The trajectory of broking in an AI-powered world

To help answer this, let’s start with one bit of important technical information. In 2016 Google’s AlphaGo beat the world champion at a strategy game called Go by learning to play using around 30 million known moves or data points.

Fast forward 10 years and modern large-language models that power online programs like ChatGPT are powered by many trillions of data points. This exponential explosion of data available to (and consumed by) platforms such as this is actually hard for the human brain to truly comprehend. To put it in perspective, it is an improvement by an order of magnitude of around 10, the equivalent of adding 10 zeros. Imagine a colleague who in 10 years got 10 billion times more knowledgeable with another 0 being added every one to two years and you will start to get a sense of the scale and pace of change!

In 2023, ChatGPT and equivalent capability from the other major technology platforms (Google, AWS, Microsoft etc) started opening up this AI technology to every software developer in the world. The companies they work for - such as your brokerage or aggregator - now have access to this AI capability.

In short, what this means is the near-magical superpowers we’re getting a taste of in tools like ChatGPT are starting to show up everywhere, including broker technology. These tools will drive the efficiency and productivity of the entire value chain, making it easier and faster for brokers, support and their clients to navigate the process end to end.

Sounds great, right? So what’s the problem?

Mortgage broking sits at just over 74 per cent of total loans and is growing, partly due to a longer term shift powered by the three persistent customer needs: convenience, choice and cost.

The issue for brokers, brokerages and aggregators is that if broker productivity goes up by orders of magnitude - and the number of loans available to be written grows nominally/as a function of population growth and market share growth - there are going to be two very different classes of operators in the market; highly productive brokers and everyone else.

Running some basic numbers, assuming around 65,000 loans are settled in Australia per month, and a good broker today can settle 10 loans a month, you’d only need 6,500 brokers to write 100 per cent of loans in the market!

At 20 per month, just 2x (not 10x,) this number goes down to around 3,500. A far cry from the 19,000 brokers who exist in the industry today.

Low productivity brokers will ultimately self-select out.

This maths doesn't suggest the disappearance of brokers, but the concentration of settlements to a smaller and smaller number. This shift isn’t only occurring in mortgage broking in Australia, though. For a glimpse of the future look at the difference between a trading floor in the late 2000s vs the make up of a highly profitable trading firm or hedge fund now.

So, what does this all mean and what can or should you do?

For brokers with established businesses and/or looking to exit the industry in the next 5 - 10 years a complete overhaul of your business and operations is either impractical or not possible. Leveraging either the emerging technology provided by your aggregator and/or off-the-shelf technology to drive productivity through efficiency and better engage the next generation of customers is going to be your highest value activity.

To achieve this, you are going to need to become or attract someone in your business who can actively operationalise these tools and lift (read: train) you and your ‘old dogs’.

Our experience at UNO, for example, is that brokers of all ages and experience levels can adapt to a new and much better way of doing things if they are trained in person and supported through that change in a sustained way.

For loan writers who are working in an existing brokerage, or working at a bank and their next move is to establish their own brokerage business (not as an employee), your business partner decision is probably the most critical decision you will make right now.

The critical technology-related questions to ask and evaluate would be:

1. What is the current state of the technology they provide to both you as a broker, the support and the customer? What is the track record of the improvement of this technology? Releasing once a year isn’t a good sign, quarterly is fine, but monthly is even better.
2. How many engineers have they got on the team, how senior are they and how long have they been with the team? Building and supporting outstanding (not just good) technology happens over many years and needs highly engaged, highly skilled people committed over that timeframe.
3. How much of the end-to-end process is done directly by the broker vs by a combination of the customer, the back-office support and the technology? Where you spend your time is key to getting leverage and your overall productivity. We believe that talking to prospective clients, doing credit and building relationships are the three highest-value activities a broker can be doing to maximise customer outcomes and settlement outcomes. Everything else takes you away from these three things.
4. How is the brokerage or aggregator leveraging AI today and what are their plans for the future? The future is already here if they are not using AI already they are already behind

It’s going to be a brave new world, make sure you’re on the right side of history.

Vincent Turner is the founder & CEO of UNO Home Loans