AI-Powered ArcGIS Monitor Implementation for the City of Elizabethtown, KY

Elizabethtown, KY

The City of Elizabethtown, KY (the city) runs its GIS operations on ArcGIS Enterprise. Bowman delivered a modern ArcGIS Monitor installation, automated plain-language reporting, and an AI agent staff can query about their own system, so they get the signal without combing through dashboards.

Challenge

The City of Elizabethtown initially came to Bowman for a free GIS assessment of its ArcGIS Enterprise environment. That assessment recommended a Monitor upgrade, among other improvements. The city returned to Bowman to implement that recommendation. They wanted to catch platform issues before end users hit them, without asking its GIS team to keep an eye on the system around the clock. Its existing monitoring ran on ArcGIS Monitor 10.8.1. Any new system needed to surface problems in plain language, not add another set of dashboards for staff to watch.

Solution

To address this challenge, Bowman implemented ArcGIS Monitor and built automated reporting and an AI analysis layer on top of it, including:

  • Installation and configuration of the current ArcGIS Monitor release (2026.1) on the city’s dedicated server, connected to ArcGIS Enterprise 11.5, replacing their 10.8.1 installation.
  • Analysis views for service and instance usage and system health across the enterprise components, including database servers, with alerts for service-down, disk-pressure and data-store conditions.
  • A deterministic Python report that pulls directly from ArcGIS Monitor and emails a weekly system health summary to GIS staff, with no AI in the path.
  • An AI agent, built in Copilot Studio and connected through Power Automate, that lets GIS staff ask natural-language questions about their Monitor data and get answers directly in Microsoft Teams.
  • A connection to Esri’s ArcGIS MCP, giving the agent structured access to Monitor’s operational data.
  • A live training session and a follow-up refinement session that tuned the analysis views and weekly report against how staff were actually using them, plus configuration documentation hand-off.

The AI reads only ArcGIS Monitor’s operational metrics, such as service usage, error rates and system health, not the city’s underlying GIS data. The approach is built to run on any capable AI assistant; for Elizabethtown it was delivered within the city’s Microsoft 365 environment, using Power Automate and Copilot Studio, connected to the on-premises Monitor server through a gateway.

Impact

The City of Elizabethtown now has proactive, low-effort observability of its ArcGIS Enterprise platform. Issues are caught before end users report them. Staff get plain-language summaries instead of raw system data. Because the agent runs in Microsoft Teams, GIS staff can query the system when something needs a closer look, from a desktop or a phone, on site or at home. The MCP connection also lets the team diagnose and correct long-standing ArcGIS Server issues, cutting down log noise and reducing the number of alerts Monitor raises. With ArcGIS Monitor on the current supported release and the training and documentation to back it up, the city’s GIS team can manage and extend the system on its own going forward.