7 questions on Open Banking

A fireside chat with Richard Shanahan, Tic:Toc’s Head of Data Science

He sits cross-legged in the armchair. Gently swirling a tumbler of scotch and rocks, he’s entranced with a brooding expression. Richard Shanahan is, above all else, a tall man. When he’s not deep in analysis as Head of Data Science for Tic:Toc, you’ll find him trimming his salt and pepper stubble. That, or enjoying the finest Cuban this side of the Torrens. Behind him, an ornate fireplace dominates the room. Despite this being a fireside chat, no fire is lit. It’s Summer. And it’s fire ban season. Blues plays softly in the background.

He shoots a glance that tells me he’s desperate to talk about Open Banking. We dive straight in.

So, in a word, it’s choice. With open banking, consumers are now going to be able to move freely from provider to provider and get better financial services products.

It’s the recognition that the data belongs to you, the consumer. It’s your data. What you do with it is up to you. And you should be able to enter into a value exchange with financial service providers to get a product that’s going to better suit your needs.

For the Australian financial services landscape, it means we’ll see a competition refresh with new players entering overtime, and competition drives better consumer outcomes.

The implementation of open banking in Australia is mandated and referred to as the Consumer Data Right (CDR). The unique thing about this legislation is it puts the consumer at the centre. So, it’s your data and you get to dictate how it’s going to be shared.

A challenge that will face financial service providers and participants in Australia is the current accreditation process under the CDR. It is rigorous and it’s really forcing providers to look at a lot of the stuff that’s fundamental to their business — their assurance and risk strategies, their info security and architecture, and their overall strategy.

Past the accreditation challenge, the main issue will be ongoing compliance. You need to demonstrate that you are managing your customers’ concerns appropriately while also maintaining the data to the high standards dictated by the CDR.

After receiving the data, the question becomes ‘what do you do now? Once you receive account transaction data, there’s still a lot of work to be done. How do you truly identify salary, adjusting for aspects such as salary sacrifice? How do you find rental income and apply haircuts or adjustments? Are your expenses discretionary and non-discretionary? That raw data in its own right is very powerful, but it’s not the end game. This is where products like XAI have helped Tic:Toc Home Loans, for example, really achieve the efficiency and automation that we see today.

In Australia, we’re now at a point where data can be shared across a range of account types, including transactions, credit cards, home loans, and personal loans. But not all data providers are actually providing that information now. Very soon they will be. The major banks and major ADIs in Australia will now be mandated to provide that information. This is the data holder's perspective. However, there is also the data recipient aspect. These are the providers or intermediaries of financial services products to consumers. Becoming a data recipient is a lengthy process.

Until more data recipients become active in the ecosystem, consumers aren’t going to be able to benefit from open banking. So what we need to see is a real concerted effort over the next 12 to 18 months to get data recipients and holders to really embrace CDR because it will give consumers a better deal.

The fundamental change is that CDR data is the consumer’s data first and foremost. That may sound like a simple distinction, but no longer is the data coming into your organisation yours where you can put it in your warehouse and do what you want with it. The data coming in through CDR belongs to the consumer and when that comes into your environment, you need to have a level of control and rigour around what you do with it. So, for example, if you want to do a simple analysis of the number of expenses in a period of time, that aggregated value is still consumer data.

The next thing companies will quickly realise is you need to ask your team once you get that data in, what are you going to do with it? A lender might realise that they do not have any functionality internally to identify and categorise data or to identify income. It’s all well and good to get the data in, which is part of the battle, but how do you extract meaning and how do you use any downstream systems? That’s where products in the industry have been built specifically to extract meaning from raw data to help drive automated decisions.

In terms of how CDR will change the way the industry thinks about data, I’d really love to say the answer is goodbye paper. We will say goodbye to a lot of paper, but I think paper is still here to stay. But it’s going to revolutionise the ability for lenders, for personal financial management providers, or any companies dealing with financial data today to offer better products to their customers.

It’s also going to be challenging for a lot of the incumbent providers of these products because everyone’s invested a lot of time and money and smarts in their processes. They’re proud of it. CDR is now coming through and challenging that again. So, you know, how are they going to react? How are they going to respond?

Following on from how incumbents should tackle the CDR shakeup, I think we’ll see more of a b

uy vs. build vs. partner approach. Once you get your data and you don’t necessarily want to build the smarts to try and extract all the insights from the data you might need, or build the integration that goes off to the data holders — because that is complex.

So it’s going to be challenging. And who can you partner with? There are people in the industry and organisations who can provide these services. So once that data comes through, you can extract all the meaning and insight and the results can be really easily surfaced through an API integration with your existing workflow, and through a user interface that is designed for lenders and operational staff to use.

For smaller lenders, the argument in the industry has been so far that the current accreditation requirements are actually against the philosophy of CDR, because it’s blocking those small players that can deliver better products. To be a recipient, you currently need to meet what’s called an ‘unrestricted’ level of accreditation. This is costly and onerous in terms of the info security requirements, your technical architecture, your assurance and risk profiles, and your ability to maintain governance over consent and data going forward. So It’s blocking them from being able to participate adequately.

The ACCC, which is the regulator at the moment, has announced recently that they’re seeking consultation on alternative levels of accreditation. And in my opinion, this is great news for the industry because it’s going to allow more people and more financial services providers to participate in CDR at different levels of accreditation. In practice, the different levels of accreditation will mean data recipients get different amounts of data or they can process data for shorter periods of time. But more players in the ecosystem will make better deals for consumers.

Australia is slowly preparing to roll out the open banking framework to other industries. So soon we’ll have open energy, telecommunications, and eventually perhaps open superannuation. I think over time these will eventuate and the Government, from a legislative point of view, can start thinking about these emerging changes now.

We’ve already seen open banking roll out in places such as the UK. The key difference with Australia is you have the ability to read and write within the UK system. What that means is you can get data, but you can also change it. So imagine you’re a small financial services provider, and you want to build your own online banking app as you’re selling your lending product. You can do that now because with this you can get the data in and you can execute payment through the app that you built against the data holder downstream somewhere. That’s really powerful. And that’s a fundamental shift. Overall, in Australia, we’re still looking at a read-only ecosystem. It will take a lot to adapt.

I think we need to really move towards the next part, which is to make Australia a ‘write ecosystem’. That is permission to both “read” and “write” certain data should be granted. Allowing us to execute that mix is going to create a whole lot of additional value. You’ve got to think about the frequency of data. Today, you apply for a financial service, say once a year, you might make a payment maybe once a week. So having a ‘write ecosystem’ is only going to rapidly increase the consumer’s engagement with the ecosystem and the trust and their willingness to use CDR data when they’re applying for a financial service product.

In terms of other opportunities, Senator Jane Hume recently spoke around the fact that the Australian CDR, the legislation, the way it’s structured, is unique in that it puts the consumer in control. Other jurisdictions are actually looking at this legislation as a model for their own open banking going forward.

So what this means is we could potentially be exporting legislative frameworks to other jurisdictions. From an Australian competitive perspective, Australian companies could be uniquely placed to move into those jurisdictions, allowing them to take advantage of additional competitive opportunities.

You need to ask how the technology is arriving at the decision. AI is at its core a way to automate something, and that something can be a decision. AI is a very broad term for anything that automates a process that a human would normally need to do manually. AI algorithms can often involve complicated transformations of data, which can be confusing in the way the AI arrives at a decision or a recommendation. The fear of the black box if you will.

When a financial service provider transitions to AI techniques they need to be able to take their customers, take their internal staff, stakeholders, and their credit and risk teams on the journey. And the really important way to do that is to always put the human in the loop. The human has to be there controlling that decision. To ease this transition, AI can be thought of as a tool, but not necessarily making the decision. In my opinion, it should help that human in the loop to arrive at a decision faster.

There will be a whole lot of processes that need to be put into place to maintain your ongoing trust in any AI solution in terms of accuracy of automation. You’ll need independent audits by humans, and a human in the loop, where required, to make sure that the automation is clear and accurate.

These can be difficult challenges for somebody who’s built an amazing machine learning algorithm and has managed to productionise it and expose it through an API. Some would think the job’s done. But they then need to take those traditional stakeholders on a journey to explain how it runs and how it comes to the decision. And to explain why it’s going to be right for the business, and how it’s going to be robust to threats and to bias which could change the outcome of those decisions over time. I think the key aspect is that human always has to be in the loop. XAI Validate is a product that does that. It will automate based on a highly curated set of labels that we learn over time from our community. It will automate insights linked to your credit policy, accurately identifying income sources, expenses, assets, and liabilities It will automate that and serve that up to you through an intuitive user interface. And the decision is then always resting in the clients’ hands. These are like recommendations or insights, but the decision ultimately will rest with that user.

Related posts:

Wondering about what is a Data Engineer? In this post, we go through the evolution of this role and how it relates to other roles in the data science field.

Recycler view is a trendy and important view in Android development, and it is a very wonderful view to use. I am going to take you through how it's done in Android development. For my simple screen…

I very intentionally didn’t learn much about this book before reading it. In fact I was so circumspect about this book before first diving in that I didn’t even know it was fiction. This was very…