There are two clocks running inside every employer.
One belongs to your HR strategy. It ticks faster every year, accelerated by AI, by talent market pressure, by every new generation's expectations of work, by the pace at which your competitors are reinventing how they hire, develop and retain people.
The other belongs to your payroll obligations. It ticks slowly, on the rhythm of legislative drafting, gazettals, transitional rules and ATO guidance. Sometimes it stops for years. Then a Payday Super lands, and it ticks loudly for eighteen months.
For most of the last twenty years, nobody cared that these two clocks ran at different speeds. Both were slow. Picking a single vendor for HR and payroll was just common sense: one contract, one data model, one throat to choke.
In 2026, that's no longer true. And the buyers who haven't noticed are about to spend the next three years paying for it.
The cycles have diverged
The most recent data from the Sapient Insights Group HR Systems Survey, the long-running annual benchmark of how organisations actually buy and replace HR technology, shows something that wasn't true five years ago. HR system replacement intent has surged. In enterprise organisations (1,000+ employees), the share of buyers planning to replace their HR Management System in the next twelve months more than doubled in a single year. Across all organisations, roughly one in ten has already decided to replace its HRMS, and another one in twelve has an active RFP.
Sapient's lead researcher, Stacey Harris, attributes this to a fundamental change in friction: the cloud has made it easier to switch HR systems than ever before, and AI is forcing buyers to ask whether the systems they bought four years ago can still carry them.
The HR clock
Strategy-driven. Replacement triggered by ambition, AI capability gaps, and CHRO-led transformation. 12–18 month innovation cycles.
The payroll clock
Risk-driven. Replacement triggered by compliance failures, on-prem end-of-life, or cloud migration. 5–10 year architectural rhythm.
Payroll moves on a different clock. Replacement intent in payroll has historically been high; Sapient's recent editions consistently report between a quarter and a third of organisations planning to replace payroll within two years, but the reasons are entirely different. Payroll replacements are triggered by compliance failures, legacy on-premises systems reaching end of life, vendor consolidation, or cloud migration. They are rarely triggered by ambition. Payroll changes are risk-driven; HR changes are strategy-driven.
Once you see the asymmetry, you can't unsee it.
What's making it worse, and what's making it permanent
Two things are widening the gap in 2026, and both are structural.
AI is rewriting the HR layer
The HR/talent layer is the epicentre of AI disruption in enterprise software. Recruiting, candidate assessment, internal mobility, skills inference, learning content, performance feedback — every one of these is being rebuilt around generative and agentic AI on cycles of twelve to eighteen months.
Gartner is forecasting that by 2027, three-quarters of hiring processes will include certifications and tests for workplace AI proficiency. Workday has spent more than $2 billion in disclosed acquisitions over the last year — Paradox, Sana, HiredScore, Evisort — buying its way into AI talent capabilities it could not build internally fast enough.
When the world's most successful integrated HCM vendor concludes that it cannot innovate fast enough on the talent layer, that is information you can use.
Payroll's relationship with AI is different. AI in payroll is real, valuable, and entirely back-office: anomaly detection, pre-submission validation, assistance with award interpretation, and employee self-service chatbots. Useful. But it does not redefine what payroll is. Payroll's first-order metric — was every employee paid the right amount, on time, with the right tax and super, under the right award? — does not get strategically reframed by AI the way recruiting does. Payroll AI makes payroll better. AI for talent acquisition makes talent acquisition different.
ANZ compliance is at a decade high
In Australia and New Zealand, the compliance load on payroll between 2025 and 2027 is the heaviest it has been in a decade. Payday Super starts on 1 July 2026 and forces every Australian payroll system through per-pay-event qualifying earnings reporting, the new SuperStream 3.0 with Member Verification Request, and a seven-business-day super remittance window. The SG rate rises to 12% at the same point.
Wage theft was criminalised in January 2025, imposing personal liability on directors and shifting payroll from an operational concern to a board-level risk. New Zealand is replacing the Holidays Act 2003 with a new hours-based regime under the Employment Leave Bill, with most employers facing a 24-month transition once it passes.
Health NZ alone has paid roughly $911 million in Holidays Act remediation to date, a useful number to keep in mind the next time anyone tells you payroll compliance is a low-stakes domain.
Compliance pressure does not speed up payroll. It absorbs the available capacity of every payroll vendor's roadmap, and it makes change risk-averse rather than ambition-driven. ANZ payroll in 2026 is on a slow, deliberate cadence, whether anyone wants that or not.
