Companies are racing to adopt large language models (LLMs), which serve as the foundation for generative AI applications that could radically transform enterprise operations and productivity. This momentum has fueled investment of over $16B in funding to LLM developers in 2023 alone.
Although financial services players have been slower to move — largely due to concerns around compliance and data quality — they are increasingly taking steps to integrate LLMs into their offerings.
However, how they do this varies from one company to the next. Incumbents are taking several different approaches to deploying LLMs, including:
- Adopting LLMs to streamline internal workflows, such as document processing
- Launching their own LLMs to offer generative AI-led products
Meanwhile, B2B fintech companies are leaning on LLMs to enhance their offerings for corporate finance teams.
We dive into each of these strategies below.
Financial services firms are adopting LLMs to streamline internal workflows
Many financial services firms are turning to LLM developers to build out their generative AI capabilities. OpenAI, Cohere, AI21 Labs, and MosaicML all list financial services players or fintechs as customers.
Initial pilots of generative AI in financial services have primarily focused on back-office workflows. To this end, Citi, Deutsche Bank, Mizuho, and MUFG have each given their employees access to LLM tools to increase efficiencies across use cases like document processing, reporting, drafting memos, and handling internal inquiries.
For example, Mizuho has rolled out Microsoft‘s Azure OpenAI service to 45,000 employees in its lending units in Japan.
Meanwhile, Citi is planning to grant access to generative AI tools to the majority of its 40,000+ coders following a pilot with 250 of them. The bank also stated that it used generative AI to comb through 1,000+ pages of new regulatory documents for the US banking sector.
