Many enterprises have been experimenting with Generative AI (GenAI) over the past 12 to 18 months, but few have successfully transitioned from pilot projects to fully integrated business solutions.
That makes OCBC a rare case study in success. Singapore’s second largest bank has been using GenAI in its customer relationship management system to resolve customer issues more quickly. This has led to 20 per cent to 30 per cent improvement in its net promoter score, which measures customer loyalty.
The bank’s “listen and respond” GenAI application was trained on several examples of typical customer feedback, significantly reducing the time and effort required to develop a traditional AI application, said Adrien Chenailler, OCBC’s head of data science.
Without GenAI, he said, it would have taken a significant amount of effort and time to collect and label the data to develop a traditional AI application for this task.
“With GenAI, it took us one week to get this application up,” he told an industry audience at a conference held by cloud vendor Cloudera in Singapore on August 7.
The bank is exploring ways to include more GenAI agents in its major IT systems. One area is streamlining the due diligence process for new wealth customers.
Currently, bank officers laboriously review numerous documents to compile reports. A GenAI application could automate this by extracting relevant information and generating reports, which officers would then validate, thus saving time and effort.
Currently, OCBC has more than 20 AI applications in its business workflow, of which about half are GenAI. The bank’s strategy, said Chenailler, is to leverage GenAI’s large language models to enhance business processes.
Scaling these applications has been challenging, involving the processing of 75 million words daily, transcribing 600 hours of calls, and managing over 1,200 daily users, he added.
At a separate media briefing at the one-day event at Marina Bay Sands, Cloudera chief executive officer Charles Sansbury gave a cautiously optimistic prognosis of market conditions for the next six months.
“The current market dislocations are due to changes in the interest rate and some speculation that US Federal Reserve will cut rates if the US economy softens,” he said.
“However, we still see some economic activity in US and Asia,” he added. “The economy is okay but fragile.”
Even if there is a slowdown in the economy, what matters is where the budgets will be cut. Cybersecurity budgets, said Sansbury, will remain stable while the data, analytics and AI sectors, because of the importance to GenAI, are likely to face minimal cuts.
The Asia-Pacific market is a fast growing region, adding about 61 per cent new customers year over year for Cloudera. Sansbury, who visited customers in the region in the last three weeks, said the region’s younger workforce and IT systems facilitate innovation, with fewer legacy systems to hinder progress.
Abhas Ricky, Cloudera’s chief strategy officer, added that enterprises like telecom operators are investing in analytics to build new products and new revenue streams.
“In other regions, it is about cost avoidance, replacing certain tools which are becoming too costly to maintain, and cost reduction,” he noted.
Asian companies may be larger than their American and European counterparts, he pointed out, but their IT budgets are lower, which forces them to innovate to get more value out of their investment.
