I admit that I am impressed with the speed with which certain companies like Moody’s have launched external-facing services for their customers that are leveraging generative AI. These are companies that are sensitive to data sovereignty, regulatory complexity, and reputational risks.
It is becoming clear that the large companies in financial services with substantial invaluable proprietary data (be it in the form of text or numbers — think financial statements, information on private and public companies, or analytics, etc.) that were advanced in their cloud migration data roadmap, are able to start leveraging generative AI and Large Language Model capabilities. The advancements from these technologies are exposing the Cloud and data laggards in our industry.
The risks in our industry are becoming more complex and are increasingly forcing digitalization. We all know that legacy systems are costly to maintain and full of obstacles in scaling products and services. Leveraging unstructured and alternative data in these environments is costly, risky with respect to cyber security and privacy, and time-consuming.
Microsoft is empowering financial services to deal with these risks and the complexity both at the Compute and at the Data & Analytics level. Microsoft Azure has empowered large financial players (e.g. Société Générale, Rabobank, MUFG, etc.) to reduce compute costs for running complex models using multiple data sources and producing AI analytics. Doing more at a lower cost while at the same time reducing risks (cyber risks, Fraud, privacy) and going to market fast with high compute services; is paramount especially when considering the deployment of these advanced Machine Learning models. Microsoft is an enterprise partner in driving innovation to meet the growing demands of the industry to innovate in Financial services with GenAI.
In late June, Bloomberg announced the Moody’s and Microsoft strategic partnership to co-create new products and services for research and risk assessment built on Azure. I spoke to Nick Reed who has been the Chief Product Officer at Moody’s, who leads the company’s GenAI Intelligence group.
Moody’s Analytics, the division of Moody’s focused on non-rating activities, has been growing in the mid double-digits is contributing 50% of the company`s revenues, which shows the increasing demand for insights for risk management and decision-making purposes.
Moody`s has been one of the early clients of OpenAI, as they wanted to explore the potential of GenAI in their internal processes. They started exploring the potential for their ratings business which has already gained productivity efficiencies in their internal workflows required in credit ratings. Early on, Moody`s realized the potential of the technology and wanted to develop a secure and compliant way to implement these LLMs internally across their businesses and explore the commercialization potential.
Moody`s partnered with Microsoft to develop fast and securely a Moody`s platform deploying Azure OpenAI 4.0 in combination with Moody`s proprietary data and analytics. This new Moody`s platform can onboard any Large Language model (e.g. Llama, Bard, etc.) and can be fine tuned for different purposes.
Currently, it is used internally by all Moody`s business units. The first customer-facing commercial service is the `Moody`s Research Assistant`.
Moody`s is focused on what they call The era of Exponential Risk which in plain words means a combined view of cyber-attacks; geopolitical tensions; sanctions and security issues; extreme weather events, persistent inflation, Supply chain failures, and more.
The `Moody`s Research Assistant` is intended to provide a 360-degree perspective on the various risks covered by Moody`s Data & Insights.
Moody`s customers can ask the `Moody`s Research Assistant` a simple question like:`Give me an overview of the financial standing and credit rating for HLP Manufacturing`
A more detailed fact-based answer on Credit, Cyber, Climate, ESG, Catastrophe, and Property risks can be provided with the relevant prompt if the customer has access to the relevant Moody`s libraries. Keep in mind that Moody`s has one of the largest databases on private companies (the Moody’s Orbis database has information on 445+ million entities worldwide). Imagine the empowerment of financial analysts which goes beyond the speed of obtaining this analysis and is more about the value-add of contextualizing risks by asking the assistant complex questions.
In the short time that the `Moody`s Research Assistant` has been available to customers in private preview, there is a noticeable new user base. In large banks, credit officers and portfolio managers are using the Moody`s co-pilot in their workflows. They are becoming better at decision-making because of the easiness of contextualizing the information typically used in their workflows.
Managing the complexity of risks inherent in deploying Large Language Models, is daunting. Experimenting internally and launching customer-facing services leveraging GenAI, requires a tech partner who can mitigate these risks and assist in going to market fast, securely, and in a cost-efficient way.
Moody`s wants to make sure that its own data is protected, and that the data of the client queries is protected. Moody`s wants to manage the costs of running these LLMs. It wants to use existing LLMs (like OpenAI but not only) and incorporate its own proprietary knowledge library. It wants to build all this in a way that new services can be launched based on the feedback from this first service offering.
The agility of the Moody`s platform requires increased risk management, and this is the core value of building on Microsoft`s Compute and AI Data & Analytics infrastructure.
This is a sponsored post. Microsoft is my client. I choose to highlight topics that I believe are valuable in understanding how technology is reshaping Financial Services.
📌 Subscribe to my YouTube Channel with my insights and industry leader interviews. New video every Wednesday: https://www.youtube.com/EfiPylarinou
📌 Apple Podcast 👉 https://buff.ly/3P1Cp1Z
📌 Spotify Podcast 👉 https://buff.ly/3xPyWaV
📌 Twitter: https://twitter.com/efipm
📌 Linkedin: https://www.linkedin.com/in/efipylarinou/
📌 TikTok: https://www.tiktok.com/@efiglobal
📌 Web: https://efipylarinou.com/