iscover 
A new Sydney-developed artificial intelligence (AI) platform, SecondBrain, and its flagship pricing module, Sharp, have officially launched to the Australian mortgage broking sector to streamline complex, time-consuming administrative workflows.
Designed to reduce multi-day pricing and administrative tasks down to minutes, the platform automates lender portal submissions, document processing, and data entry.
The tool harnesses AI to pull current rates from the lender’s pricing tools, so the broker can still compare them side by side with the escalated pricing from the nine portal lenders.
It therefore gives brokers a full market picture in one view, rather than having to manually send off pricing requests. It currently prices for new lending rather than repricing for existing loans.
Built inside the Microsoft environment, SecondBrain operates directly within Microsoft Teams, utilising Entra ID for identity verification and Azure infrastructure for cloud operations.
The technology was originated by developer James Larkey in response to operational bottlenecks experienced by mortgage brokers.
According to Larkey, the Sharp technology automates new lending escalation pricing for new-to-bank applications across up to nine lenders simultaneously.
By processing loan scenarios through attended automation, the tool returns aggregated pricing comparisons to brokers within four minutes.
The broader SecondBrain platform can also execute a range of operational workflows, including document classification and filing via OneDrive or SharePoint, lender valuations, credit report ordering, bank statement processing, and automated data entry into aggregator platforms, such as Connective.
It is aggregator-agnostic and structured on a Software-as-a-Service (SaaS) subscription model.
Speaking to Broker Daily about SecondBrain and Sharp, founder and developer Larkey revealed that the technology was first built to resolve bottlenecks at major brokerage Azura Financial, following a request from one of the brokerage’s brokers in 2025.
He said: “I was in Azura’s office late one night setting up the renaming agent when director Tom Hawley came off a client call in a foul mood. Six loan splits across five lenders, and a client who wanted it back immediately. I asked how he would normally handle it. Same way he had for 10 years: open every portal, key it in, wait. That was the whole brief.
“Azura Financial then engaged me to build it properly and I spent six months sitting in their office. Their brokers became the test bed for the pricing tool, working on real files with real deadlines, which meant feedback came back the same afternoon rather than through a formal pilot.”
The tool is now being used by more than 40 brokerages across Australia.
“A broker enters a scenario once, in Teams,” he said and noted that Teams is the interface, Entra ID handles identity, the program runs on Azure in Australia East, and credentials sit in their own Azure Key Vault.
“Sharp requests pricing from up to nine lenders at once and returns the results in a single comparison, usually inside four minutes.
“Since this is new lending rather than repricing, what we see is escalated pricing on new-to-bank applications.”
While the pricing tool (Sharp) has been built to assist brokers in running a bigger book with the same team, the SecondBrain platform can run thousands of individual jobs outside of this.
Larkey told Broker Daily that, in August alone, SecondBrain handled 31,209 individual jobs, including ID checking, document renaming, lender valuations, and payslip and living expense validation.
The founder outlined that the platform also functions with strong security and governance protocols. For example, it functions under the active session and permissions of the individual broker, requiring manual multi-factor authentication (MFA) entry by the user when accessing lender portals.
Sensitive borrower data, including income details, employment history, and credit files, is not retained by the system, operating strictly on a pass-through basis.
Encrypted lender credentials are stored in dedicated per-firm vaults and accessed via reference calls rather than stored passwords. The platform relies on Microsoft’s underlying SOC 2 and ISO 27001 certifications.
Larkey said: “The only thing Sharp retains is the de-identified rate outcome. That is what lets us publish the network report to every user on a Monday morning. Everything else is pass-through. We are not a system of record and we are not trying to become one.
“Borrower data is never stored in the system. No income, no employment details, credit files. Credentials sit in a dedicated encrypted vault per firm. For all lenders we never see the password at all, we pass a reference to the vault and the automation retrieves it directly. Every access is logged with the broker, the lender, the outcome and the time. Brokers hold no password with us; they sign in with their own Microsoft 365 identity, so their firm’s MFA and conditional access policies apply."
The SecondBrain founder said that he is now working with the banks towards formal approval.
The platform also offers validation agents that cross-reference client-declared payslips and living expenses against initial fact-find data.
Additionally, SecondBrain collects de-identified approved rate outcomes across its network to generate a weekly peer-rate report. This analytics feature provides brokers with aggregated market data reflecting actual approved interest rates across various loan-to-value ratio (LVR) brackets, repayment types, and borrower categories.
[Related: Yellow Brick Road launches new AI platform]
Want to see more stories from trusted news sources?Make Broker Daily a preferred news source on Google.
