Research Project: This is a free AI research project. No warranties, SLAs, or company associations. Learn more

Back to Guides
AI & Automation
Advanced25 min

AI Agents: Backup, Security, Capacity, Cost

Deep dive into how autonomous AI agents monitor and remediate issues across your clients

BrainstormMSP's AI agents work autonomously to monitor your clients, identify issues, and take action—all while learning from your environment.

1

Agent Overview

Available Agents

Backup Agent

Monitors backup success across all platforms

Verifies recovery capability

Predicts capacity issues

Creates tickets for failures

Security Agent

Tracks control status in real-time

Correlates security events

Prioritizes by risk level

Recommends remediation

Capacity Agent

Monitors resource utilization

Predicts when upgrades needed

Identifies optimization opportunities

Generates procurement alerts

Cost Agent

Tracks cloud spending

Identifies waste

Suggests right-sizing

Alerts on budget overruns

2

Configuration

Enable Agents

1. Go to **Settings > AI Agents**

2. Select agents to enable

3. Configure per-agent settings:

- Monitoring frequency

- Action permissions

- Notification preferences

Agent Permissions

Choose what agents can do:

**Monitor Only**: Alert but don't act

**Recommend**: Suggest actions for approval

**Auto-Remediate**: Take action automatically

Per-Client Configuration

Override global settings for specific clients:

Disable certain agents

Adjust thresholds

Change notification preferences

3

Custom Prompts

Customize Agent Behavior

Agents accept custom instructions:

Example: Backup Agent

For clients in the "Financial" tier:

Alert on any backup older than 12 hours

Require immutable backup verification

Escalate to senior tech if unresolved in 2 hours

Example: Security Agent

Ignore failed login events from known IP ranges:

192.168.1.0/24 (internal network)

10.0.0.0/8 (VPN range)

Testing Custom Prompts

1. Create prompt in staging mode

2. Run against historical data

3. Review agent decisions

4. Adjust and promote to production

4

Performance Tuning

Optimize Agent Performance

Reduce Noise

Tune thresholds to avoid false positives

Create suppression rules for known issues

Group related alerts into single ticket

Improve Accuracy

Review agent decisions weekly

Provide feedback on incorrect actions

Update prompts based on learnings

Scale Efficiently

Stagger client scans to avoid peaks

Use caching for frequently-accessed data

Archive historical data appropriately

Monitoring Agent Health

Dashboard shows:

Agent uptime and status

Actions taken per period

False positive rate

Average response time

Completed!

You've completed the AI Agents: Backup, Security, Capacity, Cost guide. Ready to continue learning?

Related Guides