Financial services professionals and end-of-the-world movie enthusiasts alike may well recoil at the idea of artificial intelligence infiltrating the home loan process, but brokers and borrowers have nothing to fear. NextGen.Net’s Sales Director Tony Carn explains how this new technology will boost your business
I’ll admit that there can be a fair bit of fear associated with new technology. Particularly around things like robo-advice and new online players.
At NextGen.Net, we’ve been focusing on ‘machine learning’ lately, which is a form of artificial intelligence based on pattern recognition.
The exciting thing for the thirdparty channel is that machine learning can allow brokers to focus on doing what they do best, which is having conversations with clients and providing sound credit advice.
One of the biggest bugbears for brokers is the amount of time spent on processing loan applications. From our perspective, streamlining the processes and the non-value-add parts of a broker’s business allows them to get on with doing their next deal.
Whether you are a sole trader processing your own loans or you have a back office of support staff, having greater efficiency in home loan processing allows you to build out your business and expand, whether that’s geographically or diversifying into a new revenue stream.
We are currently in the process of building further enhancements to our ApplyOnline lodgement system.
The current system has a high degree of sophistication built into it already. Brokers are provided with a dynamic list of documents required based on our sophisticated rules engine, and are then able to upload and file them on screen into categories based on the lender’s requirements. This allows future reworks to be significantly reduced.
What we are working on now, though, is improving the entire user interface, making it more streamlined in terms of the look and useability. The process will be quicker, easier and far more intuitive.
But it’s the machine learning part that’s really exciting. This allows ApplyOnline to recognise what documents are, based on what they have seen before. We have fed over 100,000 documents into a machine and those documents are helping the machine learn.
So the machine learns, for example, what a tax return looks like. It learns what different types of payslips look like. It can continually learn from all of those documents that are fed into it. We’re effectively training the machine to recognise the different types of documents that a customer provides to support a home loan application. That’s machine learning.
This process enables the system to make predictions. The more the machine learns, the higher confidence levels you get out of those predictions. What this means is that we add value to brokers by saving them more time. The smarter the machine is, the faster and more confident a prediction; and the machine will then read and recognise the document, and help file it in the right place for the relevant lender.
It’s all about looking at different things, like text search phrases. We use our Optical Character Recognition (OCR) to search text on every document. We’ve also got the machine learning part of it, which is from the images themselves, as well as the statistical analysis.
Those three things build the intelligence of the machine and make it smarter and more effective for the broker and lender.
In 2016, NextGen.Net will launch ‘Smart Docs’ and reveal new capabilities that machine learning can deliver to brokers and lenders.
While we provide the technology, it is really the brokers who we should give credit to, because what we have done really came as a response to the market.
Best-of-breed brokers like to get the processing piece done so they can get on with writing the next home loan and meeting their next client. So we looked at how we could streamline the supporting docs process.
It is still a real problem for the industry, particularly the lenders – the amount of Missing Information Requests (MIRs) or documents that are illegible.
These things create reworks, delayed approvals and affect approval rates and customer satisfaction levels. Supporting Docs has received a great take-up from lenders.
One of the things we need to do is continually evolve that and come up with new developments and solutions like ‘Smart Docs’, so that the process becomes as effortless as possible for brokers.
One of the big things we did last year was increase our training presence in the market. We now have training managers who specifically train brokers.
They are a fantastic source of feedback for point-of-sale problems or market issues that brokers experience.
They provide invaluable information that can be used in creating smarter solutions to do things better.
NextGen.Net also hold bi-annual technology workshops with all of the tech leaders from the major aggregators in Australia. We discuss current and possible future market problems that we can develop solutions for.
We like to get a consensus and really have that community engagement about what we develop.
The big four bank has told ASIC to consider the utility of the ...
AFCA has clarified that for historic complaints involving brokers...
Small-business lender Capify has appointed Dino Pacella as its fi...