Moody’s Corp. is rolling out new technology backed by generative artificial intelligence that it hopes will help staffers shave hours off the process of analysing vast troves of data and writing reports. 

The ratings agency is providing employees large language models from Google Cloud that will quickly sift through both public documents and the firm’s own database of information to help them write analysis, Nick Reed, the company’s chief product officer, said in an interview. The aim is to enable a range of staffers to work on projects that previously would have required expertise in fields such as coding, finance and accountancy, he said. 

“What could potentially have been a day’s long process is now literally five minutes,” Reed said. “There are hundreds of use cases. Our approach has been just to get it in the hands of our people so that they can start to work out where they can make changes to the way they work.”

It’s the latest sign of how the world’s biggest financial firms are experimenting with the ways they can use artificial intelligence. PricewaterhouseCoopers LLP said earlier this month that it had teamed up with ChatGPT owner OpenAI to offer clients advice generated by the technology, while KPMG announced a $2-billion investment in Microsoft Corp’s generative AI effort and cloud services.

For Moody’s, the software will leave an electronic trail showing where the information its citing came from, according to Philip Moyer, vice-president of artificial intelligence and business solutions within the cloud business at Alphabet Inc.’s Google. He said that should also prevent the technology from hallucinating, a term coined by developers for when large language models generate answers that sound convincing but aren’t true.  

Outside Offering

Moody’s will also offer the large language models it’s developing to other financial firms, which can also use it to analyse information for rote tasks. 

For instance, Moyer said, the software could help a bank employee quickly assess whether the bank should onboard a small business client. That’s because the staffer could instruct the large-language model to find the top three risks the potential client identified in its financial disclosures or write three bullet points summarising its most recent earnings call or identify five peers with similar carbon footprints. 

Within minutes, the bank employee could make a judgment call based on research that would have previously taken hours. 

“This is about democratising access to that information,” Moyer said. “It’s about building a model that allows people that don’t know and understand that language to be able to interact with that information.”

More stories like this are available on