Build an Automated SLA Risk Agent with Routines in Microsoft Foundry

2026-06-17 · Source: Microsoft Foundry Blog articles · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Cloud Computing & IT Infrastructure · Depth: Intermediate, medium

Summary

Microsoft Foundry offers a tutorial for building an automated SLA risk agent that provides daily AI-generated summaries of support tickets likely to breach Service Level Agreements. The process, taking about 45 minutes, involves connecting a sample dataset from Hugging Face, stored in Azure Blob Storage, to Azure AI Search for Retrieval Augmented Generation (RAG). This data forms a Foundry IQ knowledge base, which grounds an agent configured with specific SLA triage instructions. The agent identifies high-risk tickets based on criteria like "Critical or High priority," "Open or In progress status," "Low hours_until_breach values," and "Clear business impact." Finally, Microsoft Foundry's Routines feature schedules the agent to run daily at 9:00 AM, ensuring early risk identification and recommended actions.

Key takeaway

For MLOps Engineers or AI Engineers tasked with automating support operations, this Microsoft Foundry pattern offers a robust solution. You can deploy a production-ready SLA risk agent, grounded on your actual ticket data, to proactively identify potential breaches. This approach ensures your support team receives daily, AI-generated summaries of high-risk tickets, enabling timely intervention. Consider adapting this routine-based agent framework for other recurring analysis tasks like inventory alerts or compliance checks to enhance operational efficiency.

Key insights

An automated agent in Microsoft Foundry can proactively identify SLA breach risks using RAG and scheduled routines.

Principles

• Grounding agents on specific data improves accuracy.
• Automated routines ensure timely risk detection.
• Prioritize tickets by urgency and business impact.

Method

Prepare data in Blob Storage, index with Azure AI Search, connect to Foundry IQ knowledge base, build a grounded agent with SLA triage instructions, and schedule daily runs via Routines.

In practice

• Integrate Teams for notifications.
• Connect multiple data sources.
• Track SLA trends with dashboards.

Topics

• Microsoft Foundry
• SLA Management
• AI Agents
• Azure AI Search
• Retrieval-Augmented Generation
• Automated Workflows

Best for: AI Engineer, MLOps Engineer, Software Engineer

Related on AIssential

• How Veris AI and Lume Security built a self-improving AI agent with Microsoft Foundry — High-fidelity simulation and automated optimization enable AI agents to self-improve safely in production.
• SAS: AI Agents for Industry-Specific Challenges — Domain-tuned AI accelerators bridge the gap between general AI and specific industry challenges, driving measurable results.
• Tracking Every Token: Granular Cost and Usage Metrics for Microsoft Foundry Agents — Granular AI agent cost and usage tracking is achievable by integrating Azure API Management and Application Insights.
• AI Agent Sandboxing for SaaS: How Builders Let Agents Work Without Letting Them Roam — AI agent sandboxing enables safe, controlled automation by enforcing strict boundaries around agent actions and resources.
• Microsoft's AI Futurist explains how he uses Copilot — and the real-world problems enterprises are solving with agents — Microsoft's agent strategy centers on providing a comprehensive platform for context, governance, identity, and secure data access, alongside diverse model choices.
• AI tool poisoning exposes a major flaw in enterprise agent security — AI agent tool registries require behavioral integrity checks, not just artifact integrity, to counter sophisticated supply chain attacks.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Microsoft Foundry Blog articles.