Login
ISCOVER
An AI-powered data-matching algorithm has been rolled out by the brokerage to automatically pair and allocate documentation data with relevant document requests.
Online mortgage brokerage UNO Home Loans this week (17 July) launched an AI-powered initiative that connects application data to required document checklists in a bid to reduce the administrative burden of document matching for its brokers.
To continue reading the rest of this article, please log in.
Looking for more benefits? Become a Premium Member.
Create free account to get unlimited news articles and more!
Looking for more benefits? Become a Premium Member.
The technology, embedded into the broker platform used by the brokerage’s 10 brokers, works by enabling the complete set of data to be captured within the broker platform, which is then piped to the UNO proprietary rules engine that generates a dynamic document checklist specifically reflecting the data in the application.
When documents (such as ID documents, payslips, notice of assessments, contracts of sale, etc) are sent to the broker, they are automatically ingested into the client application and the AI tool examines each document in a private AI instance on Google Cloud. This extracts the type of document uploaded and its extensive metadata, such as the annual pay and employer on a payslip or the address on a purchase contract.
The extraction forms a map of every document alongside the checklist. Then, Google’s Vertex AI is used to generate matches against what document is needed and to whom (or what) it relates. For example, a payslip for John Smith or a purchase contract for a specific property.
UNO’s founder and CEO Vincent Turner has said the brokerage developed the technology to help save brokers’ time on collecting, downloading, organising, labelling, and matching documents. He suggested that the technology can reduce document collection and matching to a few seconds, freeing up the broker to focus on client-facing/dollar-productive tasks.
Speaking to The Adviser, Turner said: “Document matching is just the beginning. Once you have a strong relationship between the application data, the required documents and the documents themselves it enables the entire process to be reimagined.
“We see a world where as a client there is virtually no application form, just connect open banking and upload documents and you’re done.”
According to the brokerage CEO, the capability is based on “progressive learning”, so as the AI sees more applications, new data types, and extended metadata, it learns and improves over time.
UNO reportedly plans to extend this to support commercial applications and documents in the coming quarter and said that near-term extensions could also help with fraud detection, where subtle but telltale signs of document manipulation can be flagged to the broker.
Connective warns brokers to be vigilant to fraudulent documentation
Documentation administration and security are of increasing importance to the broking industry, with many groups focusing on data security and identifying fraudulent documentation.
Aggregation group Connective this week reminded brokers to be more vigilant and thorough when verifying client documents to identify and combat potential collusion between clients and third parties.
According to Connective’s group legal counsel Daniel Oh, brokers should be reviewing clients’ documents “carefully” as part of the application process and to identify potential collusion.
“Brokers jeopardise their accreditation with a specific lender when they submit an application without thorough review,” he said.
“Under the National Consumer Credit Protection Act, brokers have an obligation to make reasonable inquiries. However, Connective has found that brokers do not always escalate concerns sufficiently,” he said, adding brokers should not rely solely on information provided by their clients.
Connective has outlined that brokers should be alert for “inconsistencies” in payslips, including mismatched fonts, spelling errors, or any other discrepancies that may warrant further investigation.
The aggregator put forward the following steps for brokers to detect potential collusion in loan applications:
- Scrutinise supporting documents: Watch for inconsistencies in payslips, bank statements, and documents from third parties. Red flags include mismatched ABN details, missing tax agent information, unrealistic income compared to the applicant's role, formatting errors, and inconsistencies across documents.
- Multi-layered verification: Obtaining documents is only the first step, so don’t just accept documents at face value. Verify details with employers directly, check accountant credentials with the Tax Practitioners Board register, and conduct online searches to confirm the legitimacy of businesses and individuals involved.
- Escalate to be sure: Understand the concept of “reasonable inquiries” under the National Consumer Credit Protection Act (NCCP). If inconsistencies arise, escalate your concerns by contacting relevant parties and document all inquiries made.
- Be wary of pressure to fast-track: Be extra cautious around clients who pressure you to rush the application or seem reluctant to provide necessary documents. A genuine applicant should be comfortable with standard verification procedures.
“Watch for unusual client behaviour – do their actions or overinvolvement ring alarm bells? Excessive pressure to expedite the application or reluctance to provide necessary documents can be signs the applicant may not be genuine,” Oh said.
“Even a slight doubt warrants a further check until you’re satisfied – and always document your inquiries for your records.”
JOIN THE DISCUSSION