How can AI impact the future of the finance industry?

Key Concepts

• Agentic AI: Generative AI combined with the ability to take action.
• Generative AI: "Created for me" technology.
• Agentic AI: "Do it for me" technology.
• Traditional AI/Machine Learning: Works well on structured data.
• Unstructured Data: The prevalent data type in financial services.

Agentic AI: Definition and Core Functionality

Agentic AI represents an evolution of Artificial Intelligence, specifically building upon the capabilities of Generative AI. The core distinction lies in its ability to not only generate content or insights but also to take action based on those outputs. In essence, Agentic AI is characterized by three key components:

1. Goal Setting: It can understand and define a specific objective.
2. Task Planning: It can devise a sequence of steps to achieve that goal.
3. Action Execution: It can perform the planned tasks.
4. Feedback Adaptation: It can adjust its actions based on incoming feedback or changing circumstances.

This "do it for me" paradigm differentiates it from Generative AI, which is described as "created for me" technology.

The Role of Generative AI in Agentic AI

The emergence of Agentic AI is directly enabled by advancements in Generative AI. Generative AI's ability to understand and process complex, often unstructured, information is a prerequisite for Agentic AI's planning and execution capabilities.

Traditional AI vs. Generative and Agentic AI in Financial Services

• Traditional AI/Machine Learning: This branch of AI, existing since at least 1956, excels at processing structured data. Structured data is organized and easily quantifiable, such as numerical databases or predefined forms.
• Generative AI and Agentic AI: These newer forms of AI are particularly relevant to financial services because a significant portion of data in this sector is unstructured. Unstructured data includes text documents, emails, audio recordings, and images, which are more challenging for traditional AI to interpret and act upon. The ability of Generative AI to comprehend and process this unstructured data is what allows Agentic AI to function effectively in such environments.

Real-World Application: Financial Services - Payment Resolution

A prominent example of Agentic AI in action is within the financial services sector, specifically in the area of payment resolution.

Case Study: Payment Error Resolution

An Agentic AI system can autonomously handle a payment error by:

1. Initiating an Autonomous Workflow: The AI begins a predefined process to address the issue.
2. Processing Unstructured Data: It can sift through various sources of information, including emails, transaction logs, and customer communications, which are often unstructured.
3. Cross-Checking Information: The AI compares data from different sources to identify discrepancies and confirm the nature of the error.
4. Confirming the Mistake: It pinpoints the exact cause and location of the payment error.
5. Amending the Payment: The AI can make the necessary corrections to the payment details.
6. Submitting for Approval: It can then submit the amended payment for human review and approval, if required by policy.
7. Logging in Audit Trail: Finally, it meticulously records all actions taken in the audit trail for compliance and record-keeping purposes.

This example highlights how Agentic AI can automate complex, multi-step processes that previously required significant human intervention.

Report on Agentic AI in Finance

The significance of Agentic AI for financial services has led to the publication of a comprehensive report titled "Agentic AI finance and the do it for me economy." This report is publicly available and delves deeper into the implications and applications of this technology in the financial sector.

Conclusion

Agentic AI represents a significant leap forward in AI capabilities, moving beyond content creation to active problem-solving and task execution. Its ability to process unstructured data, set goals, plan, act, and adapt makes it particularly valuable for industries like financial services, where complex workflows and vast amounts of unstructured information are common. The "do it for me" nature of Agentic AI promises to automate and streamline processes, driving efficiency and innovation.