Project: [Enteprise] - [LISA Lincolnshire Information Services Agent]

Track

Enterprise Agents (M365 Agents Toolkit)

Project Name

Lincolnshire Info Services Agent

GitHub Username

@JonEricEubanks

Repository URL

https://github.com/JonEricEubanks/LISA-Lincolnshire-Info-Services-Agent

Project Description

🏛️ LISA — Lincolnshire Info Services Agent

LISA is a production AI agent built on Microsoft Copilot Studio, deployed
inside Microsoft Teams via M365 Copilot Chat. It provides Village of Lincolnshire
staff and residents with instant access to verified municipal data through a
conversational interface.

🔍 Problem

Municipal data is scattered across GIS portals, county databases, and paper permit
files. Staff and residents waste time making calls or navigating multiple systems
for basic property information.

✅ Solution

LISA consolidates property ownership, assessments, zoning, school districts,
utility providers, elected officials, and permit history into a single 30-second
chat interaction — right inside Teams.

⚡ Key Features

• 🗄️ Azure AI Search over live municipal SQL views
• 📄 SharePoint permit documents with full citation transparency
• 🃏 Adaptive Card permit lookup rendered natively in Teams
• 🤝 Connected multi-agent architecture with specialized child agent
• 🔁 Automated email triggers for property requests
• 🧪 Formal semantic evaluation suite with Compare Meaning scoring

Demo Video or Screenshots

🎥 Demo Video & Screenshots

• 📹 Demo Video: [Demo]

Primary Programming Language

Other

Key Technologies Used

🛠️ Key Technologies

• Microsoft Copilot Studio
• Azure AI Search (Semantic + Hybrid)
• Microsoft Teams / M365 Copilot Chat
• SharePoint (Permit Document Library)
• Adaptive Cards
• Power Automate (Email Triggers)
• Claude Sonnet 4.5 (Preview)
• Azure SQL Views
• Connected Agents (Multi-Agent Architecture)

Submission Type

Individual

Team Members

n/a

Submission Requirements

• My project meets the track-specific challenge requirements
• My repository includes a comprehensive README.md with setup instructions
• My code does not contain hardcoded API keys or secrets
• I have included demo materials (video or screenshots)
• My project is my own work with proper attribution for any third-party code
• I agree to the Code of Conduct
• I have read and agree to the Disclaimer
• My submission does NOT contain any confidential, proprietary, or sensitive information
• I confirm I have the rights to submit this content and grant the necessary licenses

Quick Setup Summary

🚀 Quick Setup

1. Access Microsoft Copilot Studio at copilotstudio.microsoft.com
2. Import the agent solution package from the repository
3. Connect Azure AI Search indexes and SharePoint knowledge sources
4. Configure the Semantic Hybrid Search connector
5. Publish to Microsoft Teams via M365 Copilot Chat

Technical Highlights

🏆 Technical Highlights

• Connected multi-agent architecture — LISA orchestrates Agent 1 as a
  child agent with redirect trigger logic, keeping the parent clean and scalable
• Three grounded knowledge sources — two Azure AI Search indexes and a
  SharePoint permit document library with full citation transparency
• Adaptive Card permit lookup — interactive form rendered natively in Teams
• Instruction-driven behavioral contract — enforced rules ensure consistent,
  trustworthy responses across every interaction
• Formal evaluation suite — 7 semantic test cases with Compare Meaning
  scoring for regression testing after any configuration change

Challenges & Learnings

🧠 Challenges & Learnings

The biggest challenge was grounding responses reliably across municipal data
with inconsistent completeness. The solution was enforcing Not available as
a hard instruction rather than letting the model skip or infer missing fields —
building trust with users who needed to know the difference between missing and
unretrieved data.

Contact Information

https://www.linkedin.com/in/joneric-eubanks-pmp-developer/

Country/Region

United States

This issue is a project submission for the Lincolnshire Info Services Agent and does not identify a change, source file, test, or entry point in the agentsleague repository. Review the linked project repository and its README first; the submission provides no defined contribution scope or completion criteria.