Submission type
Generic use case
Name
Rajneesh Khare
Industry category in which use case would best fit in (Select up to 2 industries)
Information technology and services
Operations
Complexity level
Intermediate
Summary (abstract)
Admin Assistant is an intelligent, conversational chatbot built to bridge the gap between non-technical users and enterprise automation tools. Using a combination of Google Gemini, UiPath Agent Builder, and UiPath Maestro, the solution allows users to interact via natural language queries. If the assistant detects that the query is related to UiPath Orchestrator (e.g., getting job statuses or failed queue items), it automatically triggers an agentic process that navigates to the appropriate tool and returns contextual results to the user — all without writing a single line of code or accessing the Orchestrator manually.
Detailed problem statement
Enterprise users frequently need to query automation-related data or trigger automation tasks (such as retrieving failed queue items, checking job statuses, or restarting processes). However, these actions typically require them to navigate through technical platforms like UiPath Orchestrator or request help from RPA developers.
This results in:
→ Operational delays
→ Dependency on IT or automation teams
→ Limited access for non-technical stakeholders
There is a growing need for a self-service automation interface that’s as intuitive as having a conversation.
Detailed solution
The Admin Assistant is a conversational prototype that enables business users to query or interact with their automations using a natural, chat-based UI. The core of the system lies in its ability to intelligently decide when a user’s request requires an automation and then dynamically invoke the right process.
How It Works:
-
User asks a question via a chatbot built using Streamlit (e.g., “Show me failed queue items from Orchestrator”).
-
The LLM (Google Gemini) analyzes the intent. If it matches a supported Orchestrator operation, the assistant:
→ Triggers an agentic process via API.
-
The agentic process is designed in UiPath Maestro, containing:
→ An agent created using UiPath Agent Builder that interprets the user’s intent.
→ Based on the intent, the agent navigates to the right tool (e.g., queue lookup, job status).
-
The agent collects the response and returns the data back to the LLM.
-
The LLM formats and summarizes the output for the user in conversational style.
Tech Stack:
→ Development Language: Python
→ UI Framework: Streamlit (used to build the user-facing chatbot interface)
→ LLM Integration: Google Gemini 1.5 Flash (via google.generativeai Python SDK)
→ UiPath Agent Builder: To develop the tool-identifying agent
→ UiPath Maestro: To orchestrate agentic workflows that execute based on user intent
→ APIs: UiPath Orchestrator APIs (used to start jobs and fetch execution data)
Video link
Expected impact of this automation
→ Empowers business users to get Orchestrator insights without technical training
→ Speeds up operational response by reducing dependency on RPA developers
→ Enables natural language interaction with enterprise automation
→ Reduces time-to-action for common admin tasks
→ Promotes adoption of agentic automation within organizations
UiPath products used (select up to 4 items)
UiPath Orchestrator
UiPath Agent Builder
UiPath Maestro
Integration with external technologies
Google Gemini, Streamlit (Python-based UI framework), REST APIs for communication between frontend and Orchestrator
TO-BE workflow/architecture diagram (file size up to 4 MB)
Other resources
N/A