Jupiter Asset Management Taps RPA for Manual Process Automation

Jupiter Asset Management is currently reviewing its manual processes to determine which of them could be automated, which the firm says would increase processing speed and remove human error.

Paul Kay, Jupiter’s CTO, says robotic process automation (RPA) is one of the tools the firm is using to re-engineer processes that could be done more efficiently.

Kay describes this as intelligence process automation. “We are probably tackling 30 or 40 processes this year, maybe even 50,” he says. “What we will do is try to remove any of the manual work and make them more efficient. We will get rid of some of them, and we will combine some of them, so maybe by the end of the year we will have half as many processes—and RPA could be supporting 20% of those in whole or in part. That is the work we are doing right now.”    

For example, Jupiter is using RPA for bid-list reports to comply with the City Code on Takeovers and Mergers, a set of binding rules and principles governing the conduct of takeovers in the UK. At Jupiter, meeting these requirements is a daily activity where a user has to manually pull together all the requisite data, with a second user then having to confirm that all steps have been correctly followed. Currently, this process takes 30 minutes to an hour. Automating it will cut that down to five minutes, and the risk of human error will be eliminated. The firm expects the process to go live in the next four to six weeks.

RPA is normally applied to less sophisticated manual tasks. When financial services firms began working on RPA a few years ago, there was a lot of hype around the technology. But it has drawbacks, one of which is that information entered into the bot needs to be very precise. Even minor changes can paralyze operations.  

Indeed, Kay says that when Jupiter first started working with RPA, the firm thought it would be applicable in more areas than it actually is.

“We thought the robots would be super-reliable,” he says. “But what is happening is they are kind of like humans. You come in on a Monday morning and the world changed. Perhaps someone has changed the website, and the robot is programmed to go to that website and download data on the Monday morning. It can’t quite pick up the data that it needs, so you have to go in and tweak the robot. It is like the robot has gotten sick that day.”   

Another problem is the amount of effort it takes to shift tasks to RPA. Kay says return on investment is not obtained if it takes too long to automate the process. If previously 50 people were doing exactly the same thing every day, Kay says that would be the type of process to automate. “But if you have someone doing a process once a week, or once a month, and you need to spend months building it, [that] doesn’t totally make sense,” he says.

The intelligence process automation team at Jupiter is composed of five people and looks after RPA and other automation tools. Kay says he was a big believer in robotics, having previously worked at banks with big operational departments. He says there are a lot of opportunities for RPA-driven applications, but he has found that Jupiter didn’t have the large size and the scale to deploy the robots and take out 10 or 15 people who are doing exactly the same jobs.  

Now he says the firm is going through various processes and asking why they are carried out. 

“Rather than just applying one tool, the RPA tool, we ask if the process should sit in one of our strategic platforms going forward, so the Aladdins or the Salesforces, etc.,” he says. “Or is it something we are going to have to do manually, and if it is, what is the right tool? It could be RPA, it could be a SharePoint, and it could be like Microsoft Forms. It could [also] be a little bit custom code. The key point is that we take the time to understand the process first and then decide how to re-engineer it, what the best target state is and what tools to use.”

Towards a Golden Source

Outside of RPA, Jupiter is also engaged in a broader project to create a single source of truth for data, Kay says. The aim is to create a golden source of trusted, accurate information for each dataset the firm creates. This information would be used by consumers of that data, whether those consumers are in operations, sales, fund management, or are on the client side.

Kay says Jupiter has not had a full handle on the overall management of the availability and integrity of data within the firm. “We need to shift that, and make sure that we are looking at the data as an asset—really looking after it properly, making sure that it is always up to date and fresh,” he says.

Kay expects the program to run for about another year. “Then what we will need to do is consider whether we need a chief data officer, or a data management office,” he says. “We have to maintain the data management framework and maintain our data to the level that we want for the rest of time. That is what we are trying to put in place—that core foundation.”