Perth mortgage broker

How a Perth Mortgage Broker Uses AI to Pre-Qualify Leads and Cut Comparison Time in Half

Composite case study based on real patterns from Australian mortgage brokers using AI. Names are illustrative.

Michael is a sole-operator mortgage broker in Perth’s northern suburbs. Accredited with around 30 lenders, he writes about 80 loans a year. The work is detail-intensive and compliance-heavy: which is exactly why he was sceptical about AI. After 12 months of using it carefully, he’s converted. Here’s what actually moved the needle.

Pre-Qualifying Leads Before the First Call

Michael’s old process: a prospect would enquire, he’d call them, spend 30-45 minutes on the phone gathering basic info, then another hour assessing whether they were genuinely serviceable before committing to a full application. Many didn’t proceed. A lot of that time was wasted.

Now he sends every new enquiry a short online form (built in Typeform) that captures income, expenses, debts, deposit, and property target. He pastes the responses into a ChatGPT prompt:

“Based on the following financial details, give me a preliminary serviceability assessment for an Australian home loan. Be conservative. Flag any obvious issues. Income: [details]. Expenses: [details]. Existing debts: [details]. Deposit available: [amount]. Target property value: [amount]. I’m a licensed Australian mortgage broker: this is for my internal pre-qualification use only, not financial advice to the client.”

The output gives him a quick read on whether it’s worth pursuing before he picks up the phone. He still makes his own assessment: the AI doesn’t replace his judgement: but the 45-minute discovery call is now a focused 15-minute conversation.

Lender Comparison Summaries

Writing a clear, readable lender comparison for clients used to take Michael an hour. He’d pull rate sheets, calculate repayments, compare features, and format it all into something a first home buyer could actually understand. Now he feeds the raw data into ChatGPT and gets a structured comparison draft: rates, fees, features, pros and cons of each option: in minutes. He reviews everything carefully before it goes to the client, but the formatting and narrative work is done.

Client Communication

Mortgage broking involves a lot of status update emails, “we’ve submitted your application,” “the valuation has been ordered,” “conditional approval received.” These are routine but need to sound personal and clear. Michael uses ChatGPT templates for all of them, personalising the key details before sending. Client feedback on communication has improved: he thinks because the emails are now more consistently warm and clear, rather than rushed and terse when he’s busy.

Compliance Notes

Michael uses AI to help structure his Best Interests Duty documentation: not to write it (the compliance details are his professional responsibility) but to organise and format the narrative around why a particular product was recommended. Saves him 20-30 minutes per file on the write-up component.

What He’s Careful About

Michael is explicit that AI never gives financial advice to clients and never touches ASIC-regulated content without his review. The AI is a drafting and formatting tool: his licence, his professional judgement, and his responsibility. He makes this clear in his AI usage policy, which he keeps on file.

He also doesn’t paste identifiable client data into public AI tools. For anything with names and financials, he uses anonymised placeholders in the prompt.

The Numbers

  • ChatGPT Plus: ~AU$28/month
  • Typeform (lead forms): ~AU$35/month
  • Total: ~AU$63/month
  • Time saved per deal: estimated 2-3 hours
  • At 80 loans/year: potentially 160-240 hours saved annually

For a sole operator, those hours are the difference between a sustainable practice and a burnout spiral.


📖 Related: Full AI guide for mortgage brokers | AI for financial planners

Related Reading

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

📋 Browse all Australian small business AI case studies →

Similar Posts

Leave a Reply

Your email address will not be published. Required fields are marked *