Connect UniversitySite to Your AI Platform

AI Knowledge Connector Extension Setup Guide

These instructions show how to connect UniversitySite to Microsoft Copilot, ChatGPT Enterprise, or another AI platform so it can answer questions about your firm’s training content and link users directly back to UniversitySite.

Estimated setup time: 15–30 minutes No custom development required


What This Integration Does

Once configured:

  • UniversitySite exposes export APIs
  • A scheduled PowerShell script extracts structured learning content
  • The data is staged in a customer-controlled storage location (e.g., SharePoint, Azure Storage, file system)
  • Your AI platform indexes that staged content as a knowledge source
  • AI responses include direct links back to UniversitySite

There are no live API calls during AI conversations.


Step 1 — Choose a Staging Location

Select where exported files will be stored before your AI platform indexes them.

Common options:

  • SharePoint document library
  • Azure Blob Storage
  • Secure file server
  • Network file share

Create a folder dedicated to UniversitySite exports.

Example (SharePoint):

AIKnowledgeExportsFromUniversitySite/

Leave the folder empty for now. The script will populate it.


Step 2 — Generate API Credentials in UniversitySite

  1. Log in to UniversitySite as an administrator.
  2. Navigate to:

InstructorSite → Global Settings → AI Knowledge Connector Setup

  1. Click Generate Credentials.
  2. In the same settings area, flip these switches to On so export text stays complete and up to date:
  • Automatically Extract Full Text From Resource Files
  • Automatically Extract Full Text From Video Courses

You will be shown:

  • Client ID
  • Client Secret

⚠️ Copy and store the Client Secret securely.

Authentication uses OAuth2 (client credentials). No end-user login is required.


Step 3 — Download the Sample PowerShell Script

From the AI Knowledge Connector Setup page:

  • Download the sample PowerShell export script.
  • Save it to a secure location on a server or workstation that can run scheduled tasks.

This script:

  • Authenticates to UniversitySite
  • Calls the export APIs
  • Writes structured JSONL files to your staging folder
  • Supports both full and incremental sync modes

Step 4 — Run the PowerShell Script (Do Not Modify It)

Use the provided script as-is. Do not edit or customize the script file. Configure execution using command-line parameters or environment variables.

Required parameters

You must provide:

  • baseUrl
  • clientId
  • clientSecret

Run example:

powershell .\Export-AiKnowledgeConnectorData.ps1 -baseUrl "https://yourfirm.universitysite.com/UniversitySiteYourFirm" -clientId "" -clientSecret ""

Environment variable option (instead of passing required params each run)

powershell $env:BASE_URL = "https://yourfirm.universitysite.com/UniversitySiteYourFirm" $env:CLIENT_ID = "" $env:CLIENT_SECRET = "" .\Export-AiKnowledgeConnectorData.ps1

Optional parameters

  • outputDir  — Write output files to a specific folder (default is current directory)
  • Verbose  — Show detailed logs for troubleshooting

Examples:

powershell .\Export-AiKnowledgeConnectorData.ps1 -baseUrl "https://yourfirm.universitysite.com/UniversitySiteYourFirm" -clientId "" -clientSecret "" -outputDir "C:\AIExports\UniversitySite"

powershell .\Export-AiKnowledgeConnectorData.ps1 -baseUrl "https://yourfirm.universitysite.com/UniversitySiteYourFirm" -clientId "" -clientSecret "" -outputDir "\\fileserver\AIExports\UniversitySite" -Verbose

Files created by the script

The script creates or updates these files in the output directory:

  • courses.jsonl
  • course_text.jsonl
  • resources.jsonl
  • resource_text.jsonl

Step 5 — Schedule the Script

Schedule the PowerShell script to run automatically.

Common options:

  • Windows Task Scheduler
  • Azure Automation
  • Enterprise job scheduler

Recommended frequency:

  • Nightly for most firms
  • More frequently if desired

Peak Hours Restriction

The export APIs are restricted during peak hours (9:00 AM – 5:00 PM server local time) to protect system performance.

Schedule your script to run outside peak hours — for example, overnight or early morning.

If you need to run an export during peak hours (for example, during initial setup or troubleshooting), contact Profiscience Support to request a temporary override.


Step 6 — Run the Script Once Manually

Before relying on automation:

  1. Run the PowerShell script manually.
  2. Confirm successful completion.
  3. Verify files appear in your staging folder.

You should see files similar to:

courses.jsonl resources.jsonl course_text.jsonl resource_text.jsonl

If files are present and populated, the export is working.


Step 7 — Add the Folder to Your AI Platform

For Microsoft Copilot Studio

  1. Open Copilot Studio.
  2. Open your agent.
  3. Go to KnowledgeAdd Knowledge Source.
  4. Choose your staging location (e.g., SharePoint).
  5. Select the UniversitySite export folder.
  6. Save and allow indexing to complete.

For ChatGPT Enterprise

  1. Create or edit your GPT.
  2. Add files or connect to your storage location.
  3. Select the UniversitySite export folder.
  4. Allow indexing to complete.

Step 8 — Test Your AI Assistant

Try questions such as:

  • “Which course covers data privacy? Include links to relevant courses and resources.”
  • “Do we have training on GDPR? Include links to relevant courses and resources.””
  • “What internal resources explain CCPA? Include links to relevant courses and resources.””
  • “Where can I find our harassment prevention training? Include links to relevant courses and resources.””

Responses should:

  • Reference relevant courses or resources
  • Include direct links to UniversitySite
  • Be grounded in your firm’s curated learning content

How Updates Work

Depending on your configuration:

Full Snapshot Mode

  • Rebuilds export files each run
  • Reflects additions, edits, and deletions automatically

Incremental Mode

  • Exports only modified content
  • Maintains efficiency for large content libraries
  • Derives the last synced point from the most recent dateEdited  in each existing JSONL export file
  • Removes records for deleted courses/resources during cleanup

Your AI platform will re-index updated files according to its own schedule.


Security Notes

  • UniversitySite is not accessed during AI conversations
  • Export credentials are application-level and rotatable
  • Data is staged in your controlled storage environment
  • You control which AI systems ingest the exported content

Troubleshooting

No Results in AI Responses

  • Confirm the script ran successfully
  • Verify export files are not empty
  • Confirm your AI platform has completed indexing

Script Fails to Run

  • Check stored credentials
  • Confirm network access to UniversitySite
  • Review PowerShell error output

Script Returns a "Service Unavailable" Error

The export APIs are unavailable during peak hours (9:00 AM – 5:00 PM server local time). Re-schedule the script to run outside those hours. If you need to run during peak hours, contact Profiscience Support to request a temporary override.


Summary

Once configured, the AI Knowledge Connector:

  • Keeps your AI assistant grounded in authoritative training content
  • Preserves UniversitySite as the system of record
  • Supports both full and incremental synchronization
  • Requires no runtime coupling to AI conversations

UniversitySite becomes a structured, curated knowledge source within your firm’s AI ecosystem.

If you need assistance, contact your UniversitySite administrator or support team.

Still need help? Contact Us Contact Us