Close Brothers unlocks RPA with Document Understanding

Close Brothers has been on an RPA journey for more than six years, and much of its work in this area has focused on automating rules-based, structured processes. But Stephen Durnin, the company’s head of operational excellence and automation, says the 2020 Covid-19 pandemic thrust automation around unstructured input, like email and documents, into the spotlight.

“We provide support, which we call forbearance within the banking sector, to customers in the form of a payment holiday if they’re struggling to meet their financial obligations,” Durnin says. “With a high number of customers furloughed at the time, it was important that we were able to lend support at scale and in a timely fashion. To deliver an effective service to our customers and support the contact center by deflecting volume from our telephony channel, we developed a scalable automated process to assess and grant forbearance requests, which integrated with chat bots and self service webforms.”

The pressures of the pandemic led Close Brothers to shift its approach to automation. Its early efforts were pitched at automating activities and making operational processes more scalable — more of a task automation method. Once the pandemic hit, though, Durnin says the company had to adopt a more end-to-end approach, including the use of AI-powered technologies to help it overcome restrictions of traditional RPA. But getting there also required rethinking its existing processes.

Consolidating strengths

In 2022, the company merged its operational excellence and automation teams. The Operational Excellence and Automation CoE established a suite of tools and methodologies to enable holistic “process transformation,” including operational excellence frameworks, Lean Six Sigma, RPA, low code workflow, and dashboarding.

“This was really driving home the point that we need to lean these processes before we automate them, that we want to make sure we get as much throughput within the automation solutions as possible,” Durnin says. “For example, a manual process that bottlenecks will still bottleneck when automated. It’s important to create flow within the process.” 

So it onboarded OutSystems, an AI-powered low code platform, to help orchestrate end-to-end processes. It runs on top of Close Brothers’ legacy platforms and integrates with AI-powered RPA platform, UiPath.

“We’ve created a vehicle to deliver at scale,” Durnin explains. “Rather than delivering a couple of processes at time, we’re automating 10 every couple of months. This is achieved through having an effective approach to process discovery and simplification, creating re-usable automated components and showing agility in your delivery approach.”

Benefits of the right partnership

Close Brothers also partnered with RPA company UiPath, and one of their first big projects was leveraging UiPath Document Understanding in the company’s wholesale finance operation. Prior to the project, the business was facing numerous challenges. Manual processes were restricting the scalability of the service, but most of the documents involved were semi- or unstructured, making them difficult to automate. The result was a long lead time for payouts as deals had to wait for administrators to process them. This was exacerbated by errors or missing information in documents provided by customers, leading to additional work downstream.

“We operate with spreadsheets, email inboxes, and invoices coming in via PDFs, which typically drive a manual process where somebody has to review the document or email, reconcile those documents to make sure they have all the required information, and then go back to the customer if some information is missing,” Durnin says.

By integrating Document Understanding with RPA, Close Brothers could automate the reconciliation of documents submitted by customers and enable automation of the end-to-end payouts process. Document Understanding can pull data from non-standard, unstructured documents like PDFs and feed it into an automated queue. If information is missing, the process can automatically request it from the customer. Then when all the information has been provided, it’s passed to the ops team for review.

The team started with a PoC with three core objectives:

1. Test the effectiveness of the AI module for things like accurately extracting and reconciling data within unstructured documents.
2. Define a process for training and monitoring the solution.
3. Establish a repeatable set of guardrails for security and preserving customer privacy.

The PoC involved training the module on about 1,000 documents and achieved accuracy scores between 95% and 100%, which led Durnin’s team to put the module into production.

“For the majority of the team who work in the sales support function, the original process had been a very repetitive task — checking vehicle registrations, invoice amounts, and dates, among other items,” says Richard Gregson, commercial director for Close Brothers Wholesale Finance. “With the implementation of Document Understanding, the team, who are highly skilled in the delivery of funding and securing title, can now focus on more of the main risks to this process and review some of the key differences in more complex deals.”

Gregson says the team has been enthusiastic in embracing the automation, and notes that blending robotic processes with human in the loop review has been a primary enabler of success.

Close Brothers is taking the power of Document Understanding beyond its Wholesale Finance operation, too. The banking group provides lending facilities to independent contract hiring rental companies, and augments RPA with Document Understanding to provide a faster, more efficient payout service to its contract hire customers. Durnin says Document Understanding will also help it unlock new capabilities within the UiPath Suite, including communication mining and GPT integrations.