Skip to main content

Analyze Sessions JSON with AI

The Sessions JSON export can be used to review completions, expiry, progress, repeat attempts, and, when included, learners' recorded responses. The downloadable analysis instructions tell an AI assistant how to interpret the export and produce a management report. They work with session-only and response-inclusive exports.

info

For changes across retaken training attempts, use Advanced mode > Report options > Include all sessions before downloading JSON. The default report includes only the latest relevant sessions.

In this section​

Prepare the files​

  1. Open Reports for the content and population you want to analyze. Switch to Advanced mode. For retake comparisons, enable Include all sessions under Report options.
  2. In Detailed results, choose Download JSON. Select the session and user fields you need. Include User ID for person-level, cohort, and retake analysis. Interactions is selected by default; keep it selected to analyze learner responses, or clear it for session-only analysis. Fields marked Contains personal information must be selected explicitly.
  3. Save the JSON export and download the analysis instructions (.txt). You do not need to edit the instructions.

Use an AI environment approved for the user data you selected. The JSON download does not offer password encryption.

Use an AI assistant​

ChatGPT​

  1. Start a chat that supports file attachments.
  2. Attach the JSON export and sessions-json-analysis-instructions.txt in the same conversation.
  3. Send: “Analyze the attached JSON file according to the attached training result analysis instructions.”

Claude​

  1. Start a chat or project where you can add files.
  2. Add the JSON export and sessions-json-analysis-instructions.txt to the same conversation.
  3. Send: “Analyze the attached JSON file according to the attached training result analysis instructions.”

Codex​

  1. Save the JSON export and instruction file in one local folder.
  2. Open that folder in a Codex local project so Codex can read both files.
  3. Ask: “Read the Sessions JSON export and sessions-json-analysis-instructions.txt in this folder. Analyze the JSON according to the instructions and provide a management report.”

Copilot​

  1. In Microsoft Copilot, start a chat that accepts both files and attach the JSON export and instruction file. In GitHub Copilot Chat, put both files in your working folder and add them as chat context.
  2. Confirm that the chat can read the contents of both files.
  3. Send: “Analyze the Sessions JSON export according to sessions-json-analysis-instructions.txt and provide a management report.”

For any assistant, you can add the intended audience, reporting language, course scope, or output format to your request. If you omitted Interactions, ask for operational and completion analysis; answer-level findings require recorded responses. If you are comparing retakes, confirm that Include all sessions was enabled in the report before exporting.

Request a visual PDF​

If the assistant can create files, you can follow up with: “Create a presentation-ready PDF of this report with clear, attractive graphs for the most useful findings. Label counts and denominators, keep the graphs consistent with the report, and check every page before sharing it.” You can also request a chart-ready data file or other useful output. If PDF creation is unavailable in your chosen tool, ask for the report and chart data instead.

Review the report​

Check that the assistant used the actual exported files, identified the report's population and date, showed counts alongside percentages, and separated findings supported by responses from those supported only by session data. Review recommendations against the evidence before sharing them.