Insurance firm QBE saves 50,000 hours annually with Pega Robotic Process Automation

Derek du Preez Profile picture for user ddpreez March 10, 2022 Audio mode
QBE European Operations has used Pega bots to help automate the underwriting and claims process, saving the company hours and improving customer satisfaction.

Close-up of robot examining financial report with calculator on desk © Andrey_Popov - Shutterstock
(© Andrey_Popov - Shutterstock)

QBE Insurance Group is one of the world’s leading insurance and reinsurance companies, operating in 37 countries across the globe and brings in $17 billion in revenue annually. The organization uses Pegasystems globally as a workflow platform, but its European Operation has been leading the way in adopting Pega’s Robotic Process Automation (RPA) to drive efficiency, improve the value of employees’ work and boost customer satisfaction. 

To give you an idea of how successful the use of Pega bots has been thus far, some of the headline returns from their deployment include: 

  • Automating an estimated 30,000 customer claims tasks per week

  • Saving approximately 50,000 hours, the equivalent of 25 full time employees annually

  • Dramatically speeding up the process for its customers - with customer complaints halving in that time

We got the chance to speak with Amit Dixit, head of operational excellence at QBE Europe, who provided detail on how QBE started with RPA, how it’s being managed, and what potential it holds for the future. Central to this is how Pega is used to manage the flow of work across the organization and between different stakeholders. As Dixit explained: 

We are an intermediary business. We work with a lot of brokers and are involved with anything to do with the underwriting process, whereby we are evaluating the risk, pricing and then eventually handing it over to our shared services to be booked onto the system. The entire workflow is managed through Pega for almost all the lines of business. 

One of the core reasons why I was brought into the organization was to make it more efficient and to make sure the processes are lean and genuinely adding value - removing the waste from the processes and reducing complexity. 

As part of that I did a lot of work with underwriting, across claims, credit control, to understand these processes. And as I was going through it, we realized that some of these processes deal with pretty structured data and a lot of the employees are involved in doing repetitive, non-value added work. 

A growing bot workforce

As such, front of mind for QBE and Dixit was how the company could automate this highly repetitive, non-value work. How could it release the capacity so that the frontline employees are using their time servicing the customer, rather than inputting data. 

Claims soon emerged as the most obvious use case, where QBE pays approximately $2 billion in claims every year. Dixit said: 

The volume of transactions, paperwork, and the stuff that goes along with that - the view was, can we look into our claim payment process and automate that? So that was our very first use case of deploying Pega Robotics.

QBE gets a lot of information and data from third party providers that it uses to service the claims, which it needs to process and transpose onto its own systems. Employees were doing this work previously, but it seemed like an obvious area for bots to be deployed. Dixit said: 

What we did was work with our suppliers to get data in a slightly more structured format and then we put automation on top of that, so it is extracting the data from these Excel or Word documents, and then eventually transposing the data onto our systems. 

Once we got our first use case up and running there was genuine excitement. This really works! The product has the capability to actually do what we want. So from then on it just accelerated. 

After initial pilot success in claims, Pega was quickly rolled out to credit control and underwriting operations, with RPA solutions now operating across the business, where the key focus is pulling data out of systems and replicating it onto QBE systems to save employees’ time. Dixit said: 

We have about 80 bots now across claims and underwriting. They probably do about 40,000 transactions a week. And there are about 100 users who have automation bots running on their desktops. We have numerous cases whereby we move data across systems, and those are legacy systems, they’re not API enabled. We have built bots where they move data from one system to another, but there are also bots that move data from our systems to external systems. 

I have a dedicated team of four individuals that service these bots. Bots don’t break, but what happens is the underlying system changes. So it’s about coordination with the leader of these systems to make sure that everyone is in the loop and on top of the system changes that are happening. We need to be aligned to the changes that are happening, so there is a bit of maintenance that’s required, but it’s not massive. 

Automation is one of the key enablers for QBE’s digital transformation - and the work with Pega has laid the foundation for future rollouts. The bots have already resulted in the organization paying a motor claim, on average, 46 days quicker than it was previously. And customer complaints have halved too. 

On plans for the future, QBE is now sharing the success it has had with its North America and Australia divisions, to see how they can go about identifying similar use cases in their verticals. And Dixit is also thinking about how bots could be deployed internally across the company’s finance and HR functions.

How to succeed

Dixit has two key pieces of advice for other organizations that are considering the use of RPA. Firstly, he said, use cases should be focused on processes that involve repeatable tasks and use structured, high volume data. However, he also added:

When we started this journey there was a debate around: do you need to actually redesign the processes before you can deploy the automation? Or can you do it without doing that? My take on this is that you can redesign the process, but what’s absolutely critical is understanding the end to end process. 

The last thing I wanted is to look into a small subset or a segment of the process and just deploy a bot on top of them. So one of the things we worked really hard on is having that end to end understanding of the process and then making a very conscious call that yes, this is where a bot should be deployed, and this other set of activities should be done by humans. So not all processes were redesigned, but all processes were well understood. 

And secondly, business engagement is key. He said: 

Typically what I would advise is pick a small use case and pick a business partner that is genuinely willing to work with you. If you are pushing this tech onto certain teams, in my view, that doesn’t work. Make sure that the first use case is successful and then promote it heavily. Once we did that we were snowed under, getting requests from numerous teams. 

A grey colored placeholder image