January 28, 2025 · 10 minutes read · Tutorial

How to Use Autonomous Agents in SnowCoder: A Complete Guide

By Kumoco

Editor's note (2026-06): SnowCoder now ships 8 MSP Agents (6 scheduled + 2 on-demand). This guide originally covered six; the six scheduled agents are Instance Health Monitor, SLA Guardian, Incident Manager, License Optimizer, Patch Manager, and AI Readiness Advisor, and the two on-demand agents are the Instance Audit Agent and Upgrade Readiness Agent. The newer agents are summarised in the section below and have dedicated posts at Upgrade Readiness Agent and Zurich to Australia: Upgrade Readiness Agent in action. MSP Agents are an Enterprise tier capability.

What If Your ServiceNow Instances Could Monitor Themselves?

Managing ServiceNow instances is a constant balancing act. You're tracking SLA compliance, watching for performance issues, optimizing licenses, planning upgrades, and triaging incidents, often across multiple instances simultaneously.

What if AI could handle the monitoring and analysis for you, 24/7, and alert you only when action is needed?

That's exactly what SnowCoder's autonomous agents do.

What Are SnowCoder Agents?

Agents are AI-powered automations that continuously monitor your ServiceNow instances and provide intelligent recommendations. Unlike simple scripts or monitors, these agents use Yeti AI to analyze patterns, predict issues, and suggest specific actions.

Each agent specializes in a critical area of ServiceNow operations:

Agent Purpose Default Schedule
Health Monitor Track instance performance and predict issues Every 15 minutes
License Optimizer Identify unused licenses and cost savings Daily at 2 AM
SLA Guardian Prevent SLA breaches before they happen Every 5 minutes
Patch Manager Plan upgrades and assess risks Weekly
Incident Manager Auto-categorize and route incidents intelligently Every 10 minutes
AI Readiness Advisor Score how AI-ready your instance is and flag remediation work Weekly
Instance Audit (on-demand) 500+ checkpoint best-practice analysis Manual
Upgrade Readiness (on-demand) Pre-upgrade risk and buildable remediation backlog Manual

The 8 MSP Agents Explained

1. Health Monitor Agent

What it does: Continuously monitors your instance's vital signs, including active sessions, transaction volumes, response times, and incident patterns.

What you get:

  • Health score (0-100) with status classification
  • Anomaly detection for unusual patterns
  • Predictions for potential issues
  • Specific recommendations to maintain performance

Best for: Operations teams who need early warning before problems impact users.

Example insight: "Transaction volume increased 45% over the past hour while incident creation rate doubled. Recommend investigating potential service degradation before peak hours."

2. License Optimizer Agent

What it does: Analyzes user activity across your instance to identify license waste and optimization opportunities.

What you get:

  • List of dormant licenses (users inactive 60+ days)
  • Over-provisioned users (high-tier licenses for low-usage patterns)
  • Monthly cost savings potential
  • Specific reclamation recommendations

Best for: IT managers and finance teams focused on ServiceNow cost optimization.

Example insight: "Identified 23 Fulfiller licenses assigned to users with no login activity in 90+ days. Reclaiming these licenses could save approximately $4,600/month."

3. SLA Guardian Agent

What it does: Monitors all active SLAs in real-time, identifying at-risk items before they breach.

What you get:

  • At-risk SLAs with risk scores
  • Predicted breaches for the next 24 hours
  • "Silent" tickets (no activity, likely forgotten)
  • Specific actions: reassign, escalate, or reprioritize
  • Systemic issues (understaffed teams, process bottlenecks)

Best for: Service desk managers and anyone accountable for SLA compliance.

Example insight: "INC0045678 has 23% time remaining but no activity in 4 hours. Current assignee has 12 other critical tickets. Recommend reassigning to the Network-Tier2 group which has available capacity."

4. Patch Manager Agent

What it does: Assesses your instance's upgrade readiness and creates detailed upgrade roadmaps.

What you get:

  • Current version and support status analysis
  • Customization risk assessment (low to critical)
  • Plugin and store app compatibility review
  • Step-by-step upgrade recommendations
  • Detailed "Patch Deployment Plan" reports

Best for: Platform administrators planning major upgrades or maintaining compliance.

Example insight: "Instance is on Yokohama (extended support ending). Found 847 customised artefacts including 12 modified base system scripts. Recommend upgrading to Zurich in Q2 with a 3-phase approach to minimise risk. The Upgrade Readiness Agent has produced a buildable remediation backlog with human-review flags."

