/
boisestate.ai System Performance Details

boisestate.ai System Performance Details

This page contains technical facts, behaviors, and quirks related to http://BoiseState.AI — our university-built generative AI platform. It is meant for users who want specific, often operational details about the system.

Access & Login

  • Login Method: Uses Boise State Single Sign-On (SSO).

  • Username Format: Boise State Username and Password (same as other campus services).

  • Session Timeout: tied to your My.Boisestate.edu login cookies; estimate 8 hours

Session & Performance Details

  • Idle Timeout: [e.g., Sessions automatically end after 1 hour of inactivity].

  • Concurrent Sessions: [e.g., Up to 2 sessions per user at once].

  • Response Time Targets: [e.g., Under 5 seconds for most queries].

Model Behavior & Capabilities

  • this stuff all exists on the website, should we duplicate here?

  • Base Model Version: [e.g., LLM-X 2.1 fine-tuned for Boise State context].

  • Training Data: Includes campus policies, course catalogs, AI best practices.

  • Limitations: May not answer questions about third-party platforms unless integrated.

Update & Maintenance Schedule

  • Regular Maintenance Window: Sundays from 12am–4am MST.

  • System Status Page: [Link to status.boisestate.edu/ai]

  • Version History: [Link or brief bullets: e.g., “May 2025 – v1.2 released with Nearpod integration”]

Data & Privacy

  • Data Retention: [e.g., No prompt or response data is stored long-term].

  • PII Handling: The system does not process or retain personally identifiable information (PII).

  • Audit Logs: Maintained by OIT for up to 90 days for support diagnostics.

Integrations & Add-ons

  • Canvas: Integrated for content-aware responses.

  • Nearpod: Pilot support for AI-assisted lesson planning.

  • Other Tools: [e.g., Google Workspace integration in testing phase].

User Feedback & Troubleshooting

screenshot showing the profile menu opened, the option for Feedback is highlighted

  • Report Issues: The Boise State Help Desk has been trained to received calls and troubleshoot/document and escalate issues regarding this tool.

  • Known Bugs: [e.g., Sometimes delays when processing large PDFs].

Future Fact Categories to Anticipate

(Use these as placeholders or expand as needed)

  • Mobile App Behavior

  • Accessibility Compliance Info

  • API Access or Developer Tools

  • Language Support & Localization

  • Data Source Transparency

  • Faculty & Staff Use Patterns or Case Studies