Turn Limited Learning Insights into Action with ChatGPT in HR

Aggregate LMS, Quiz and Feedback Data into a Single View for ChatGPT

The first tactical step is to bring your scattered data into a format that ChatGPT can interpret. Export course-level data (ID, title, type, duration), learner-level status (role, department, completion, time spent) and assessment data (quiz scores, attempts) from your LMS. Add optional columns for NPS or satisfaction scores, plus free-text feedback from surveys or post-course comments.

Combine these into a single CSV or Excel file where each row represents a learner-course interaction (or, alternatively, a course with aggregated statistics by cohort). This doesn’t require a data lake; your BI or HRIS team can usually create a recurring export. When using ChatGPT, paste or upload a representative sample first (e.g., one quarter’s data for a specific program) to validate your prompts before scaling.

Example prompt to structure the data analysis:
You are an HR learning analytics assistant.
You will receive a tabular export from our LMS with the following columns:
- course_id, course_title, skill_tag
- learner_role, learner_department, tenure_months
- completed (yes/no), time_spent_minutes
- quiz_score_before, quiz_score_after
- satisfaction_score, feedback_text

First, summarize the data schema you see and confirm any assumptions.
Then propose 5-7 analytical lenses that would be most useful to
understand skill improvement and content effectiveness by role.

This approach lets ChatGPT understand the data structure before you ask more targeted questions about performance and impact.

Use ChatGPT to Identify Skill Improvements and Weak Spots by Role

Once ChatGPT understands your dataset, you can direct it to quantify learning impact. Focus on before/after quiz scores, practical assessment results, or certification outcomes. Ask the model to slice these by role, tenure band and department to reveal where content works and where it does not.

Example prompt to detect impact:
You are an AI learning analyst.
Using the uploaded dataset, please:
1) Calculate average quiz_score_before and quiz_score_after by
   course_title and learner_role.
2) Identify the top 10 courses with the largest average score
   improvement for each role.
3) Highlight any courses where there is high completion but
   minimal or negative score improvement.
4) Present results in a concise table and a narrative summary
   that a non-technical HR stakeholder can understand.

Expected outcome: a ranked view of which modules actually improve skills for each role, and which ones look ineffective despite high completion rates.

Mine Open-Text Feedback with ChatGPT to Refine or Retire Content

Free-text feedback hides some of the richest learning insights but is often ignored because nobody has time to read thousands of comments. ChatGPT can cluster this feedback, detect recurring themes and link them to specific modules or formats. Start by extracting only the columns course_title and feedback_text for a pilot program and upload them as a text file or table.

Example prompt for feedback analysis:
You are a learning experience researcher.
You will receive a list of course_title and feedback_text entries.
Tasks:
1) Group the feedback into 8-12 themes (e.g. too theoretical,
   unclear examples, great practice tasks, outdated content).
2) For each course_title, summarize which themes are most common.
3) Flag courses with a high proportion of negative themes, and
   suggest specific improvement actions (e.g. add practice
   exercises, shorten length, update examples).
4) Provide 5 anonymized example quotes per key theme.

This gives HR and L&D a prioritized list of modules that need redesign, retirement or promotion based on actual learner voice rather than gut feeling.

Build Simple, Repeatable Insight Dashboards with ChatGPT

Instead of manually building presentations for every L&D steering committee, let ChatGPT help you generate consistent insight packs. After analyzing your data, ask the AI to produce structured summaries by program, role and skill tag. You can then convert these into slides or use them to brief managers.

Example prompt for dashboard-style output:
Act as a learning analytics reporting assistant.
Based on the previous analysis, create a structured summary for
our 'Onboarding Sales Academy' program:
- 1-page executive summary (plain language, max 300 words)
- Key metrics: completion, time-to-complete, avg score lift,
  satisfaction, by role
- Top 5 most effective modules and why
- Top 5 modules to improve or retire and why
- 3 concrete recommendations for next quarter
Format the output with clear headings so we can reuse it in
slide decks.

Expected outcome: consistent, data-backed reports that can be refreshed monthly by re-running the same prompts on updated exports, dramatically reducing manual reporting time.

Create Adaptive Learning Recommendations with ChatGPT

Beyond analytics, ChatGPT can help you personalize learning paths based on identified skill gaps. Use quiz results and course metadata (e.g., skill_tag, difficulty, duration) to generate recommended next steps for different cohorts. This can be done as a batch process that outputs recommendations per role or even per individual, which you can then review before upload to your LMS or communication channels.

Example prompt for recommendations:
You are an L&D recommendation engine.
You will receive a dataset with learner_id, learner_role,
quiz_score_after per course, and course metadata
(skill_tag, difficulty, duration).

1) For each role, identify the 3 most critical skill_tags where
   average quiz_score_after is below 70% across completed courses.
2) For each of these skill_tags, recommend 2-3 existing courses
   from our catalog that best address the gap, considering
   difficulty and duration.
3) Output a table with columns: learner_role, skill_tag,
   recommended_courses (with short justification).

Expected outcome: practical recommendations that let HR move from generic catalogs to targeted development plans at role or cohort level, without building a full custom recommendation engine on day one.

Use ChatGPT to Simulate Budget and Portfolio Decisions

Once you know which programs drive the most skill improvement, you can use ChatGPT to model the impact of reallocating budget. Provide the AI with per-course cost estimates (development and delivery), participation numbers and measured impact (e.g., average score lift or certification rates). Ask it to simulate scenarios like “what if we cut the bottom 20% of low-impact courses and reinvest in the top 10%?”

Example prompt for portfolio optimization:
You are an L&D portfolio strategist.
You will receive a table with course_title, annual_cost,
number_of_learners, avg_score_improvement, and satisfaction.

1) Classify courses into 3 groups: High Impact, Medium Impact,
   Low Impact based on cost vs. score improvement.
2) Estimate potential cost savings if we discontinue Low Impact
   courses and reduce Medium Impact courses by 30%.
3) Suggest how to reinvest these savings into High Impact
   courses (e.g. more cohorts, localization, blended formats).
4) Provide a concise narrative that HR can use to justify
   budget shifts to finance and business leaders.

Expected outcome: clearer decisions on which content to scale, fix or stop, along with narratives that help HR credibly argue for smarter L&D investment.

If implemented step by step, organisations typically see a 20–40% reduction in time spent on manual learning reporting, significantly better visibility into which 10–20% of the catalog delivers most impact, and stronger arguments for reallocating — or increasing — L&D budgets based on real learning effectiveness data.