At a glance
Business: Metro Finance
Industry: Commercial non-bank lending and asset finance
Product: Affinda Extract, document automation platform
Challenge: Rising settlement volumes meant more manual document review, which put pressure on turnaround times and broker service
Solution: Document automation that splits and classifies settlement documents, extracts and validates data, while working within Metro’s existing systems and regulatory requirements
Results:
3,700+ deals processed, with 97.5% passing validation
Almost 80,000 fields extracted and checked
Near-instant settlement for some introducers and customers
Metro Finance is built on people. Fast turnaround, reliable finance options and high-touch service with a national network of brokers who know they’ll get a real answer, fast. But as deal volumes grew, every settlement came down to a bottleneck of data stuck across multiple documents the team had to review, classify and validate by hand.
Metro didn’t want more manual work, and they didn’t want AI for the sake of it. They challenged Affinda to help scale the settlements team while keeping the accuracy, controls and human judgement a regulated lender needs. With Affinda Extract, routine document checks now run automatically and only exceptions that need human review are routed to the team. For some introducers and their customers, the settlement processing is now near-instant.
“We don’t measure success by whether we implemented AI within Metro. We measure success by what changed for the business and the customers.”
– Ninos Yacoub, Head of AI, Architecture & Data, Metro Finance
The challenge
Growth without compromising lending standards
Metro Finance is a leading Australian non-bank lender providing vehicle and equipment finance. Since 2011, it has funded more than $10 billion in loans and supported over 120,000 customers through its national introducer network and is known for fast turnaround and personal service brokers can count on.
As deal volumes increased, that reputation came under pressure. Settlement staff reviewed, classified and validated large amounts of document information by hand, and the process became a constraint on scalability and turnaround times.
“Growth should never be at the expense of service,” said Ninos Yacoub, Head of AI, Architecture & Data at Metro Finance. “Rather than responding to higher volumes by adding more manual work, we’re using automation to remove repetitive administration and create capacity for our teams.”
But Metro doesn’t automate for its own sake. “At Metro, we don’t start with AI – we start with the business problem,” said Ninos. Any new technology had to integrate with Metro’s existing technology landscape, meet the standards and regulatory requirements of Australia’s lending environment, and answer three core criteria:
Would it improve the experience for customers, brokers and staff?
Would it remove repetitive work?
Would it deliver a measurable business outcome?
Accuracy was especially important because the team needed to trust data was extracted and validated before being put to work in their lending operations.
The approach
Validated document data, with people in control
Metro builds in-house where technology creates an edge, as it has with its MyMetro platform. Document automation was different. Metro chose Affinda Extract, Affinda’s document automation platform, for its OCR and AI-driven document intelligence and the way Affinda worked with the team throughout the process.
Metro’s settlement packs contain a mix of documents, from bank statements and supplier invoices to driver’s licences and certificates of currency. Affinda Extract splits each pack into its individual documents and classifies them, so the right extraction and validation rules can be applied. The required fields are extracted, with every value traceable to its source. The data is then checked against Metro’s business rules before moving into its existing systems, with exceptions routed to the team for review.
Getting the technology right was only part of the process. Ninos described the implementation as smooth and highly collaborative, with Affinda working closely with Metro's team to understand its settlement requirements and integrate the platform into its existing technology environment.
“Document automation is something every lender needs. It’s not where we create an edge. We will leave that to Affinda.”
– Ninos Yacoub, Head of AI, Architecture & Data, Metro Finance
The Affinda difference
Accuracy the team could trust, exceptions it could own
Metro tested Affinda against its own settlement documents, including the variations that come from different brokers and customers. At the time of Ninos' interview, more than 3,700 deals representing almost 80,000 fields had been processed through the system, with reported extraction accuracy above 97%.
The proof of concept also demonstrated a low exception rate, with the cases flagged for review generally requiring genuine attention from the settlement team. That meant Metro could automate routine checks without losing visibility or control.
Affinda Extract's instant learning capabilities also helped shorten the feedback loop. Corrections and examples from the team could inform future extractions, helping the system adapt to different document formats without repeatedly retraining the underlying models. Metro was able to reach its accuracy targets sooner, with less manual effort during testing.
“Affinda gave us the confidence to automate the routine work while maintaining the controls, visibility and human oversight you would expect in a lending environment.”
– Ninos Yacoub, Head of AI, Architecture & Data, Metro Finance
The outcome
Greater capacity to serve brokers and customers
Routine checks now run consistently. The settlements team spends its time where judgement counts. For some introducers, settlement is now near-instant. Their customers get into a new vehicle or put new equipment to work sooner.
“Our team spends less time manually reviewing documents and more time dealing with exceptions, supporting our brokers and improving customer service,” said Ninos.
Lending decisions stay with Metro’s people. “Our principle is simple: automation supports our people,” said Ninos. “Technology is excellent at processing documents and handling repetitive tasks, but lending decisions, assessments and customer outcomes require human judgement.”
As its network grows, Metro can scale settlements without scaling manual work. Its technology team stays focused on the broader lending journey.
“The best result comes from combining the efficiency of automation with the experience of our people.”
– Ninos Yacoub, Head of AI, Architecture & Data, Metro Finance