By Nick Lim, Founder, Switchboard Finance
The fintech industry loves the neobank versus traditional bank narrative. Slick app against branch network. AI-driven credit scoring against manual underwriting. Digital onboarding against paperwork. It makes for good conference panels and clean LinkedIn posts.
But from where I sit, writing deals for self-employed Australians every day, neither side is winning. Because neither side was built for the people I work with.
The traditional bank problem: a system built for PAYG income
Banks built automated credit systems optimised for PAYG income. If your earnings arrive by payslip, the machine works beautifully. If your income comes from a BAS statement, a trust distribution, or six client invoices landing at unpredictable intervals, the machine was not designed to assess you. It returns a decline, not because you cannot service the loan, but because you are the wrong shape for the algorithm.
I see it every week. A business doing half a million in revenue, profitable and growing, declined because the ABN is under two years old. That is not a credit risk. That is a policy gap.
And this is not a small group of people. More than 2.7 million actively trading businesses were operating in Australia as of June 2025, and close to two-thirds of them have no employees, which means the owner is the business. The big four have spent a decade automating themselves out of serving them. Every year the credit policies tighten for non-standard income. The retreat from self-employed lending is not a bug. It is a feature of how these institutions chose to scale.
Why neobanks have not fixed self-employed lending
Here is where the narrative falls apart. Neobanks were supposed to fix this. Faster decisions, smarter data, no legacy systems. In theory, a neobank powered by real-time transaction data should assess a cafe owner’s actual cash flow better than an analyst reading a two-year-old tax return.
In practice, most neobanks built their credit models on the same PAYG assumptions the banks use. They made the onboarding faster and the interface prettier, but the underlying risk assessment still penalises variable income, seasonal revenue, and complex entity structures. A self-employed borrower declined by a major bank’s algorithm gets declined by most neobank algorithms too, just with a nicer rejection screen.
The neobanks that have gained real traction in business lending, the unsecured working capital platforms like Prospa and Lumi, are genuinely different. They read bank statement data instead of tax returns and approve in hours rather than weeks. But they operate at the short-term, higher-rate end of the market. They solve the cashflow gap, not the home loan gap, the asset finance gap, or the development finance gap. The core products that build wealth for business owners stay locked behind the same traditional credit policies that banks and most neobanks share.
The real winner: non-bank lending and private credit
While banks and neobanks argue over who owns the digital experience, the non-bank sector has quietly become one of the most significant parts of Australian finance. Private credit has grown to roughly $235 billion in assets under management, up from a fraction of that a few years ago. Non-bank lenders are approving deals that both banks and neobanks decline, not because they have lower standards, but because they kept the human underwriters the automated systems replaced.
A human credit analyst can read a cafe owner’s BAS and understand that the January dip is seasonal, not structural. They can look at a trucking operator’s bank statements and see that the lumpy deposits are progress payments on a government contract, not irregular income. They can assess a developer’s file on the strength of the completed asset and the exit strategy, rather than on whether the borrower has two consecutive years of tax returns.
That is not technology. It is judgment. And it is the thing both banks and neobanks automated away in the pursuit of scale.
Human underwriting is the real differentiator
This is the part the technology debate keeps missing. The question was never whether the app is good. A faster, prettier front end on the same PAYG risk model still says no to the same people. What changes the outcome for a self-employed borrower is a person who can interpret a non-standard file, not a smoother way to submit it. The RBA has noted that private credit has grown strongly as banks retreat from lending that does not fit tighter capital and credit settings. That retreat is exactly the space business owners fall into. The lenders filling it compete on understanding, not interface.
Where self-employed lending goes next
The future of lending for business owners is not more automation. It is better human judgment supported by better data. The institutions that combine real-time transaction analysis with experienced human underwriting will win the self-employed market. Nobody has cracked that combination at scale yet.
Until someone does, the millions of Australians who do not fit either the bank’s template or the neobank’s algorithm will keep finding their way to specialist brokers and non-bank lenders. Not because the technology is not good enough, but because the risk models behind the technology still were not built for them.
The neobank versus bank debate is a technology argument. The real gap in Australian lending is a risk-model argument. And until a credit engine can genuinely assess variable, seasonal, multi-entity business income at scale without penalising it, the non-bank sector will keep growing, and the debate about who has the better app will keep missing the point.
Sources
Business numbers: ABS, Counts of Australian Businesses (June 2025)
Private credit AUM: EY-Parthenon, Australian Private Debt Market Overview (2026)
Bank retreat and private credit growth: RBA Bulletin, Credit Markets (February 2026)
