AI Usage Boundaries & Controls

Overview

AI Agents in SEOP operate within defined boundaries — they cannot access systems outside their configured scope, cannot take autonomous action without logging, and cannot override human decisions.

What AI Agents Can and Cannot Do

Can Do	Cannot Do
Run automated tests against defined test suites	Modify production systems without human approval
Review code and surface quality issues	Access client data outside the defined workspace
Generate documentation from code and sprint data	Take irreversible actions autonomously
Flag delivery risks and surface analytics	Override engineer or client decisions
Integrate with connected tools (Jira, GitHub)	Connect to external systems not authorised in settings

Data Sovereignty

Client code and delivery data processed by AI Agents remains within the Scrums.com secure infrastructure. Data is not used to train external AI models. Enterprise clients can specify data residency requirements.

Audit Logs

On Enterprise plans, all AI Agent activity is captured in audit logs — including what the agent did, when, against which resources, and what the outcome was. Logs are available in Settings > Security > Audit Logs.

Overview

Catalog classes

Agent guidance

AI Usage Boundaries & Controls

Overview

What AI Agents Can and Cannot Do

Data Sovereignty

Audit Logs

​Overview

​What AI Agents Can and Cannot Do

​Data Sovereignty

​Audit Logs

Overview

What AI Agents Can and Cannot Do

Data Sovereignty

Audit Logs