Logo Reinfer

Deutsche Bank: Scaling service and improving the client experience

Deutsche Bank: Scaling service and improving the client experience
Company
Deutsche Bank is one of the world’s largest banks and most widely recognised financial services providers. It provides services in investment banking, trading, private wealth, asset management and retail banking to over 27 million customers in 2,500 branches across the world.
Industry
Finance
Location
Global
Employees
83000
Customer since
2021
“With requests coming in through multi-channels, unstructured comms is a huge part of day to day life at Deutsche Bank. With that backdrop, how do we respond quickly to clients? How can we help clients facilitate payments and transactions? Being able to navigate and create structure out of unstructured data makes it easier to respond to our customers, faster.”
<small>Rushabh Shah, Program Director Intelligent Automation, Corporate Banking Technology, Deutsche Bank<small>

The challenges

The client experience is one of the most important differentiators on which investment and corporate banks compete. While Deutsche Bank is renowned for its customer experience and portal strategy, its leaders knew they had to do more to stay ahead of the competition.

Deutsche Bank’s clients expect fast, reliable and high-quality service at all times. Yet this isn’t easy to provide. Deutsche Bank maintains a corporate banking service workforce of 12,000 people spread across more than 50 branches globally. These agents spend their time performing comms and admin work - speaking to, responding to and processing customer requests. The quality and responsiveness of this service team can’t be scaled up without a massive increase in headcount - something that’s unsustainable in the long term.

The banking leader receives hundreds of thousands of customer messages every day, primarily taking the form of emails, calls and chats. Many of these requests are complex and require a service person’s expertise, but the majority are simple, purely transactional requests that require only a couple of responses. If Deutsche Bank could find a way to automate some of this transactional workload, it would not only enhance the client experience but greatly amplify employee productivity as well.

70% of inbound message volume to Deutsche Bank is email.

The solution

Automation is present at every level of Deutsche Bank. The banking leader uses a comprehensive suite of intelligent automation tools, including OCR, computer vision, case management and BPM platforms to tie everything together. To date, Deutsche Bank manages and maintains over 600 different bots across its operations to deliver greater speed and efficiency.

The bank wanted to introduce its automation pedigree into more of its customer-facing service workflows. However, customer requests and client conversations have been difficult to automate due to a lack of structured data. Furthermore, the bank knew that any form of service automation had to engage clients and speak to them in their own language, with the accuracy and personalisation of a human conversation. Customer portals and chatbots could not provide the level of personalisation and accuracy desired.

Deutsche Bank had previously found some success with natural language processing (NLP). It had built and maintained its own NLP models for integration with tools like Salesforce to enable self-service and automation. However, it was having difficulty scaling NLP beyond certain limited use cases. It takes time, resources and expertise to develop, train and maintain NLP models. But the context at the bank was always changing, with people moving teams and budgets fluctuating. The data science expertise required was expensive to scale and growing harder to maintain.

Ultimately, sustaining this process of model development across the bank’s many thousands of service desks simply wasn’t feasible. That’s why Deutsche Bank adopted the Re:infer enterprise NLP platform. Re:infer provides businesses with simple, powerful tools to mine, monitor and automate their service and communications workflows. It does this through a fully no-code platform, allowing users to rapidly train and use their own NLP models, regardless of technical ability. With multilingual models able to understand and interpret the world’s most popular languages, Re:Infer was the ideal solution for the global bank.

The data science expertise needed to develop NLP models at the bank was expensive to scale and growing harder to maintain.

The results

Re:infer has helped Deutsche Bank scale NLP across its service function, improving agent productivity and enhancing the client experience. Deutsche Bank found Re:infer to be fast and cheap to get into production. Thanks to Re:infer Active Learning, models were quick to train in the platform’s no-code user interface. After minimal training, its service teams have been able to create their own custom NLP models and get them into production within days rather than months. This is all possible without the requirement for expensive data science expertise.

With Re:infer Communications Mining, Deutsche Bank’s service teams have been able to gain unprecedented customer insight. By analysing their inbound communications en masse, they are given a clear picture of what’s driving urgency, complaints and client demand. Rather than experimenting with changes to see if that improves the client experience, agents can immediately see and target problems at the source.

Deutsche Bank is also laying the foundations for message and email automation. Re:infer is able to extract entities - key concepts and phrases like dates or customer numbers - and pass them down to automation tools for downstream processing. This will enable service teams to automate, from end to end, many of their highest-volume client requests. This will free agents to focus on only the most valuable and complex work, while also reducing handling and response times for all types of requests.

We’re using Re:infer to unlock insights and kill waste. I personally see it as a process mining tool for our email data and unstructured comms. A tool like Celonis is good for analysing workflow logs, but Re:infer takes it to a new level by helping us see the data in our emails - to understand the intents of our clients and their reasons for contact.
<small>Rushabh Shah, Program Director Intelligent Automation, Corporate Banking Technology, Deutsche Bank<small>

Watch the Deutsche Bank story

Deutsche Bank’s Rushabh Shah joined Intelligent Automation Week to explain how Re:infer is helping his team to scale service and improve the client experience. Watch his full presentation:

Turn any conversation into actionable data

Drive operational efficiency and optimise the customer experience with Re:infer’s Conversational Data Intelligence Platform.