iamchamp - stock.adobe.com

How Microsoft is easing GenAI adoption into financial services

Microsoft is supporting enterprise deployment models, addressing the risks of the technology and leveraging its industry cloud capabilities to ease generative AI adoption in the financial sector

At FWD Group, plans are afoot to utilise generative artificial intelligence (GenAI) to enhance customer service and marketing, as well as in underwriting and claims operations.

The Hong Kong-based insurer recently partnered with software giant Microsoft to leverage the latter’s Azure OpenAI Service and advanced AI models to drive its GenAI initiatives.

Since embarking on its AI journey in 2019, FWD has deployed nearly 200 AI models across its business, encompassing more than 600 use cases. Additionally, it is an early adopter of Copilot for Microsoft 365, an AI companion that aids employees in their day-to-day tasks.

Bill Borden, Microsoft’s corporate vice-president of worldwide financial services, noted that employee experience, customer experience, operations – including fraud management – and the development of new business models are top priorities for financial services firms implementing GenAI initiatives.

While it is still early days before the industry sees new business models emerge from the use of GenAI, Microsoft has been focused on driving and easing adoption of its technology, including its reportedly $13bn investment in OpenAI.

Three years prior to its alliance with OpenAI, the company had begun to transition its entire product development team towards incorporating GenAI capabilities and large language models (LLMs) into its products, according to Borden.

“We’ve done that with a guiding set of principles on how we design AI in a safe and responsible way, which we share with our customers – that’s why we’re able to move forward with speed and pace to get these products on the market,” he added.

With issues around toxicity, hallucination and appropriateness – risks that by and large had not existed in traditional AI – having design principles to address such issues is key to easing adoption of GenAI in regulated industries, especially in use cases with higher financial or reputational impact.

To that, Borden said the responsible development of technology is part of Microsoft’s heritage: “We’ve continued to build on that as we put our frameworks together for development around AI, working with Microsoft Research, using our world-class systems and engineering processes to develop products, and then looking at governance and policy management around that.”

For example, when it comes to copyright issues around LLMs, Borden said for developers who use Copilot Studio to extend or build their own copilot capabilities, Microsoft is confident of its guardrails and controls that it will defend customers who face AI copyright challenges.

Microsoft also works closely with regulatory and government bodies to understand the power of what GenAI can be, and shape the management and use of the technology to benefit society, Borden said, adding that it has written whitepapers on responsible AI that it has shared with industry, academia, governments and civil service communities.

“We’ve been working with regulators around the world for the past 10 to 11 years as we rolled out cloud technology, and now as we start to roll out GenAI, we’re extending those conversations with regulators to educate them and hear their point of view.

“They haven’t come out saying that there’s necessarily a need to regulate – there’s talk about it, but right now, the conversation is that institutions should continue to leverage their risk frameworks and control structures to build and use AI responsibly.”

With Azure OpenAI Service, financial institutions can harness GenAI capabilities in an enterprise-oriented model where they can apply their own control structures, development procedures and cloud capabilities, Borden noted.

“You can also bring in APIs [application programming interfaces] from OpenAI’s ChatGPT, open-source models like llama from Meta, your own models or small language models,” he said, adding that no customer data will be used to train external models.

On how Microsoft’s GenAI capabilities fit into its broader cloud portfolio, Borden said data and AI has been core to its integrated industry cloud offerings, which its partners can build on to provide differentiated services.

“Microsoft has always been a platform-oriented company, allowing other technology providers to build on our tech stack,” Borden said. “The thousands of financial services and application providers that are building on our technology stack, not just compute but also collaboration, leveraging data and AI, is what makes us unique to serve the financial services industry.”

Read more about artificial intelligence in APAC

  • The Sea-Lion large language model was built to cater to the language and cultural diversity of Southeast Asia, which is currently underserved by existing models that mostly originate from the West.
  • Google Cloud has teamed up with the Singapore government on a slew of initiatives to drive AI adoption, build an ecosystem of AI startups and expand the pool of AI talent in the city-state.
  • AI can be transformative in government, but its implementation has to be done in a way that drives positive outcomes while mitigating its downsides, according to the head of Australia’s Digital Transformation Agency.
  • Culture Amp is building a generative AI capability that summarises employee survey responses, automating a process that typically takes HR admins up to hundreds of hours to complete.

Read more on Artificial intelligence, automation and robotics

Data Center
Data Management