So, the gap between the two clocks isn't a transient phenomenon. AI is speeding up the HR clock. Compliance is slowing down the payroll clock. They are designed, right now, to be out of sync.
The integrated suite trap
Gartner has published a case study worth careful reading. An organisation chose a single cloud HCM vendor and started by deploying its two anchor modules, core HR and payroll. Two years later, digital talent management became critical to the business. The suite's learning, development, and skills analytics capabilities proved inadequate. The organisation ended up buying a separate, best-of-breed talent management suite at a substantial additional cost, then had to carry the integration complexity between the two.
The pattern is so common that Josh Bersin has written about it for the better part of a decade. Integrated HCM platforms, in his framing, run on five-to-ten-year architectural cycles, a deliberately slow rhythm shaped by the need to keep payroll and core HR stable across thousands of customers and dozens of jurisdictions. The talent and employee experience layer above the core now runs on twelve-to-eighteen-month cycles, driven by AI competition. The two cannot stay in sync inside a single product, no matter how integrated the marketing says it is.
How the trap closes
You choose an integrated suite for the operational case: one contract, one data model, simpler integrations. Two or three years later, your CHRO comes to the executive team with an AI-led talent strategy that needs capabilities your suite vendor cannot deliver on the schedule you need.
Your CFO has just signed a five-year deal. Your IT team has just finished a painful implementation. You bolt on the best-of-breed module. You now have the integration problems you signed up for a suite to avoid, the licence cost of two products, and a change conversation you cannot have for another three years.
What good buyers are doing now
The buyers we see making good choices in the ANZ mid-market are no longer asking "integrated suite or best-of-breed?" as a binary. They're asking a sharper question: what runs on which clock, and what should be coupled to what?
The answer most of them are landing on looks like this. Payroll sits on its own clock, chosen for jurisdictional depth, compliance discipline, award and EBA interpretation, reliability under the Payday Super stress test. HR, talent acquisition, learning, performance and employee experience sit on a faster clock, chosen for AI roadmap, employee experience, and the ability to change in two years if the strategy changes. The two are connected by published, well-documented, two-way API integrations, the boring, professional plumbing that lets each layer move at its own pace.
The two-clock architecture
- Payroll layer: chosen for jurisdictional depth, compliance discipline, and stability under regulatory stress.
- HR / talent layer: chosen for AI roadmap, employee experience, and the ability to swap out in two years if strategy moves.
- Integration layer: published, well-documented, two-way APIs — the boring, professional plumbing that lets each side keep its own time.
This is not a rejection of integration. It is a rejection of the idea that integration must come from a single product. The Australian Payroll Association's 2025 industry survey found that poor integration between HR and payroll systems is the single most-cited operational pain point, ahead of legislation, skills shortages, and award interpretation.
That is real, and it has to be solved. The good news is that in 2026, it can be solved with documented APIs and disciplined contracts between systems, rather than squeezing both functions into a single product whose architecture cannot serve them both well.
When bundling still makes sense
We're not arguing that integrated suites are wrong for everyone. They aren't. For roughly 150 employees in a single jurisdiction, with a simple award structure and a modest talent program, a well-built, integrated platform like Employment Hero or Xero Payroll is genuinely simpler and cheaper. The integrated story works at that scale because payroll is, effectively, the strategic HR system. There isn't a separate clock yet.
Bundling also works for organisations that have made a deliberate, eyes-open decision not to move quickly on talent strategy and to stay on a single, slow, stable rhythm for both HR and payroll. That's a legitimate choice. It is just rarely the choice anyone explicitly makes. Most organisations default into it and then wonder, three years later, why their HR strategy keeps stalling.
The mistake we see most often is the unexamined assumption that integration is always worth paying for, even with vendor lock-in. It isn't. Integration is worth paying for when the things being integrated genuinely need to move together. Payroll and HR analytics need to move together. Payroll and learning content do not.
The principle
When the two clocks in your business run at different speeds, the worst thing you can do is force them to share a single watch.
The HR clock is going to keep accelerating. The payroll clock is going to keep ticking on the rhythm of legislation. Architect your employer stack to let each one keep its own time and use the integration layer between them to do the work that, in 2016, you would have asked a single vendor to do.
That's what Affinity has been built to do in the ANZ market: be the payroll engine that holds steady under compliance pressure, integrates cleanly with the HR system you actually want to use, and lets your HR strategy change without forcing your payroll to change with it.
Two clocks. Two systems. One sensible architecture.
Affinity is the ANZ payroll engine designed to hold steady under compliance pressure and integrate cleanly with the HR system you actually want to use. Let your HR strategy change without forcing your payroll to change with it.