- During Cloudera’s Evolve in New York, executives discussed how the company helped OCBC Bank integrate AI across the organization.
- This 91-year-old bank is one of the first in the world to implement generative AI tools on a large scale.
- It refuses to use the Open API, so it retains control over input and output.
This month, OCBC Bank, after a six-month trial, unveiled OCBC ChatGPT, becoming the first bank in the world to embrace generative artificial intelligence on a large scale. The head of the bank’s group data office, Donald MacDonald, said OCBC is deploying the chatbot to 30,000 employees in 19 countries. Chatbot, created in collaboration with Microsoft Azurestarting this month, it should help employees in the bank’s 420 branches and offices around the world with tasks such as writing, researching and brainstorming.
While OCBC ChatGPT is just one part of the bank’s overall effort involving generative artificial intelligence, OCBC is said to be using or piloting four other functions. They are generally categorized as: “Wingman”, who helps his team of coders write code; “Whisper,” which transcribes voice calls and creates summaries for its contact centers; “Buddy,” which extracts information from 150,000 pages of company documents and records meetings for staff; and “Document AI,” which digests documents like financial reports.
The Singapore-based bank wants to use generative artificial intelligence to personalize customer interactions, suggest stock purchases, and detect fraud and suspicious transactions for the bank. According to MacDonald, AI already makes more than four million daily decisions about risk management, customer service and sales for the bank. Ultimately, he expects that number to grow to 10 million by 2025 generative artificial intelligence takes over more functions.
But behind all these possibilities is a platform that makes it all possible.
OCBC and Cloudera AI collaboration
The challenge for enterprises in adopting generative AI lies in the fact that organizations must provide third-party AI tools with access to their specialized knowledge and exclusive data in order for the model to deliver accurate results. However, there is a risk of exposing confidential information without proper precautions.
This highlights the importance of optimal hybrid data management for organizations that use third-party AI solutions proprietary data. In OCBC, a hybrid cloud platform such as Cloudera has been helping the bank gain value using AI and ML for years. “Their (OCBC) success delighted us. I don’t think we expected that. Those guys took our platform, took it apart, pushed it to the limit, integrated with other ecosystems and created their own platform,” said Remus Lim, Cloudera’s vice president of Asia Pacific and Japan. Tech Wire Asia on the sidelines of Evolve, New York.
Lim explained that OCBC had already been working on this a few years ago before generative artificial intelligence became a mainstream topic. Apply Cloudera’s blog shows that in 2015, OCBC began a multi-phase initiative with Clouder, aimed at providing customers with access to its banking services through a simple, convenient user interface that provided targeted and customized products and services.
“They started five years ago, setting up an AI Lab and recruiting the right people. Today, they have about 200 data scientists,” Lim pointed out. OCBC eventually migrated to Cloudera Data Platform (CDP) and CDP Machine Learning in 2022 to drive several solutions that increased operational efficiency, enabled new revenue streams and improved risk management.
“One of the key things that OCBC emphasized is that it does not use the Open API, because it cannot control what goes out and what users input. This means that OCBC GPT is confined within a highly secure and controlled local environment. And that’s in our platform, CDP and our machine learning,” said Lim TWA.
According to Clouderawhat OCBC did was build a single entry point for all of its LLM use cases: a hybrid framework that could seamlessly integrate multiple data sources, including inputs from thousands of users, and a private cloud data lake that would keep users’ data secure, how would become real – time insights adapted to company standards.
The bank built fast microservices to access LLMs stored on its local servers as well as LLMs available in the public cloud: a cost-effective model that allowed it to use public LLMs in the cloud and host open source LLMs, depending on functionality and customization that is required. By deploying and hosting its code assistant, customized for 2,000 users, OCBC saved 80% of the cost of using a SaaS solution.
The platform integrates with the bank’s ML operations and fits into its larger ML engineering ecosystem. This cloud-based ML platform enables OCBC to build its applications and use the tools and frameworks of its data scientists’ choice. OCBC could also, with ML models, send more than 100 different personalized incentives on its mobile banking app, informing customers about financial opportunities, including eligibility for a new credit card or loan—achieving click-through rates of up to 50%.
The initiative led to a more personalized customer experience, higher campaign conversion rates, faster transactions, reduced data center downtime and an additional SGD 100 million ($75 million) in revenue annually.