- AI tools for finance teams
- Back office AI disruption
- How CFOs are using AI
In 1985, Microsoft Excel was disruptive.
A finance team could build a cash flow model without a calculator or a filing cabinet. It was one of the most significant shifts the back office had ever seen and people were excited about it. (Imagine a world where you were actually excited to open a spreadsheet…)
Forty years later, 89% of finance teams still use Excel as their primary tool. The CFO's back office has run on a myriad of legacy ERP systems, spreadsheets, and manual processes held together by institutional knowledge and human effort.
But we are at the point of another disruptive shift. AI is rewriting the operating model of the finance function at the structural level and a new tension is emerging. CFOs know AI adoption is no longer optional, but transforming legacy systems remains the single biggest barrier to change, compounded by analysis paralysis from the volume of AI tooling now available.
We talk to CFOs every day and based on the insights we are gathering on how teams are executing with AI - we have formed six predictions on how AI will disrupt the back office first and where teams need to be investing to remain competitive.
1. Accounts payable will be fully autonomous in 12 months
Accounts payable has always been an unglamorous function as a cycle of invoice chasing, manual data entry, reconciliation errors, and timing mismatches. That tedium is precisely what makes it the process susceptible to disruption.
Best-in-class organisations are now processing invoices in 3.1 days at $2.78 per invoice, compared to an industry average of 17.4 days at $12.88, reflecting a 78% cost reduction driven almost entirely by automation.
The reduction is dramatic because AP is fundamentally a rules-based system. Every step follows a defined logic, from document capture, data entry, three-way matching, exception handling, approval routing and payment scheduling. Rules at scale are exactly what AI does best. Robotic Process Automation (RPA) eliminates manual document handling. Invoice automation removes data entry errors. AI validation reconciles invoices against purchase orders and contracts in seconds, catching discrepancies before they become costly mis-payments. Fanatics Betting & Gaming's CFO Andrea Ellis described reducing their vendor journal entry process from 20 hours to 2 hours each month, from adopting Generative AI tools.
28 days of annual savings
The disruption for CFOs: Maintaining a traditional, manual AP function will soon be the equivalent of running a paper-based filing system in a cloud-first world. The cost centre won't be justifiable and the AP manager role as we know it won't exist in five years.
2. The best use of AI will be finding money you are losing
Although automating AP will save organisations months of work, the most under-hyped use case is anomaly detection and leakage identification: AI surfacing cash that was silently haemorrhaging through contract non-compliance, miscategorised spend and duplicate payments.
This is the commercial upside in automation: consider a global biotech company introduced invoice-to-contract compliance using agentic AI. The system ingests contracts and invoices year round, checking that terms like early payment discounts, tiered pricing, and volume rebates are correctly applied. It uncovered contract leakage of approximately 4% of total spend. For a business with $1 billion in procurement, that's $40 million in recurring annual margin improvement.
The same logic extends to opportunity cost. Lyka's CFO and COO, Gabriel Guedes, explained this in relation to sourcing investment opportunities for their idle cash: friction in manual investing processes means the default is not to invest. When you're talking about roughly 5% on a series round, that inaction is its own loss.
The disruption for CFOs: Within 12 months, the CFOs generating the highest measurable ROI from AI will be those who pointed AI at their spend data and procurement contracts. This is the use case that turns AI from a cost centre into a profit driver.
3. Cash forecasting will be replaced by continuous intelligence
The leakage problem above is compounded by a forecasting problem. Even when finance teams know their numbers, they often don't know them fast enough. As one CFO put it at a recent industry roundtable:
We say it takes a quarter to forecast a quarter.
Finance teams waste hundreds of hours manual cash forecasting, and the outputs are often unreliable. As one of the Primary clients recounted from their experience working with middle market companies in M&A: revenue guidance would be well thought out and typically land within close range of accuracy. Cash forecasts, by contrast, were either never created due to bandwidth constraints, built of flawed assumptions or wide of the mark entirely.
The conditions of the past five years have made cash forecasting hard. Before 2020, stable interest rates and predictable markets meant backward-looking forecasts could be reliably used as inputs for forecasting. However, since 2021, financing costs have more than doubled and inflation has reshaped input costs across every sector. Geopolitical instability also remains completely unpredictable. The result is that static financial models can’t keep up with how quickly the external environment impacts them.
There is going to be a shift from periodic forecasting to continuous financial intelligence. AI models will ingest real time bank feeds, ERP data, AR and AP positions, and external market signals to produce a live view of liquidity. Early AI-powered forecasting models are reducing error rates in cash predictions by up to 50% compared to traditional methods, simply by processing more data at a continuous rate.
