EXCLUSIVE: "The Subledger Solutıon" - Adrian Mustafa, SAP Fioneeer and Mike Dolan, EY in 'The Fintech Magazine'

Has the GL had its day? No, but if you want to future-proof your finance team and give them access to granular, real-time information, a new accounting architecture is the way to go, say SAP Fioneer and EY For decades, the general ledger has been the place where the numbers come together.

It has provided order, control and a trusted record of what has happened. But it’s becoming harder for organisations to fit every detail they need access to into that familiar structure, especially financial organisations. More products, more systems, more jurisdictions and more demanding reporting requirements have created a need to understand transactions at a granular level that the traditional general ledger was never designed to provide – and becomes bloated and unmanageable if organisations try.

Boards want to know what sits behind the numbers. Regulators want microscopic auditing. Finance teams need to be able to trace a figure back to the individual transaction that created it with as much context around it as possible. The answer is not to abandon the general ledger; but it is to stop asking it to do everything. Instead, the concept of a digital subledger is back on the agenda.

The idea is straightforward: retain the discipline and controls of the general ledger, but move detailed accounting logic and transaction information into subordinate ledgers before the complex numbers reach it. The result is a multi-tiered accounting architecture in which the general ledger provides the wider view, while subledgers support it with the detail underneath. And there is another reason this approach is now attracting attention. The same characteristics that make subledgers useful to finance – structured, granular and traceable data – are those also required to make sure artificially intelligent systems deliver on their promises. 

For banks and insurers in particular, the challenge is becoming less about producing numbers and more about knowing precisely what those numbers contain, structured in such a way that they are easily accessible to humans – and their AI assistants.

Mike Dolan, Partner at accounting giant EY (formerly Ernst & Young) helps organisations evaluate and implement product accounting subledgers and sees the daily pressures finance teams are under.

“Every year, the CFO function is being asked to do more and more and more with less and less and less,” he says. “Their budgets are going down, but their expected output is going up.”

That might mean they’re being asked to add another basis of accounting, deal with a merger or acquisition, introduce a new product or respond to an accounting standard.
“The only way they’re going to really be ableto stay ahead of all this is through investments in technology,” Dolan says.

In his view, a general-ledger-centric approach is no longer enough. The GL is fine for providing the summary view of the accounting book of record and the controlled financial statements on which organisations depend. 

“But what if I need 20 chart fields?” says Dolan. “If I try to jam all that into the general ledger, I get this fat GL syndrome where there are too many accounts, too many thingsto reconcile.”

Subledgers offer another way of handling that complexity: preserve the accounting principles and controls that finance teams trust, but apply them at a more detailed level, he says.

“Let’s take everything we like about the general ledger, but push it down to subordinate ledgers underneath the GL and help to take some of that burden away.” Adrian Mustafa, Product Director for User Innovation at SAP Fioneer – a global software and technology company that has worked with EY since April 2024 to develop Cloud accounting solutions – agrees with Dolan.

“General ledgers still fulfil the core mission, but they were designed for a different era,” he says. “An era with fewer source systems, less complexity, much lower demand for granularity.

“The subledger was built to capture, classify and control at every level of detail. This is less a question of scale, and more of the right architecture.”
As such, the subledger acts as a bridge between the operational activity of a financial organisation and its reporting. It can draw together events from core banking systems, trading platforms and other sources, apply accounting rules and create the journals that ultimately flow into the general ledger.

“It creates a structured finance layer between all of your source systems and the GL,” Mustafa explains.

That layer also gives other parts of the organisation, such as risk and audit teams, access to a consistent, shared, real-time version of the truth.

“If you ask three different colleagues, especially from different teams, for the book value of the same financial instrument, you may get five different answers [if you are solely relying on a general ledger],” adds Mustafa.

That does not necessarily mean anyone has made a mistake. Definitions, calculations and processes can simply become dispersed across systems and departments. The subledger’s role is to bring those elements together and make the resulting figure traceable.

“What is the KPI, for instance? What does it consist of? And which postings led to that number in the report or financial statement?” adds Mustafa.

That ability to trace a number back through the accounting chain becomes particularly important during month-end and year-end close because a modern subledger should go beyond showing where a problem has occurred and help finance teams resolve it, too. It can reduce the reliance on spreadsheets and manual investigation, and gives accountants a structured route from identifying an error to correcting it.

“One clear dashboard should show you the open issues, the likely causes and the next correct steps,” says Mustafa.

When they can drill down through financial statements to reach successive layers of detail – not just one or two, but a dozen clicks deep, from the statement itself, through the accounting entries and processes, back to the transaction – finance teams gain a much clearer view of how a reported number has been constructed. That creates a different relationship with the numbers. The financial statement is the end point of a long chain of calculations and reconciliations; with subledgers, organisations can make the chain itself visible.

It can also make the accounting function more flexible. New products, accounting standards or bases of accounting can be accommodated through changes to rules and transaction types rather than by redesigning an entire collection of systems. And when transparency and flexibility improve, efficiency tends to follow. Less manual reconciliation means less time spent producing and validating information, while a clearer audit trail gives finance teams greater confidence in the figures they are reporting.

While every CFO will be cheering the efficiencies, the fact that they can trust the numbers is priceless.

Context and control

So why isn’t the subledger being universally adopted, then? According to Mike Dolan, there’s a temptation, particularly as financial institutions build increasingly sophisticated data platforms, to question the need for subledgers if they have already put all the information in one data warehouse to which they can apply AI. But, if anything, says Dolan, that doesn’t diminish the need for subledgers, it reinforces it because for AI to deliver the expected results the data needs context – accounting rules, processes and controls.
“It’s tempting to think that we’ll just use big giant data dumps, put AI on top of them and everything will be solved. But you still need your system of record, and the orchestrated process to work the right way in order to generate that results output,” he says.

Dolan sees data platforms as systems of analysis rather than replacements for systems of record.

“They’re great for the results, and they’re great for being able to analyse and run analytics off, but they don’t eliminate the need for control over the business logic,” he says.
And that business logic can, after all, be highly specific to an industry. For financial services organisations, for instance, accounting is tied to valuations, calculations and changing regulatory requirements. Storing the underlying data in a warehouse does not mean an organisation has captured the rules governing how that data should become accounting information.

The subledger can sit at the point where accounting data and accounting processes meet. Rather than simply recording the outcome, it can retain the calculations, rules and transaction context that explain how that outcome was reached, providing not just finance, but also audit and control functions with a more complete picture of what sits behind the numbers.

The benefits are not limited to reporting, either. They can have a marked impact on business culture. A more structured approach can reduce the dependence on a handful of people who understand an organisation’s historical accounting workarounds.

Mustafa describes the familiar situation of the long-serving specialist who knows how the data warehouses, spreadsheets and legacy processes fit together.
“They hold the complete picture of all the complex data warehouses and Excel workbooks and accounting workarounds,” he says. “But they are essentially alone in that knowledge. By standardising the content and making everything fully traceable, it opens the door for our younger colleagues – and increasingly for AI as well.

“It bridges the gap between the long-serving experts and the new generation who want to contribute but cannot because the complexity is simply too high. That is one of the most important, and I think most underrated, effects a subledger has.”

The implementation itself can help transfer knowledge, because teams have to translate historical spreadsheet and data warehouse logic into a more structured model.
“What I always see during those projects is that teams that never really communicated with colleagues before, suddenly start collaborating,” says Mustafa.

With this in mind, Mustafa also points to the inherent difficulties in building a subledger internally over time. While accepting that a strong IT team with accounting expertise could build the basic functionality – turning a data platform into a subledger is a ‘hard task, but not an impossible one’, he says – the more challenging part is keeping it resilient as the organisation changes.

New accounting standards arrive. Products change. Business models evolve. People who understand the original architecture leave.

“What tends to happen over five to 10 years is it deteriorates,” says Mustafa. “It becomes a black box that only a few select people in your organisation understand and can maintain.”

A dedicated software platform can instead spread that expertise across a product and the organisation supporting it. That brings the discussion back to one of the strongest arguments for revisiting subledger architecture now: AI. The technology may be new, but the prerequisite is familiar – clean data.

CFOs are increasingly keen to introduce AI into finance systems, but Mustafa’s experience developing software has reinforced how difficult that can be when the underlying architecture was not designed for it.

“We’ve had to retire entire modules that were working basically fine for the last few years, simply because they were not AI-ready,” he says. AI readiness, he argues, comes down to two things. 

“First, you need a structured data model that the AI agent can work with and understand. And secondly, and this is underrated, you have to have a clean process model.”

The distinction matters. A warehouse can hold the data, but may not expose the accounting logic and process context in a form that an AI system can safely work with. That makes the subledger potentially more than a solution to today’s finance problems.  It can become part of the foundation for tomorrow’s technology.

Dolan has already seen the difference between organisations that have begun the subledger journey and those still considering it. Some clients started with investment securities, then moved into debt and derivatives, loans, insurance policies and claims. Each project expanded the use of the architecture and reinforced the benefits. Those further along are now experimenting with AI.

“Because the data is there, it’s clean,” says Dolan, “it automatically ties to the general ledger. They’ve got the drillable financials. When they put AI on top of it, those teams are really cooking with gas.”

Preparing for an AI future

For organisations that have not addressed the underlying data architecture, however, simply bolting on AI will not provide a shortcut.

“Yes, you can put AI on top of a messy environment,” Dolan says, “but it’s a garbage in, garbage out problem.”

So how do organisations make the transition without another weighty transformation project that ends up consuming years and resources? Dolan’s answer is to start small. Rather than attempting to redesign the entire balance sheet, organisations can begin with a single asset class and build from there. Investment securities might be the starting point, followed by loans, hedges or other areas where the limitations of the existing architecture are particularly apparent.

That also gives finance and technology teams time to understand the new model and work out which requirements are common across the organisation and which are specific to individual products.

“Approaching it in a piecemeal fashion gives people in the finance organisation and finance technology team the ability to learn the basic principles,” says Dolan.

It is a pragmatic answer to a problem that can otherwise become overwhelming. The aim is not to replace everything that already works, but to introduce a more appropriate layer where the existing architecture is struggling with complexity. The general ledger still has a central role.

But as the volume and complexity of financial information continues to grow, the case for putting more of the detail, control and accounting logic beneath it is becoming harder to ignore. The important thing is not to try to solve everything at once.

“We hardly ever see that as being successful,” says Dolan. “You can’t boil the ocean. But get started somewhere, get your organisation comfortable with it, realise the benefits and then let the people loose.”