Integrated risk assessment company Moody’s Analytics is deploying agentic AI tools and helping bank clients adopt the technology. 

The agentic AI market is expected to swell to $196.6 billion by 2034 with a compound annual growth rate of 43.8%, according to a March report from data analytics company Market.us.

“Most of the banks have started already moving toward agentic solutions,” Dimitrios Papanastasiou, global head of gen AI solutions at Moody’s Analytics, told Bank Automation News. “We are also deploying an agentic AI framework to automate key banking workflows that rely on data gathering, consolidation of information, analysis and content generation.” 

Moody’s aims to deploy agentic AI for: 

• Writing credit memos; 

• Risk monitoring; 

• Sales and marketing; and 

• Compliance and KYC. 

For example, when analysts write a credit memo, they can define in the agentic AI tool what the report will look like, what sections it will have and where to get the information to insert in the report, Papanastasiou said.  

“Every report or workflow that follows a predetermined process that you can think about automating can be done through the agentic AI solution,” he said. 

The platform will have access to Moody’s vast data estate and will utilize numerous specialized agents, each one combining domain expertise with specific instructions on how to interpret data and generate content, Papanastasiou said.  

The platform was developed by Moody’s, leveraging a range of large language models, such as those developed by OpenAI, Anthropic and Google, he added 

While still in its pilot phase, the agentic AI tool can yield more than 50% efficiency gains depending on the task and use case, Papanastasiou said, adding that the tool gives much higher efficiency gains for complex workflows that rely on the consolidation and analysis of large amounts of information like credit report generation. 

Deploying agentic AI solutions rather than general AI models is more effective for Moody’s because agentic tools are experts on the task at hand, have the capability for better re-enforced learning and have guardrails to avoid hallucinations, he said. 

Agentic AI point solutions 

As agentic AI gains market share, agentic AI point solutions are likely to be the first to hit the market, David Brudenell, executive director at AI operating system Decidr, told BAN. 

“What that means is that they’re taking either a single task or … a small workflow, and putting agents in there,” Brudenell said. 

Other fintechs are also deploying agentic AI within financial services. These agentic AI point solutions hit the market in May: 

• Temenos’ FCM AI Agent, a compliance tool that reduces false positives, allowing more payments to process without interruption.  

• Fenergo’s FinCrime Operating System, which uses agentic AI to increase compliance while reducing operational costs. The six AI agents launched under FinCrime OS are data sourcing, screening, documenting, significance of data, auto-completion and insights. 

• Salesforce’s Agentforce for financial services, which provides templates that can be used as financial adviser and banker agents, banking and insurance service agents and digital loan officer agents.