Making employees "super human"

UBS is placing a big focus on automation. As an example, the firm as a whole has more than 1,000 robots performing mundane, manual, repetitive tasks alongside their human counterparts. More than 70% of them are being used to automate back-office processes for the Operations function. In an interview conducted before the coronavirus pandemic took hold, we talked to Chris Gelvin, Head of Group Operations, on Operations’ journey to transform digitally, the value this is bringing to his team, and how they're just at the start of what will be a far-reaching digital transformation. In recent weeks, the Operations team has delivered an automation solution that enables UBS to process thousands of applications in the Swiss government's loans program for small and medium sized businesses, designed to cushion the impact of COVID19.

Gelvin: Our function sits at the heart of a very complex landscape. We support the needs of every UBS client in every business in every market. We touch more than 1,000 technology applications on any given day and many of our processes still involve handling physical paperwork. Right now is a pivotal moment for us. Given the increasingly sophisticated technology that's becoming available, we have an amazing opportunity to reimagine how we work and move into a much more digital, automated direction.

We're working on a number of different initiatives that are designed to enhance the client experience. For example, we've partnered with the Investment Bank on a dashboard that gives us near real-time securities settlements information. Having that digital data available to us allows us to understand 'normal' behaviors so we can spot likely exceptions early and step in to prevent them. As another example, we're looking at deploying chatbot technology to assist clients with simple enquiries. We're also very focused on ways to automatically extract data from the various forms and documents that we currently have to process manually.

What does this involve?

Basically: moving from paper-to-data. Right now we receive a lot of unstructured data information coming in on paper forms, e-mails, and PDF attachments. We want to get as much of it as possible into a structured from that is digitized, so that this data can feed seamlessly into our downstream applications. The more data we can have in structured form, the more opportunity we have to let machines do what they do best: analyzing that data and giving us insights humans would never see.

Does this increase in automation mean you're reducing the number of humans?

Gelvin: Actually, no. Automation is having a very different effect. As the different technology possibilities become clearer and clearer, we want to be able to take advantage of them by using our human capital in a new way. Our vision is to get to a place where we are creating "super humans" out of our UBS Operations team.

I want us to bring human and technology capabilities together to create something that uses the best of both worlds. We humans are really good at certain, irreplaceable things: leading, empathizing, judging. Meanwhile, machines have other skills: calculating, analyzing and working with huge sets of data. I see an enormous opportunity for us in UBS Operations, to become the thinkers, designers and custodians of how the bank operates – but now with tools that make us more effective than ever before. But we need to get the foundation in place first – and paper-to-data is a major part of that.

What are some of the technologies you are using for your 'paper-to-data' focus?

We're exploring a number of options. We're already using Optical Character Recognition technology to extract data from scanned paper forms, and we're exploring the use of Natural Language Processing to mine the many millions of e-mails we receive and to categorize them. The more we can minimize the manual handling of information, the more we reduce error and risk. We've also been an early adopter of using robots to automate data handling, and have quite a lot of success since starting this work two years ago.

What kind of robots are you working with?

Gelvin: We're using the term robot, or bot, as shorthand for Robotic Process Automation (RPA). RPA is coded computer scripts that have been programmed to follow a set of rules and to mimic the actions of a human in our systems. In other words, they can toggle, cut and paste, across applications, Outlook, and Excel just like a human can. So, although the work of each bot is monitored by a supervisor, and they have personnel identification numbers assigned to them in the same way as our employees do, they're more like a macro than a person.

What kind of work are these bots performing for you?

Gelvin: We're using RPA to execute repeatable processes at the key-stroke level, the way a human would. Say someone has a task that involves going into two or three systems to cut and paste data into a report, which is then PDFed and emailed to a client: a bot can now do all those steps. When you watch these things at work, the bot is simply opening and closing windows like we would, cutting and pasting and PDFing and emailing in the same way you or I would. Only it's doing it at a speed that is basically relentless, and a robot never makes a typo, they can operate all night, and they don’t need to take a break.

What are the main benefits you're seeing?

Gelvin: As well as improved timeliness, the fact that these tasks are completed with high levels of accuracy and completeness has a material effect on our ability to enhance the UBS client experience because reducing errors means reducing risk, rework, and frustration from our clients. But in many ways, the most important benefit for us is that it's creating additional capacity in our teams, which we can use to focus on more meaningful work instead – such as interacting directly with clients to solve problems. We're also taking some of the additional capacity we gain with RPA and using it to fix the underlying process issue that caused us to build the bot in the first place.

How many bots do you plan to have in production?

Gelvin: This answer may surprise you, but ideally the number would be zero! The best type of robot is the robot that we didn't have to build because we could address the issue through a process change or a traditional technology change. The second best type of robot is the robot we could turn off after we made that process or IT change. The third best robot is one that operates effectively, efficiently and transparently.

As part of the overall robotics program, we track both the number of bots implemented as well as the number of bots that have been decommissioned. It's a great day when we can turn a robot off! And in UBS Operations we've turned off more than 200.

Gelvin: The DNA of our UBS Operations team has already changed in a really short time thanks to our enthusiastic adoption of RPA and other early-stage technologies. More than 2,000 of our staff are already trained and able to run robots as part of their role. After only two years, it's just a business-as-usual activity for a huge number of our people. Meanwhile, we've built out a core of around 350 developers who are designing, building and maintaining robots full-time.

What have you learned from this transformation within your team?

I think one of the most interesting things we've learned is how fundamental – and irreplaceable – specialist Operations knowledge is. While many of our RPA developers have been hired in, about a quarter of them were previously in traditional roles in UBS Operations. Whatever technology solution you are applying, it's critical that you have people involved in its deployment who understand our processes, our flows, our risks, and our clients. So, we're leveraging people with deep Operations expertise alongside new hires with specific technical expertise. It's a whole new career path.