Australian Lenders Are Running AI on Bad Data — And Your Small Business Probably Is Too

Australian lenders are rushing to adopt AI. Seventy-two per cent are already using agentic AI to help underwriters make decisions. Only 3% say their data is ready to support it.

New research from Experian, drawn from its global Connected Intelligence: Scaling AI with Trusted Data and Decisioning report, found that gap is the defining problem in Australian financial services right now. The technology is in production. The foundations aren’t.

The SMB angle is sharper than it looks.

What the numbers actually say

Experian surveyed Australian financial institutions and found:

• 72% are using AI for underwriting recommendations or decision support
• 75% cite faster decisions as the core benefit
• Only 3% describe their data as “fully AI-ready”
• 67% say their data is either not ready or only partially ready

The top three barriers: fragmented data with no unified customer view (45%), poor data quality (42%), and a lack of trust in AI outputs (31%).

Mathew Demetriou, Experian’s ANZ managing director of software solutions: “AI is already in underwriting workflows with the research showing 72% are using agentic AI for decision support, but the harder question being asked is how to trust and govern it at scale. The foundations underneath AI still need to catch up.”

These are institutions with IT departments, data teams, and compliance budgets. Sixty-seven per cent of them still don’t trust their own data enough to run AI on it properly.

What this means if you run a small business

If you’re using AI tools for quoting, forecasting, customer follow-up, or any kind of decision support, what you feed in determines what comes out. Clean data isn’t a nice-to-have. It’s the precondition for AI doing anything useful.

Most small businesses are starting from a much messier place than a bank. A few things worth checking:

• Where does your customer data actually live? Multiple systems that don’t talk to each other is the exact barrier this report flags most frequently.
• When was your data last cleaned? Duplicate contacts, stale records, and wrong categories don’t just slow AI down, they send it in the wrong direction.
• Do you trust what your AI tools are telling you? If you’re second-guessing the outputs, that’s usually a data problem, not a model problem.

Regulatory pressure is heading downstream

The report also notes that Privacy Act reforms, the new Scams Prevention Framework, and the Consumer Data Right are reshaping how institutions handle data. Eighty-four per cent of respondents said transparency of analytics was highly valuable. And only 51% were comfortable letting AI make decisions without human review, even for low-risk cases.

That caution will flow downstream. Small businesses that touch financial data, bookkeepers, mortgage brokers, accountants, payroll providers, will be caught up in compliance expectations that are currently being set at the institutional level.

Getting your data in order now is preparation, not just good hygiene.

Start somewhere concrete

Pick your worst data problem. One system, one cleanup. Your CRM contacts, your invoice records, your job costing spreadsheet, whatever is most broken. Fix that first, then layer AI on top of it.

The banks are working on the same problem at far greater scale. They’ve just got more people doing it.

More hot takes: Hot Takes →

If this interests you

• The AI Tools Australian Cafes and Food Businesses Are Actually Using Right Now
• Australia Is Building Data Centres. But Can Anyone Actually Use AI Yet?
• Claude Now Watermarks Everything It Writes. Should Your Business Care?

More hot takes: Hot Takes →

Sources

• Mina Martin, Australian lenders race to adopt AI, but their data isn’t ready for it, Australian Broker, 11 August 2026
• Experian, Connected Intelligence: Scaling AI with Trusted Data and Decisioning (global report, 2026)

📖 More deep dives: News Deep Dives

📊 Compare AI tools side by side | 💼 Free resources & AI prompt packs

📖 More deep dives: News Deep Dives →