Automated Audit Agent

Agent Description:

The Automated Audit Agent streamlines internal compliance checks by fetching records through a GET API. It validates the data and automatically evaluates it against predefined rules, such as compliance scores or contract dates. The agent identifies and flags non-compliant items (e.g., low scores, missing documents). Finally, it generates a comprehensive audit report with summaries and narrative findings for review.

Purpose and Components

Purpose: To automate the internal audit process by fetching compliance records from a data source, evaluating them against specific rules (e.g., compliance scores, contract dates), and generating a comprehensive report that flags non-compliant items and provides a summary for review.
Components:
- Data Retriever: An agent to fetch and validate the raw compliance dataset.
- Report Generator: An agent to analyze the validated data, apply audit logic, and create a human-readable report.
- API Connector (GET): A tool to retrieve the compliance data from a remote endpoint.

Supported Capabilities

Initiating a secure GET request to retrieve a compliance dataset.
Verifying the presence and integrity of required data fields (e.g., record_id, department, policy, amount, approval_status).
Parsing the dataset to extract key compliance metrics (e.g., Vendor_ID, Contract_ID, Compliance_Score, Risk_Level).
Automatically identifying and flagging vendors that fail compliance checks, such as:
- Compliance_Score < 80
- Expired or missing Contract_Dates
- Unverified or missing documentation
Summarizing key audit outcomes (Total Vendors Reviewed, Percentage Compliant vs. Non-Compliant, Top 3 Risk Categories).
Generating a human-readable narrative summary of findings (e.g., "Vendor ABC shows partial compliance due to missing KYC records...").
Returning a final, structured audit report.

LLM Used

GPT_4O_MINI

Note: To learn more about the LLM and to modify its behavior, refer to the Configuring LLM settings section.

Sub-Agents

1. Data Retriever

Role: Data Collector
Scope: Fetches all compliance records via secure GET API calls, validates the fields, and forwards structured data for evaluation.
Description: This agent starts the audit. It uses the GET tool to fetch the entire compliance dataset. Its primary job is to verify that all required fields are present and correctly formatted, ensuring data integrity before passing the cleaned dataset to the Report Generator for analysis.
LLM Used: Default (Inherits from parent agent).
Tools Used: Request - Get

2. Report Generator

Role: Data Presenter
Scope: Compiles flagged and approved records into structured reports, generates summaries, and outputs for dashboards or internal review.
Description: This agent receives the cleaned data from the Retriever. It parses the data to apply specific audit rules, such as checking compliance scores, contract dates, and documentation status. It then compiles all findings into a final report, including a statistical summary and a human-readable narrative, before outputting the result.
LLM Used: Default (Inherits from parent agent).
Tools Used: None

Tools Used:

Request - Get Tool: Fetches the compliance dataset from a remote JSON endpoint.

Note: For details on modifying the Tools, refer Tools Library section.

Agent Workflow Behavior Summary

The Data Retriever (start node) is triggered and uses the Request - Get tool to fetch the compliance dataset.
It verifies the dataset for completeness and data integrity (e.g., all required fields are present).
The cleaned and validated dataset is then passed to the Report Generator.
The Report Generator (end node) analyzes the data, flags non-compliant vendors based on business rules (like scores or dates), and compiles a final audit report. This report includes both a statistical summary and a narrative explanation of the findings.

Sample Questions:

Provide summary of compliance review for vendor contracts
Which departments have pending audit actions?