Skip to main content

Behind the scenes: Building an Open Banking decision engine

Q&A with Kitty Sadler, Product Manager

by Credit Kudos1 September 2021

Off the back of the recent launch of our Open Banking decision engine, Assembly, we sat down with Kitty, the lead Product Manager, to find out more:

So, Kitty, firstly tell us a little bit about you and your role at Credit Kudos?

I’ve been a Product Manager at Credit Kudos for just under three years. I look after our core insights and tools for consumer lending: our classification engine and prediction models; our income, affordability and fraud products; our risk insights and score; and the platforms for delivering these - our Reports API and dashboard, Atlas.

We’ve recently launched an Open Banking decision engine, tell us in your own words what pain points it is solving for credit providers?

Assembly is an Open Banking decision engine that enables you to create, deploy and adjust automated credit decisions with minimal or zero technical integration.

There are a few pain points we’re solving with this new technology. Many lenders have increased automation at the centre of their roadmap but they struggle to lock in the technical and analytical resources - firstly to design and build automated decisioning, then secondly to adjust it on an ongoing basis as they learn from data.

For some, existing technology and complex internal systems make it difficult to implement new processes and be sufficiently agile to continually adjust these to optimise performance.

Others don’t have any resources in-house and rely on their loan management system’s team to make policy changes for them, which can be costly.

Others have identified the opportunity to target new, underserved populations using Credit Kudos data but have difficulty prioritising this. On top of the additional credit risk (lending to a new population), they need to invest significant technical and analytical resources in setting up a trial, then put aside further resources to continually adjust and refine it. This means the strategy can end up being deprioritised in favour of bets that are safer though potentially less lucrative.

Assembly enables these lenders to design, test and deploy credit policies on existing or new populations without any initial technical or analytical outlay. Once they’re happy with a policy, it enables them to integrate it into real-time decisioning with just one small piece of technical work. Then they can continually adjust these policies without using any technical resource at all, and far less analytical time - no data exports or spreadsheets.

What has been the most exciting part about developing Assembly?

So many of our clients had told us about these pain points that are holding them back from growing and reaching underserved populations, so it was very exciting to present early concepts to some of them and for them to respond “wow, yes, this is exactly what I needed but I didn’t realise it.” We knew very early that we’d solved the right pain with the right solution, it was just a case of refining it.

What was the greatest challenge when developing Assembly?

It’s a complex product, so we invested a lot of time in the product research phase to make sure that we developed an intuitive product with all the right language so people could build their policies in Assembly without a big learning curve. We invested a lot in testing different options with research participants to help refine this. We also had to ensure that the rules builder was simple to use but gives lenders complete flexibility, so there was a lot of sketching out of policies on paper and whiteboards.

How have your early access clients reacted to Assembly?

We’ve been super pleased with the reaction - the early access group has been able to get up and running with policies very easily. They’ve also been involved in helping us build out the roadmap to support them as their use cases grow and evolve, which is a great sign.

Any hidden benefits you want to shout about?

Firstly, Assembly enables you to manage different policies for different parts of your funnel - all through one single integration point. Once that integration point is set up, you can create new policies for different segments with no integration work at all. For example, you could get started with an automated affordability policy for any applicants that are referred for affordability. Then, later, you could decide you want to explore growing your loan book and set up a different policy for thin-file applicants. You wouldn’t need to do any extra integration work to trial and run this alongside.

Secondly, you can combine other data sources, for example, a bureau score or applicant-stated income. This allows you to manage all aspects of the policy in one place. It’s very easy to supply this data programmatically - you just add it to your existing Connect Flow integration.

What's on the roadmap for the next steps with Assembly?

Lots! You’ll soon have lots of flexibility to segment each of your populations - for example having a different income verification policy for self-employed and PAYE applicants.

We’re also planning to let you use Assembly to build your own metrics and properties, for example, a custom risk indicator or affordability calculation based on your particular strategies.

We’re also planning to give Assembly even more data science power - helping you to analyse and manage risk across your whole profile. You’ll be able to understand your pass and default rates over time and break this down by segment.

In short, this technology is a powerful and flexible tool that helps you to transform your operations and grow your loan book with minimal or no resource, and there’s lots more to come!

For more info on Assembly, visit the website here.