The disruption for CFOs: Companies will need to automate cash management and use tools that factor in or predict external volatility to survive in the current climate. In an environment where a single timing mismatch can trigger a liquidity event, not investing in real-time forecasting tools may become an existential risk.
4. Treasury management becomes a profit centre, not a cost centre
If continuous cash intelligence tells you where your money is, smart treasury management puts it to work. This is the prediction we're most bullish on: treasury will become a strategic differentiator.
Under traditional models which typically involves managing multiple banking relationships, manually tracking cash, investments, and FX trades, finance teams can leave millions on the table every year. The friction of setting up a new banking relationship for better yield or FX rates is high, that often it becomes an opportunity cost.
Technology is collapsing that friction. One Primary customer noted that onboarding a new banking relationship for idle cash investment took 3 days, compared to a typical 3 weeks to set up a new term deposit. Another customer, Nexl, generated an extra 3 months of runway through access to best-in-market yield rates.
Here’s the math: a $4M cash balance earning 4.5% institutional yield generates an additional $140,000 AUD annually, compared to $17,500 at a traditional bank rate. That directly translates to extended runway, additional headcount, or reinvestment capital. This is treasury management as a direct profit centre.
The disruption for CFOs: Investment in your treasury function could be one of the highest-yielding, yet easiest integrations available. For finance functions to succeed in 2026, the mindset needs to shift from marginal cost savings to revenue accretion and without modern treasury management systems, the revenue growth opportunities may remain invisible.
5. “Month-end close" becomes obsolete
Autonomous AP, spend intelligence, continuous forecasting, and active treasury management all point to a single structural conclusion: the traditional month-end close is becoming obsolete.
Finance teams can deeply resonate with the frustrations of the month end financial close - a compressed, stressful sprint of reconciliation, accruals, and reporting. But integrated, automated systems will fundamentally eliminate this process, replaced by a continuous close, where AI agents reconcile daily, prepare entries weekly, and analyse trends in real time, leaving only final adjustments at period-end.
What makes this prediction non-obvious isn't the speed gain, but the structural change in how organisations need to operate. Finance teams achieving continuous close in 2026 are adopting and training agents to handle execution, while humans focus on supervision, governance, and strategic analysis.
EQL's head of finance is seeing first-hand how AI is reshaping the structural ways organisations operate. The company is an example of the shift away from excel, using Primary instead to view real-time AUD and USD balances across two entities in a single view. The prime motivator is to have a continuous consolidated cash view so that leadership can be more efficient at decision making at any given time.
The disruption for CFOs: The finance teams reporting calendar is going to start to look very different. The close process shifts from a periodic event to a continuous state with key decision makers demanding insights when decisions arise, rather than when month end occurs.
6. The CFO who ignores AI governance will become a liability
As AI takes on more autonomous decision-making across AP, treasury, forecasting, and the close, a new risk emerges: who governs the machines?
Recent discussions with CFOs have highlighted that there is a high level of confidence that AI will deliver measurable ROI in 2026, but some of the key execution blocks are prioritisation and talent and capacity. Most companies are lagging in the creation of mature governance systems and processes for autonomous AI, even though the deployment is a top priority for most organisations.
This matters for many reasons, but one of them is that AI will be producing work that helps inform important decisions and in some cases, will automate them. When those decisions sit within regulated financial processes, the accountability still falls on the CFO. Decisions made on biased or ungoverned models create regulatory and reputational risk that lands squarely on the CFO's desk.
The disruption for CFOs: The CFO's emerging role isn't just to deploy AI, but to govern it and the finance function that is using it. Without governance frameworks, the biggest risks organisations will face will revolve around compliance and trust.
The new era of disruption
In 1985, Excel didn't replace finance teams but gave them the ability to execute at greater scale and speed. The calculator didn't disappear, but the people who refused to open a spreadsheet eventually did.
We're at that same inflection point. AI won't eliminate the finance function, but it will fundamentally reshape every role within it. It will be crucial for CFOs to be on top of key AI transformations to ensure their organisation can remain competitive.
The difference this time is pace. The disruption of 1985 gave us forty years of Excel dominance. The disruption of 2026 won't wait nearly that long to separate the leaders from the laggards.
About Primary
Primary provides modern treasury management solutions for complete cash visibility, idle cash optimisation, and FX risk management - all in one platform.
