Learning paths/Put AI to work in the tenant

Put AI to work in the tenant

AI or platform engineer

Choose Trace vs Autonomic Agents vs skills, and apply MCP guardrails safely.

Outcome: You can configure the right agent style for a job and avoid unsafe tool exposure. · 8 stops · ~70 minutes

1.AI agents and MCP
Open stop
Maps the agent landscape so you do not confuse chat agents with unattended workers.
2.Trace AI Agent
Open stop
Trace is the interactive MCP chat agent with on-demand skills.
3.Tenant skills user guide
Open stop
Skills are how you extend Trace without hard-coding every tool.
4.Autonomic Agents
Open stop
Unattended workers for schedules, mailboxes, and event wakes.
5.MCP guardrails
Open stop
Guardrails keep agent tools from over-reaching tenant data.
6.Platform Engineer Agent
Open stop
Patterns for MCP-enabled engineering assistance against the platform.
7.Trace crafting operations-aligned work
Open stop
Connects AI help to Operations My Work publishing patterns.
8.API endpoints index
Open stop
Autonomic Agents and related HTTP contracts live under the API tags.

1.AI agents and MCP