming back to, and I am genuinely intrigued by, is a concept I have come to call the Growth Lever Model: a financial model that doesn’t just tell us what’s likely to happen, but tells us what to do about it.

That distinction matters more than it sounds. Most of what AI has done for finance so far makes us better at describing the business. It tells us what happened and what will probably happen next, but it hasn’t changed what the business actually does differently as a result.

Where AI has already delivered

In my experience, AI has moved the needle for finance functions in five areas over the past few years:

Predictive analysis: turning company data into meaningful reporting and running forecasts off historical performance.

Transactional automation: a significant reduction in manual tasks and streamlining the leakage that sits in working capital and lifting gross margin as a result.

Financial Reporting: significant improvement in monthly, board and regulatory financial reporting.

Cashflow predictability: forecasting shortfalls and risks before they occur.

Large-scale data analysis: doing in minutes what used to take a team days.

Underneath these, I would point to three fundamental benefits finance has realised from AI adoption:

Improvement in finance operations. Automating AP, AR and procurement has taken a real load off transactional teams, cut manual error rates, and closed off margin leakage that used to slip through in contracts almost unnoticed.

Improvement in financial and board reporting. Reporting has moved from a monthly scramble to something closer to continuous. The gap between a transaction happening and it showing up, properly categorised, in a board pack or regulatory filing keeps shrinking.

Improvement in budgeting and forecasting models. What used to be a static annual forecast is now a live, driver-based model that flexes as historical performance comes in, with far less manual rework.

These are real wins, and I don’t want to undersell them. But they’re all, at heart, about producing better outputs: better numbers, faster, with less effort. None of them change what the business does differently.

Introducing the Growth Lever Model

Here’s the idea I keep sitting with, and what I have started calling the Growth Lever Model: a financial model, sorry, a business model, that focuses on changing how the business operates in order to deliver the outcome it wants, rather than just forecasting what will probably happen.

Today’s driver-based models are good at giving us a number: revenue growth of 25 percent, or 15 percent, or minus 5 percent, depending on the assumptions someone dials in. What if, instead, the model worked backwards from the growth target the business actually wants, and told us the operational plan to get there: where to focus sales effort, which conversion approach to use, what KPIs to set for which teams, and why we missed the target last time, financially and operationally?

Two examples show what this could look like:

Example 1: A Professional Services Firm

Take a mid-size accounting or law firm. It builds a predictive financial model that links revenue to utilisation rates, chargeable hours, and headcount by role: partner, senior associate, junior. Monthly revenue is modelled as billable hours × realisation rate × charge-out rate, flexed for seasonality (tax season spikes, for instance) and planned hires or attrition.

That’s a solid forecast. It gives the partners a revenue number they can adjust by changing the drivers. Apply the Growth Lever Model to it, and the firm gets a concrete operational plan to deliver 25 percent growth, consistently. As an example, it might suggest:

• Marketing on the platforms that have historically generated the best leads.
• Converting those leads using the approach that’s worked in the past.
• Setting specific KPI targets for each team.
• Explaining why 25 percent growth wasn’t achieved last time, and being clear that the reasons aren’t just financial, but operational too.

Example 2: A Dental Group

A dental group builds its model around chair utilisation: chairs × hours available per chair × fill rate × average fee per appointment. This is split by clinician type (principal dentist, associate, hygienist), since fee-per-chair-hour varies a lot between them.

What if the model could work out how many leads are needed to fill the appointment book required to hit that revenue-per-chair target, and then translate that straight into an operating plan? It could:

• Adjust the marketing and sales effort feeding the booking funnel.
• Adjust the clinic roster to match predictable patient inflow, fortnight by fortnight.
• Allocate clinicians to the revenue mix the business needs: general, cosmetic, and so on.
• Pre-order the lab materials needed for the resulting treatment schedule.

In both cases, the model stops being something finance produces for the business to look at, and becomes something the business runs on, capable of setting weekly or monthly targets and adjusting its own drivers as things change. That, to me, is what the Growth Lever Model really means: finance finally using AI to deliver outcomes throughout the business, not just outputs from finance.

Why this matters now

This isn’t just a nice idea. The data backs up the urgency. McKinsey’s 2025 State of AI survey found that nearly two-thirds of organisations haven’t yet begun scaling AI across the enterprise, and only 39 percent report any impact on earnings at all, most of those under 5 percent.^1 The businesses pulling ahead aren’t running more pilots. They’re redesigning workflows around AI and pointing it at growth, not just efficiency, which is exactly the mechanic behind the Growth Lever Model.

I think this is one of the few AI opportunities finance is uniquely placed to own. No other function sits across the full data set of the business, from revenue drivers to cost structure to cash, the way finance does.

The opportunity for finance leaders

Finance has spent the AI era so far getting better at describing the business. I think the next stage is using that same data and modelling capability to help run it: turning forecasts into operating plans, and operating plans into the growth the board is asking for. That’s the promise of the Growth Lever Model: a bigger mandate than finance has traditionally held, but one that’s genuinely within reach now.

---

Author: Sandeep Bordia

Sandeep is senior finance executive with over 14 years of experience as a CFO, Financial Controller and Director across high-growth, PE-backed, listed and government organisations, including Fairfax Media/Domain Group, Compare and Connect, City of Melbourne and Carnival Australia. 

Across his career, Sandeep has led M&A transactions and divestments valued at up to $140M, raised more than $20M in debt and equity funding, and built finance functions that turn financial complexity into clear, commercial decisions. His experience spans capital raising, transaction execution, governance and business transformation across technology, media, utilities, professional services, government and education sectors.

---

^1 McKinsey & Company, “The State of AI in 2025: Agents, innovation, and transformation,” QuantumBlack, November 2025.