Montgomery County Monty 2.0 Azure OpenAI constituent chatbot

L2 · Supervised Verified · Production Municipal / Local Government Operations

Montgomery County MD launched Monty 2.0 (Azure OpenAI + Zammo.ai RAG) covering 3,000 topics in 140 languages, facilitating 20,000+ conversations and cutting unanswered queries to 10–15% (reported by Microsoft / Montgomery County)

Problem

Prior Monty bot covered only ~20 topics; constituents needed scalable multilingual answers without hallucinations from the open web (reported by Microsoft / Montgomery County)

Workflow

Constituent asks Monty 2.0 on county channels
Azure AI Search RAG retrieves only county knowledge base content
Zammo.ai front end handles multi-turn chat and translation (140 languages)
GIS-aware answers for trash pickup/voting precinct; feedback loop improves prompts

Human-in-the-loop: 311/staff handle unresolved or sensitive issues; IT continuously tunes prompts from analytics.

Tools & stack

Tools

Azure OpenAI Service
Azure AI Search
Zammo.ai

Integrations

County knowledge base
GIS address personalization

Frameworks

None listed

Prerequisites

Curated knowledge base
Hallucination reduction beta
Translation UX

Risks & limitations

Risks

Stale knowledge
Low CSAT if expectations high
GIS personalization errors

Limitations

50% CSAT is modest; Microsoft customer story framing

Outcomes & evidence

Reported outcomes

20,000+ constituent conversations; 3,000 topics; 140 languages
Reported by Microsoft / Montgomery County · Source (microsoft.com)
Unanswered queries reduced from 35–45% to 10–15%; ~50% CSAT
Reported by Microsoft / Montgomery County · Source (microsoft.com)

Evidence

vendor-case-study (named customer) (2025-03-26)
Monty 2.0 has facilitated 20,000 constituent conversations

Vendors & implementers

Review history

Last reviewed 2026-10-14 by Research Chief (agent).

2026-10-14: Added in batch-05; source fetched and key figures confirmed; figures attributed.