AgentHack submission type - Type 1 - Enterprise Agents
Name - Mahalakshmi S
Project Name - Compliance check Agentic Automation Usecase
Team details - DevPost Team
Complexity level - High
Summary - The process involves analyzing and resolving all alerts related to Sanctions, Group Blacklist, Internal Blacklist within 10 days. Based on the analysis, a decision is made to link, whitelist or close each alert.
FTE benefits - 131
UiPath products used -
UiPath Apps
UiPath Studio
UiPath Agents
UiPath Action Center
Solution and Code - https://drive.google.com/drive/folders/112CPuxlPibEHDTlyPy6PvOdBJ2CAnLQR
Agentic Solution Architecture -
The Manual Process: Compliance team manually vets individuals by cross-referencing names and dates of birth (DOB) across search engines, filtering out irrelevant social media clutter, and reading through the top 10 external news or legal sites to assess adverse media risk.
The Core Bottleneck: This process requires heavy cognitive lifting. Traditional rules-based RPA fails here because it cannot accurately identify what a “social media site” is across infinite variations, nor can a standard bot “read” a complex news article to determine if a mention is a criminal match or just a shared name.
The Agentic Solution: An autonomous compliance agent that mimics human reasoning. It filters out digital noise (social media platforms), extracts core contextual web data, and uses Semantic Reasoning to determine if a person poses a genuine financial crime risk—delivering an audit-ready background decision in seconds.