An honest reflection on three decades of promises, progress, and the stubborn persistence of the same old problems.
Thought Leadership · Founder Perspective · 9 min read · dataline.com.au
Dataline was founded in 1987. When we started, ‘automation’ in finance meant running batch processing overnight and hoping the output files were clean by morning. The idea that a machine could read an invoice, match it against a purchase order, route it for approval, and post it to a ledger — in real time, at scale, with 99.9% accuracy — would have seemed implausible to most finance professionals of that era.
Thirty-five years later, that is exactly what we do. And yet, if you walked into a mid-sized Australian business today and asked to see their AP process, there is a reasonable chance you would still find someone manually keying invoices into a system, chasing approvals by email, and reconciling the ledger by hand at month end.
That gap — between what the technology can do and what organisations are actually doing — is the most important story in finance automation. This piece is an attempt to tell it honestly.
What has genuinely changed
The technology is no longer the barrier
In the early years of finance automation, the technology was legitimately hard. Document capture was unreliable. ERP integrations were bespoke, brittle, and expensive to maintain. Cloud infrastructure did not exist. Getting a system to work reliably required significant technical investment and specialist expertise that most finance teams simply did not have.
That barrier is gone. Cloud-native AP platforms integrate with major ERPs in weeks, not months. AI-driven document extraction can handle a wider range of invoice formats than any human team. Workflow automation is configurable by non-technical users. The technology cost has dropped dramatically. The implementation time has compressed. The capability has expanded.
The technology is not the reason finance automation has been slow to penetrate the market. Something else is.
The accuracy story has fundamentally changed
When Dataline started, data accuracy in automated processing was a genuine concern. Early OCR technology was inconsistent. Structured documents were manageable; unstructured ones were not. The answer, for many years, was human review of everything — which defeated much of the efficiency gain.
Today, the combination of AI extraction and human validation bureaus — the model Dataline has built and refined over decades — delivers 99.9% data accuracy at scale. That is not a benchmark figure. That is a repeatable operational outcome, achieved across more than 10 million invoices and 100 million financial transactions. The accuracy problem, for organisations that choose the right approach, is solved.
AI has changed what is possible at the exception layer
Historically, the hard problem in AP automation was not the easy invoices — it was the exceptions. The invoice that doesn’t match the PO. The supplier who sends in a format no one has seen before. The multi-currency transaction with a tolerance question. Humans handled these, slowly and inconsistently.
Today, AI agents can be deployed specifically to handle exception management — identifying the nature of the exception, suggesting resolution, escalating appropriately, and learning from patterns across the exception history. This is a qualitatively different capability from anything available a decade ago, and it is still being absorbed by the market.
What has not changed
Organisations still underestimate the process problem
In 1987, finance automation projects failed because the technology was not ready. In 2026, finance automation projects fail because the process was not ready — and the organisation expected the technology to fix it. This is the single most consistent pattern we have observed across three and a half decades.
Automation accelerates the process you have. If that process is inconsistent — irregular PO usage, fragmented approval structures, a supplier base that sends whatever format they feel like — automation will make those inconsistencies faster and more visible. The organisations that achieve the best automation outcomes are the ones that did process work first: standardised their PO usage, cleaned their supplier master data, defined their approval matrix clearly. That discipline is still the exception, not the rule.
Change management remains the most underinvested phase
Finance teams can resist automation longer than other operational functions. Not because they are Luddites — but because they are professionals with deep institutional knowledge, and they have watched enough ‘transformation’ projects promise to simplify their work and deliver complexity instead.
The organisations that navigate this well are the ones that treat their AP team as partners in the implementation, not subjects of it. They involve the team in the project design. They acknowledge what the team knows that the system does not. In 35 years, that approach has never failed to produce a better outcome.
The gap between what is possible and what is deployed is still enormous
Across the Australian market – and globally – there is a vast number of organisations running AP processes that are years behind what the technology can offer. Not because they cannot afford to modernise. Not because the technology is too complex. But because no one has prioritised it, no one has championed it internally, or a previous implementation disappointed and the organisation retreated.
This is not a criticism of those organisations. Finance automation is not glamorous. It does not have the profile of a CRM implementation or an ERP upgrade. AP is the back of the back office. But the financial impact of modernising it — reduced cost per invoice, faster month-end close, better working capital management, stronger supplier relationships — is as significant as almost any other operational investment a CFO can make.
The one thing that surprises us about 2026
After 35 years, the thing that surprises us most is not how much the technology has advanced — that trajectory was always clear. It is that the fundamental human and organisational dynamics around finance automation.
The underinvestment in process design is the same. The tendency to select technology before defining outcomes is the same. The surprise when a working system does not fix a broken process is the same.
What has changed is our ability to address those dynamics directly. From a flexible configurable AP automation software solution to a managed service model — AP as a Service — now changes the proposition entirely: The organisation now simply need to be clear about the outcomes they want. That is a fundamentally different conversation from the one we were having in 2005.
What we think the next decade looks like
The next decade will not be about technology capability. The platforms already exist to automate the vast majority of accounts payable, procurement processing, and financial reconciliation work. The question is adoption — and the barrier to adoption is no longer cost or complexity. It is organisational will and leadership attention.
We expect to see AI agents to take a much larger role in exception management and financial decision support — not replacing human judgement, but providing structured intelligence that makes human judgement faster and better. We expect PEPPOL e-invoicing to become the default in Australia within the decade, which will eliminate a significant portion of the data capture problem. And we expect that complex AP processes will require a partner that has the experience and ability to manage the required automation.
The automation that removes friction will win. The automation that removes control will be rejected — and rightly so. That has been true for 35 years. It will be true for the next 35.
A final note
Dataline has been part of Australian finance operations since 1987. We have survived four recessions, three waves of ‘disruptive technology’ that were going to make us obsolete, and the complete transformation of the technology stack we work with. The reason we are still here — and still growing — is not that we have always had the best technology. It is that we have always cared more about the outcome for our customer.
That is still what we do.
→ Speak to Dataline about your finance automation journey: dataline.com.au/contact