5. Incident Manager Agent

What it does: Analyzes incoming incidents to improve categorization, routing, and resolution speed.

What you get:

  • Auto-resolution candidates (90%+ confidence)
  • Category and subcategory recommendations
  • Intelligent routing suggestions
  • Trending issues and potential problems
  • Knowledge article recommendations

Best for: Service desk teams looking to reduce MTTR and improve first-call resolution.

Example insight: "15 incidents in the past hour match pattern 'VPN connection timeout after Windows update.' Auto-resolution confidence: 94%. Suggested knowledge article: KB0012456. Recommend creating a major incident if count exceeds 25."

6. AI Readiness Advisor

What it does: Scores how AI-ready your instance is - looking at data quality, ACL hygiene, scoped-app structure, naming conventions, and known anti-patterns - and turns the gaps into a prioritised remediation backlog.

Best for: Teams preparing to roll out Now Assist or SnowCoder Build Agent and who want a clean baseline before they start.

7. Instance Audit Agent (on-demand)

What it does: Performs a comprehensive best-practice audit across all aspects of your instance.

What you get:

  • Executive summary with overall health grade
  • Detailed analysis across six areas:
    • Instance health and performance
    • License utilization
    • SLA compliance posture
    • Upgrade and patch status
    • Flow automation governance
    • Customization complexity
  • Prioritized recommendations with impact ratings

Best for: Architects, consultants, and anyone preparing for audits or major initiatives.

Example insight: "Overall Instance Grade: B+. Critical finding: 34 flows have no error handling configured, creating silent failure risk. High-priority: 8 ACLs grant admin-equivalent access to non-admin roles."

8. Upgrade Readiness Agent (on-demand)

What it does: Inspects your current release, customisations, plugins and store apps and produces a buildable remediation backlog for the next ServiceNow upgrade, with human-review flags on the risky items. Pairs with the Patch Manager agent for end-to-end N-1 compliance.

Best for: Platform teams running upgrades from Zurich onwards. See Zurich to Australia: Upgrade Readiness Agent in action for a walk-through.

Getting Started with Agents

Step 1: Connect Your ServiceNow Instances

Before creating agents, ensure your ServiceNow instances are connected to SnowCoder. Each agent needs at least one instance to monitor.

Navigate to Settings → ServiceNow Instances and verify your connections are active.

Step 2: Create Your First Agent

Option A: One-Click Auto-Setup

The fastest way to get started:

  1. Go to MSP Agents in the navigation
  2. Click Auto-Setup
  3. SnowCoder creates all six scheduled agent types for every connected instance (the two on-demand agents you trigger manually when you need them)

This is ideal for getting comprehensive coverage immediately.

Option B: Create Individual Agents

For more control:

  1. Click Create Agent
  2. Select the agent type
  3. Give it a name and description
  4. Choose the instances to monitor
  5. Configure the schedule (or use defaults)
  6. Click Create

Step 3: Start the Agent

New agents are created in a stopped state. To activate:

  1. Find your agent in the list
  2. Click the Start button
  3. The agent begins its scheduled monitoring

You can also click Execute Now to trigger an immediate run.

Step 4: Monitor Results

Once running, your agent dashboard shows:

  • Status: Running, stopped, paused, or error
  • Last Run: When the agent last executed
  • Activity: Recent findings and recommendations
  • Token Usage: Cost tracking (24h, 7d, 30d, all-time)

Step 5: Review Reports

Agents generate detailed reports after each execution. Access them in the Reports tab:

  • View reports directly in markdown format
  • Download for offline review or sharing
  • Reports are stored securely and accessible anytime

Best Practices for Agent Configuration

Match Schedules to Your Needs

Scenario Recommended Configuration
Production instance, business-critical SLAs SLA Guardian every 5 minutes
Development instance, low priority Health Monitor every 30 minutes
Cost-conscious organization License Optimizer daily, others weekly
Pre-upgrade assessment Run Instance Audit manually before planning

Start Conservative, Then Tune

Begin with default schedules and adjust based on:

  • Token consumption (visible in the Metrics tab)
  • Value of insights (are you acting on recommendations?)
  • Instance criticality (production vs. development)

Assign Multiple Instances Strategically

You can assign multiple instances to a single agent. This is efficient when:

  • Instances have similar configurations
  • You want consolidated reporting
  • Token budget is limited

Create separate agents when:

  • Instances need different schedules
  • You want isolated activity logs
  • Different teams own different instances

Use Pause for Maintenance Windows

During planned maintenance:

  1. Pause agents (keeps schedule, skips execution)
  2. Complete your maintenance
  3. Resume when ready

This prevents false alerts during expected downtime.

