AI-Driven Innovation: Financial Services Simulation

Simulation

Rapidly identify ingrained ways of working through a pressured Simulation

We selected a group of 20 leaders from the product team of +800 and ran a simulation to inspire them towards a new way of thinking. Using motorsport to disruptive their way of thinking about data, we broke them up into competitive teams each with diferent areas of data responsibility and diferent ‘business missions’. Using our technology, we deliberately exposed them to silos where they could only communicate through an inbuilt IM system and gave them access to an AI chatbot that could answer all the questions they had during the race relating to their strategy.

Here, the team were exposed to real life behavioural challenges that they see each day. Poor alignment of analytics to business priorities, weak questions to analytics that simply reinforced their plans and (in a fun way) exposed them to AI where they were confronted with issues ranging from trust to thinking of AI as a colleague.

Empowering the front lines to make AI driven decisions.

The result was a clear set of behavioral gaps and ways of working friction points, that we were able to focus on during the next phase of the project; indeed most of the issues encountered in the Sim, were also issues that the leaders faced day to day at work too.

Tackling the business problem - Late Payments

We selected a high value business challenge: late payments. This was a strategic issue to the executive team and was seen as a priority to solve and in doing so prove the value of what an efective data culture can do. Whilst leaders knew what the problem was, they were unclear about how they could make analytics work hard towards increasing their understanding against it. This was all about formulating better questions.

Cross functional agile teams

We set up a cross functional group- people from diferent parts of the product team who didn’t typically work with each other everyday but close to the business challenges and importantly had the right level of domain knowledge. Importantly, we included 2 members of the Enterprise Analytics Centre throughout the whole project. They were the custodians of the data and were to be set the challenge of answering the questions that we formulated.

Putting business value front and center

We identified key themes to investigate and formulate questions against:

• Payments outside of the billing cycle
• Late payments across other channels
• User behaviour in terms of payment channel selection
• Billing shortages
• Notification preferences

We worked in short sprint cycles to keep momentum. This involved a 1-2 week sprint against each business area, during which the project team either joined a stand up call led by one of the team members or got together to run a short meeting exercise. We supported through effective coaching around the Team AI Code, providing performance support for the translators so that they could deploy at any stage when they got stuck.

Reveal hidden connection between issues

For example, we ran a Key Assumptions check during the sprint that in a 60 minute meeting was to expose the team to faulty logic, identify some weak assumptions and as a result (most importantly) surface some high value questions:

Hypothesis

Clients are unintentionally late on their payments

Weak assumptions, highlighting gaps in understanding

• Clients know they need to make a payment and when that payment is due.
• Clients are receiving effective communication as to the status of their payment.
• Clients know their options for making a payment.
• These options are sufficient for making a payment how they want.
• Clients always have the funds available to cover the payment amount.
• Late payment notice, triggers action.

Better Questions

• Email vs Physically mailed bill notification. Which one is more effective?
• What communications are clients receiving today about their bill status?
• Are clients that pay less frequently late most often?

Insight

The method with which customers receive notices and pay their bills does impact the amount of late payments made monthly, and moving customers to electronic payments has increased the incidence of late payment by 50%

Making change stick

Rather than leaving the team with some powerpoint templates and fun memories, we provided them with a tool that they could then take back to their teams and deploy against other projects. This for us really is the difference between sustaining success and scaling data culture change and just a one-of success.

Business value

Modification of e-payment channels & late payment notifications

Summary

Through inspiring a data culture with a project team, against a demanding business challenge, we were able to achieve value in the following ways:

• Build a capability of Translators to bridge the gap between business and analytics.
• Improve company wide demand for the services and resources of the Enterprise Analytics Centre by asking better questions.
• Demonstrate how working in a cross functional agile way stimulates alternative perspectives.
• Provide a scalable way of working for analytics and business team interactions.
• Drive hypothesis-led thinking and provide structure to business and analytics collaboration.