AgentHack - Support

AgentHack submission type

Enterprise Agents

Name

Daniela Rosenstein

Team name

Bilie Team

How many agents do you use

Multiple agents

Industry category in which use case would best fit in (Select up to 2 industries)

Customer service

Complexity level

Advanced

Summary (abstract)

SupportAgent, an intelligent automation solution that combines UiPath orchestration, LangGraph agent management, and AWS Bedrock AI capabilities

Detailed problem statement

Slow support response times
Inconsistent answer quality
Limited support hours (business hours only)
Inability to scale with growing ticket volume
Manual data gathering from multiple systems

Detailed solution

SupportAgent creates an intelligent support automation ecosystem with clearly defined component responsibilities:

UiPath - Primary Orchestration Layer
Serves as the main orchestration platform for the entire process

Triggers and manages process execution based on business rules

Handles all enterprise system integrations:

Zendesk for ticket management

CATO MCP for network data retrieval

Jira for ticket/knowledge base searches

CMA integration for additional client data

AWS services for media processing

Manages error handling, retries, and exception workflows

Provides monitoring and reporting through Orchestrator

LangGraph - Agent Orchestration Platform
Manages intelligent agent workflows and decision routing

Orchestrates specialized agents:

Request Analysis Agent

CATO Specialist Agent

Jira Specialist Agent

Response Generation Agent

Research Agent

Maintains conversation state and context

Implements complex decision trees and routing logic

Enables agent collaboration for complex issues

AWS Bedrock - AI/LLM Layer
Provides Claude AI models for intelligent processing:

Claude 3 Haiku for fast analysis and routing

Claude 3 Sonnet for comprehensive response generation

Powers natural language understanding and generation

Enables context-aware decision making

Ensures response quality and relevance

Transcribe: Audio-to-text conversion

Extract: Document text extraction

Recognition: Image text detection

Demo Video

Expected impact of this automation

Customer Impact: Slow responses despite some automation, limited hours coverage, inconsistent solutions

Operational Impact: High cost per ticket, difficult to scale, engineer burnout from repetitive tasks

Strategic Impact: Limits growth, competitive disadvantage, poor customer experience despite automation attempts

Automation Gap: Current automated tools (KB Chatbot, auto-messages) help but can’t access real-time data or provide comprehensive solutions

UiPath products used (select up to 4 items)

UiPath Robots

Automation Applications

ZenDesk,Slack,Jira

Integration with external technologies

Bedrock,python langgraph

Sample inputs and outputs for solution execution

zendesk comments, jira tickets, kb - outputs a final comment as response

Other resources

n/a