Understanding Agent Output

Activity Logs

Every agent execution creates activity log entries with:

  • Severity: Info, Warning, Error, Critical
  • Type: Execution, recommendation, alert
  • Message: What the agent found
  • Timestamp: When it happened

Filter logs by date range, severity, or instance to find specific events.

Metrics

The Metrics tab shows:

  • Execution frequency: How often the agent runs
  • Token consumption: Input and output tokens per run
  • Cost tracking: Cumulative spend over time
  • Success rate: Percentage of successful executions

Reports

Reports provide structured analysis in markdown format:

# Health Monitor Report - Production Instance
Generated: January 28, 2026 14:30 UTC

## Executive Summary
Overall Health Score: 87/100 (Healthy)

## Key Findings
- Transaction volume stable (+3% week-over-week)
- No critical anomalies detected
- Memory utilization trending upward (monitor recommended)

## Recommendations
1. Review scheduled jobs running between 2-4 AM (high concentration)
2. Consider archiving incidents older than 2 years (450K records)
...

Cost Management

Agents consume tokens for each execution. Here's typical usage:

Agent Tokens per Execution Monthly Cost (Default Schedule)
Health Monitor ~3,500 ~$15-20
License Optimizer ~4,500 ~$2-3
SLA Guardian ~4,000 ~$50-60
Patch Manager ~7,000 ~$1-2
Incident Manager ~4,500 ~$30-40
Instance Audit ~14,000 Manual (per execution)

Tips for managing costs:

  1. Adjust schedules: Run less frequently on non-critical instances
  2. Pause unused agents: Don't pay for monitoring you're not reviewing
  3. Use Auto-Reload: Set token thresholds to avoid interruption
  4. Review Metrics: Identify agents with low ROI

Real-World Scenarios

Scenario 1: Preventing SLA Breaches

Setup: SLA Guardian monitoring production instance every 5 minutes

What happens:

  1. Agent detects INC0098765 has 18% time remaining
  2. Notes the ticket has been silent for 2 hours
  3. Checks assignee workload (8 other critical tickets)
  4. Generates recommendation: "Reassign to Database-Team which has capacity"

Your action: Review recommendation, reassign with one click, SLA saved.

Scenario 2: Quarterly License Review

Setup: License Optimizer running daily

What happens:

  1. Agent identifies 45 users with no login in 60+ days
  2. Calculates $8,200/month potential savings
  3. Flags 12 users with Fulfiller licenses who only view dashboards

Your action: Export the list, coordinate with managers, reclaim licenses before renewal.

Scenario 3: Upgrade Planning

Setup: Run Instance Audit manually before major upgrade

What happens:

  1. Agent analyzes entire instance (takes 2-3 minutes)
  2. Identifies 23 customized Business Rules affecting upgrade
  3. Flags 5 deprecated API usages that will break
  4. Generates detailed upgrade roadmap with phases

Your action: Use the report to scope upgrade project and allocate resources.

Troubleshooting

Agent Shows "Error" Status

  1. Check the Activity tab for error details
  2. Common causes:
    • ServiceNow instance connection lost
    • Insufficient token balance
    • API rate limiting

Agent Not Running on Schedule

  1. Verify agent status is "Running" (not "Paused" or "Stopped")
  2. Check that assigned instances are connected
  3. Review the configured schedule in Settings

Reports Not Generating

  1. Ensure execution completed successfully
  2. Check token balance (report generation requires tokens)
  3. Wait a few minutes (report generation is asynchronous)

What's Next?

Agents are most powerful when integrated into your operational workflow:

  1. Set up email notifications: Get alerted when critical issues are found
  2. Schedule weekly report reviews: Make agent insights part of your routine
  3. Combine with conversations: Ask SnowCoder to explain or act on agent findings
  4. Scale across instances: Add agents as you connect more ServiceNow environments

Summary

SnowCoder agents transform ServiceNow operations from reactive to proactive. Instead of discovering issues when users complain, you're alerted before problems impact service levels.

Start with the SLA Guardian on your production instance. It's the fastest path to measurable value. Then expand to Health Monitor and License Optimizer as you see the benefits.

Your instances are running 24/7. Now your monitoring can too.

AI experts for ServiceNow

See how Kumoco and SnowCoder can accelerate your ServiceNow delivery.

Talk to our experts