EN

How AI Agents and Decision Agents Combine Rules & ML in Automation

#AI Tools
2025-11-03 19:3811 min read

Content Introduction
Ask Questions

The video discusses the concept of agentic AI, which integrates large language models with other automation technologies to create more adaptable and transparent systems for solving complex problems, particularly in banking. It illustrates this through a loan application process, showcasing how a chat agent uses large language models for customer interaction. The framework involves orchestration agents to control processes, document ingestion agents to extract data from various sources, and decision agents to ensure consistent eligibility assessments. By combining these agents effectively, the bank can streamline customer queries and decisions, maintaining regulatory compliance while ensuring user-friendly experiences. The process culminates in resolving potential loan issues and making informed lending decisions efficiently.

Key Information

  • Agentic AI integrates large language models with other automation technologies to create adaptable and transparent systems.
  • Multi-method agentic AI is necessary for solving complex problems, such as lending money by banks.
  • Chat agents powered by large language models can interact with customers to understand their queries about loan policies.
  • Orchestration agents manage requests and facilitate communication between different agents in the AI framework.
  • Decision agents determine customer eligibility and loan origination decisions based on consistent logic and documented policies.
  • Workflow technologies keep track of the loan application process, enabling state management and customer interactions.
  • Document ingestion agents convert documents into processable data for decision making.
  • Companion agents assist customer service professionals by providing necessary information about loan applications.
  • Explainer agents help interpret decision-making processes into understandable language for customers and service agents.

Timeline Analysis

Content Keywords

Agentic AI

Agentic AI utilizes large language models and combines them with other automation technologies to create adaptable and transparent systems. This multi-method approach aids in building solutions that can withstand regulatory scrutiny.

Multi-method Agentic AI

This approach integrates large language models with proven automation to address complex problems like decision-making processes in institutions such as banks.

Chat Agent

A system that uses large language models to interact with customers dynamically, understanding queries and providing responses, particularly relevant for services like bank loan inquiries.

Orchestration Agent

This agent manages requests within the agentic framework, leveraging large language models to search for appropriate agents to handle specific customer inquiries.

Loan Policy Agent

An agent that interprets and delivers information about a bank's loan policies using retrieval augmented generation (RAG) to provide intelligible answers based on a range of documents.

Workflow Technology

This technology tracks customer interaction states and manages processes such as loan applications, ensuring that customer needs are accurately addressed over multiple sessions.

Decision Agent

Utilized for consistently evaluating loan eligibility based on predefined business rules, ensuring decisions are made transparently and can be explained to regulators.

Ingestion Agent

A specialized component that processes documents related to loan applications, extracting necessary information to aid in decision-making effectively.

Companion Agent

Serves to assist customer service representatives by providing quick access to customer information and application status during loan processing.

Explainer Agent

This agent translates the decision-making logs of other agents into intuitive explanations for customer service representatives, enabling them to communicate effectively with customers regarding loan decisions.

Related questions&answers

What is agentic AI?

Agentic AI is an approach that combines large language models with other proven automation technologies to build more adaptable and transparent systems.

Why are large language models not sufficient alone?

While large language models are powerful tools, they have limitations and constraints. They need to be combined with other technologies to handle complex problems effectively.

What is multi-method agentic AI?

Multi-method agentic AI is a framework that combines large language models with other automation technologies to create robust solutions that can endure regulatory scrutiny.

How does a bank use agentic AI to lend money?

A bank can use agentic AI to interact with customers through chat agents, orchestration agents, decision agents, and data agents to handle loan applications and queries efficiently.

What role does a chat agent play in this process?

The chat agent interacts with customers to understand their requests or questions regarding loans, converting their inputs into actionable data for the bank's system.

What is an orchestration agent?

An orchestration agent identifies and utilizes various agents needed to fulfill customer requests, such as finding a suitable agent that can handle loan policy inquiries.

How does the loan application process manage state?

The loan application process uses workflow technology that maintains the state of each application, allowing customers to resume where they left off even after interruptions.

What is RAG in the context of AI?

RAG stands for Retrieval-Augmented Generation, a technique used to enhance AI responses by retrieving relevant information from a large set of documents.

What is the purpose of a decision agent?

A decision agent evaluates the eligibility of applicants by applying business rules consistently and transparently to ensure fairness and regulatory compliance.

How do ingestion agents work?

Ingestion agents process and transform documents into structured data, enabling the system to use that information in making loan decisions.

What happens when a loan application requires further clarification?

If a loan application is uncertain, a customer service representative is alerted, and they may work with a companion agent and an explainer agent to gather additional context and resolve issues.

More video recommendations

Share to: