IgniteAI Agent is designed to save time and reduce repetitive work, while protecting instructional quality, data integrity, and institutional trust. Some limits are intentional. Others reflect features that need to be developed.
Understanding these boundaries helps you use the Agent effectively and confidently.
Intentionally won’t do
These limits are by design and help ensure responsible, transparent use of AI in education.
- It won’t replace educator judgment
- IgniteAI Agent does not make final instructional decisions. You review and approve all content, summaries, and changes before they are shared with students.
- It won’t act outside your permissions
- The Agent can only take actions you are authorized to perform. It respects role-based access, course settings, and institutional controls at all times.
- It won’t bypass privacy or security safeguards
- IgniteAI Agent operates within Instructure’s privacy, security, and compliance standards. It does not use institutional or student data to train models without explicit consent.
- It won’t modify grades in the Gradebook.
- Grading must be handled with utmost sensitivity. While the Agent won’t grade student work, we’re considering other IgniteAI features to support those workflows.
Can’t do yet
These capabilities are not available today. The Agent continues to evolve, but for now, it works within the boundaries below.
- It can’t work directly with uploaded or external content
- The Agent can’t yet process uploaded files or pull in information from outside your Canvas environment. It also does not provide real-time or external web information.
- It can’t execute complex workflows in bulk
- While the Agent can complete individual or scoped actions, it doesn’t yet support multi-step operations across large sets of content or courses in a single request.
- It can’t coordinate across other tools or integrations
- The Agent can’t yet route tasks to other IgniteAI tools, connect to question banks, or integrate directly with third-party tools, LTIs, or institution-built language models.
- It can’t reflect changes instantly in the UI
- Updates made by the Agent may not appear immediately. A page refresh is required to see changes reflected in your course.
- It can’t continue working if you navigate away
- The Agent cannot continue a task if you refresh the page, navigate to a different page, or close the Agent. If the Agent is thinking, wait until it finishes before leaving the page.
Use the Agent with confidence
IgniteAI Agent is built to support educators, not replace them. Its boundaries are there to protect trust, accuracy, and control, while still delivering meaningful time savings where it matters most.
As capabilities expand, updates will be shared through release notes and product communications.