Australian mortgage brokers spend 11 per cent of their time educating and preparing each customer for a mortgage application, according to Deloitte Access Economics research commissioned by the MFAA (2025). The other 89 per cent goes to consultation, data collection, verification, assessment, application preparation and file management. Brokers are paid for advice and spend most of the week on process.
The idea
The commercial opportunity in Australian mortgage broking is not better AI-generated advice, it is the removal of the administrative load sitting between a client conversation and a lodged application. Mortgage Choice, part of REA Group, reported in December 2025 that moving from document collection to open banking data import saves around 90 minutes on every loan application. That 90 minutes comes out of three specific tasks, and each one is separately measured.
| Task in the loan file | Manual process today | With data import and AI-assisted checking |
|---|---|---|
| Collecting client income and expense data | ~20 minutes of manual data entry | Imported directly from the client’s bank feed |
| Verifying the documents supplied | ~40 minutes of checking | Automated matching, with exceptions flagged to the broker |
| Chasing missing paperwork | ~30 minutes of client follow-up | Prompted automatically against a required-document list |
| Total per loan application | ~90 minutes | Recovered |
Source: Mortgage Choice (REA Group), December 2025.
Borrowing capacity assessment, lender policy matching and compliance file notes sit alongside those three tasks and are the obvious next candidates in an Australian brokerage, though Alvo has not found published Australian time measurements for them, so no figure is claimed here.
Why it matters
Recovered administrative time only becomes commercial value in a mortgage brokerage if it converts into more lodgements or fewer support hours. The MFAA Industry Intelligence Service, 19th edition (2025), put the average number of applications lodged per active broker at 19.1 per six-month period, or roughly 38 a year. That gives an owner a calculation they can run against their own file numbers this quarter.
Worked example, five-broker brokerage:
| Step | Calculation | Result |
|---|---|---|
| Annual application volume | 5 brokers x 38.2 applications a year (19.1 per half year, doubled) | 191 applications |
| Time recovered | 191 applications x 1.5 hours saved per application | 286.5 hours a year |
| Expressed as working weeks | 286.5 hours divided by a 38-hour week | 7.5 working weeks |
| Expressed as support cost | 286.5 hours x $45 fully loaded support cost | ~$12,900 a year |
The $45 hourly rate is an Alvo assumption for the reader to replace, not a published figure. Every other input is sourced above. The number that matters is the 7.5 weeks rather than the $12,900, because most brokerages will not remove a support role, they will absorb growth without adding one. The same recovered hours are what close the operational gap behind why Australian brokers are losing settlements they have already won.
McKinsey & Company (2025) found that relationship managers in many commercial banks spend just 25 to 30 per cent of their time in client dialogue, and projected that early agentic AI use cases could reduce manual workloads by 30 to 50 per cent. Both McKinsey figures are global rather than Australian. The Australian read-across is that the ratio problem inside a mortgage brokerage is the same problem the major lenders have already costed, and lenders are acting on it first. It is also the same shape as the margin question in other document-heavy Australian industries, which is why AI is changing margins for Australian construction firms through the office rather than the site.
The Alvo take
Australian financial services is already one of the country’s higher AI-adopting sectors, so a broking principal deferring the decision is not being cautious, they are conceding turnaround speed to firms that lodge cleaner files faster. The Australian Bureau of Statistics reported in Characteristics of Australian Business 2024-25 (released 2026) that 24 per cent of businesses in financial and insurance services used AI, against 12 per cent across all Australian businesses. That is the same adoption gap driving where AI creates real margin in real estate, in the industry sitting on the other side of most broker referral relationships.
Whilst the sensible pathway starts with data and documents rather than with advice, the commercial case has to be measured at the file level. Broker Pulse, run by Agile Market Intelligence, reported in August 2026 that the average time to an initial credit decision in June 2026 was 3.7 business days across the large ADI segment and 7.1 days across non-ADIs, so lender turnaround is not something a brokerage controls. Preparation quality is, and a clean file that does not come back for rework is the only lever an owner holds on that timeline. Working out which files are costing you that time is a structured assessment of where the hours actually sit rather than a software decision.
Worth flagging the compliance boundary before anything is switched on. The Privacy Act 1988 automated decision-making transparency obligation commences on 10 December 2026 in Australia, and best interests duty under the National Consumer Credit Protection Act sits with the broker rather than with the software. AI belongs on data handling and file preparation, whilst the recommendation stays documented and human.
Brokers are paid for advice and spend most of the week on process. The recoverable time sits in data collection, verification and chasing paperwork, which is measured at around 90 minutes a loan file. For a five-broker firm that is seven and a half working weeks a year, and most owners will use it to absorb growth rather than remove a support role.
Sources: Deloitte Access Economics for the MFAA, The Value of Mortgage and Finance Broking, 2025. Mortgage Choice (REA Group), reported by The Adviser, December 2025. MFAA Industry Intelligence Service, 19th edition, 2025. Australian Bureau of Statistics, Characteristics of Australian Business 2024-25, 2026. McKinsey & Company, 2025. Broker Pulse (Agile Market Intelligence), August 2026. OAIC on the Privacy and Other Legislation Amendment Act 2024. ASIC on best interests duty under the National Consumer Credit Protection Act 2009.
Common questions
How much time do mortgage brokers spend on admin instead of advice?
Deloitte Access Economics research for the MFAA (2025) found Australian mortgage brokers spend 11 per cent of their time educating and preparing each customer for a mortgage application. The remainder covers initial consultation, verification, borrowing capacity assessment, application preparation, managing the file to settlement and post-settlement contact. Most of a broker’s week is administration, not advice.
Can AI actually reduce the time it takes to process a home loan application in Australia?
Yes, in the data collection and verification stages. Mortgage Choice, part of REA Group, reported in December 2025 that moving from document collection to open banking data import saves around 90 minutes per loan application: 20 minutes of manual data entry, 40 minutes of verification and 30 minutes chasing clients for paperwork. Credit decision time still sits with the lender.
How many Australian financial services businesses are using AI?
The Australian Bureau of Statistics reported in its Characteristics of Australian Business 2024-25 release (2026) that 24 per cent of businesses in financial and insurance services used AI, against 12 per cent across all Australian businesses. Financial services is one of the three highest-adopting sectors in Australia, so a brokerage competing on turnaround speed is competing against adopters.
Will AI let a mortgage broker write more loans without hiring support staff?
Time returned converts to capacity only if the broker fills it with client work. The MFAA Industry Intelligence Service (2025) put average applications lodged at 19.1 per active broker per six months. A five-broker firm lodging around 191 applications a year, saving 90 minutes each, recovers just over 286 hours, close to seven and a half working weeks.
How long do Australian lenders currently take to give an initial credit decision?
Broker Pulse, run by Agile Market Intelligence, reported in August 2026 that for June 2026 the large ADI segment averaged 3.7 business days to an initial credit decision, the small ADI segment 6.3 days and the non-ADI segment 7.1 days. Lender turnaround is outside a brokerage’s control, which is why the time a broker can control is worth recovering.
What are the compliance risks of using AI in an Australian mortgage broking business?
The Privacy Act 1988 automated decision-making transparency obligation commences on 10 December 2026 in Australia, and reaches systems that materially influence decisions affecting a client. Best interests duty obligations under the National Consumer Credit Protection Act sit with the broker, not the software. Use AI on data handling and file preparation, and keep the recommendation documented and human.