At DBS Bank, customer service officers (CSOs) perform a plethora of tasks each day, from searching for product information to responding to customer queries and writing call summaries based on recall and call recordings.
In addition, they need to review high volumes of offline messages and respond within the bank’s turnaround time. It was crucial to reduce these time-consuming and laborious tasks, which can have an impact on both employee and customer experience.
To alleviate the workload of CSOs and improve customer service, DBS built an an-house generative artificial intelligence (GenAI) tool to streamline workflows and reduce operational burden. Dubbed CSO Assistant, the tool helps CSOs transcribe live calls, access the bank’s knowledge base to address a specific query, triage offline messages and draft responses to customers.
DBS started deploying CSO Assistant for its Singapore customer centre in 2024 following a pilot before extending the use of the tool to other locations including Hong Kong, Taiwan and India within the same year. Powering the tool are the latest GenAI models, including Google’s Gemini and Anthropic’s Claude.
So far, CSO Assistant has benefitted some 1,000 CSOs across the region. The bank’s average call handling time was slashed by up to 32 seconds with reduced manual input, thanks to live call transcriptions and summarisation. The tool’s email triaging capability was also used to categorise over 48,000 offline messages.
And with the tool’s message drafting capabilities, offline messages can be attended to much faster, with high-priority cases responded to in a day with a standard quality. Calls are also tagged automatically to provide managers with insights on customer experience and demand management in the long term.
The tool has been well-received by the bank’s CSOs who have cited benefits such as reduced manual toil and handling time. In Taiwan, within three months of a full roll-out, the average handling time was reduced by 22 seconds with reduced menial tasks. This has enabled CSOs to be more focused on providing customer service while improving employee experience.
DBS’s in-house technology team implemented CSO Assistant in phases, focusing on systems, processes and people. It kept close tabs on service availability and promptly addressed any production defects. Constant communication was also key to ensuring that issues were fleshed out, boosting the accuracy of the CSO Assistant in providing reliable and accurate outputs, such as call or knowledgebase summaries.
Standard operating procedures were also developed to guide CSOs on the use of the tool. And with CSO Assistant being used more widely, the bank monitors operational risks daily and has established guardrails through a risk governance framework.
Plans for refresher training for all CSOs were critical prior to roll-out, and this was paired with daily performance monitoring to ensure the tool was utilised as intended. Change management was put in place, with incentivisation as a key to drive a shift in behaviour and adoption of the tool.
CSO Assistant is part of wider efforts by DBS to drive the use of AI, including GenAI, across its business. In 2024, it had more than 20 use cases that have progressed from experimentation to production and are at various stages of being scaled across the bank.
These include DBS-GPT, an employee-facing version of ChatGPT, to help employees with content generation and writing tasks in a secure environment. It is currently available to more than 25,000 DBS employees.
Other AI initiatives include an enterprise knowledge base that will give employees the ability to search and synthesise unstructured information for various tasks, and the use of GenAI for code and test generation which is expected to speed up software development.
