How to document AI-assisted work in your appraisal
In this article (7 sections)
Related programme: the Data Analytics with Generative AI course (3–4 months; ₹90,000 (inclusive of GST) · EMI from ₹4,000/month).
To document AI-assisted work in your appraisal, describe the outcome first, then your role in it — the question you framed, the checks you ran, the errors you caught — and mention the AI tool as the method, not the achievement. "Cut the Monday report from four hours to ninety minutes by rebuilding it with AI-assisted SQL, with a reconciliation check that caught two source errors" credits you. "Used ChatGPT to make reports" credits the tool. The difference comes from what you record during the year, not from how you write at appraisal time.
Why this is harder than it looks
Appraisals reward visible contribution. AI-assisted work creates two problems:
- Invisible effort. A task that now takes an hour looks easy, even though you spent weeks learning to make it reliable.
- Credit confusion. Reviewers may wonder how much was you and how much the tool.
Both are solved by recording what you did, as you do it.
The weekly log
Keep a private note with one entry per week for any AI-assisted work that matters. Five lines are enough:
Week of 22 Sep 2026
Task: Monthly vendor spend summary
AI used for: drafting the SQL joins; first-pass commentary text
My part: defined "active vendor"; wrote reconciliation against finance total;
rewrote commentary for category heads
Checks: totals matched finance ledger to the rupee; 1 duplicate vendor found
Result: 3h instead of about 7h; duplicate reported to procurementEntries like this take two minutes each. At appraisal time you'll have dozens of concrete, dated examples instead of trying to remember.
Turning log entries into appraisal lines
Use a simple pattern: outcome → your judgement → method → evidence.
| Weak entry | Strong entry |
|---|---|
| "Used AI to automate reporting." | "Reduced the weekly sales report from about 4 hours to 90 minutes by rebuilding it with AI-assisted SQL; added a reconciliation step that caught 2 source-data errors in Q2." |
| "Explored GenAI tools." | "Tested an approved AI assistant on 30 past customer-comment summaries; found it missed negative sentiment in regional-language comments in 7 of 30 and recommended keeping manual review for those." |
| "Helped team with AI." | "Ran two 30-minute sessions showing four teammates how to check AI-drafted SQL; two now use the checklist weekly." |
| "Saved time with automation." | "Freed about 3 hours a week, which I used to investigate the Q3 returns spike; the analysis led to a supplier change." |
Notice the fourth row: saved time matters most when you say what you did with it.
What to emphasise
Reviewers want to know you can be trusted with the output. Highlight:
- Verification. The checks you designed and the errors they caught.
- Judgement about limits. Places where you decided not to use AI, and why. A good example is keeping exact totals and reconciliations out of the tool; see analyst tasks you should not hand to an LLM.
- Policy compliance. That you used approved tools and kept sensitive data out.
- Spread. Anything you taught others or turned into a reusable template.
What to avoid
| Avoid | Why |
|---|---|
| Time-saved figures you didn't measure | Easy to challenge; weakens the rest |
| Implying you did alone what the tool drafted | Honest attribution protects your credibility |
| Mentioning tools your company doesn't permit | Creates a compliance problem in your own appraisal |
| Listing tools without outcomes | Tells the reviewer nothing |
| Sharing company data in examples | Keep examples at the level of descriptions and counts |
If you estimated rather than measured time, say "about" and describe how you estimated — for example, from your calendar or the report's timestamps.
A self-check before submitting
Read each AI-related line in your appraisal and ask:
- 1Does it state an outcome someone cares about?
- 2Does it show a decision or check that was mine?
- 3Could my manager verify it?
- 4Is the tool mentioned only as the method?
If any answer is no, rewrite the line.
Preparing for the conversation
Your manager may ask, "How much of this was the AI?" Have a calm, specific answer ready: "The tool drafted the SQL and the first commentary. I defined the metrics, wrote the reconciliation, found the duplicate and rewrote the commentary for the audience." That answer shows you understand exactly where your value lies.
If you ran a formal pilot with your manager's agreement, the demo results belong here too; see how to ask your manager for AI upskilling time. The Data Analyst programme, which runs 3–4 months, teaches AI-assisted analysis with verification built in, so the checks you describe are part of how you work.
Frequently asked questions
Should I mention that I used AI tools in my appraisal?
If your company permits the tools and you used them to produce better or faster results, saying so is usually a strength, provided you describe what you decided and checked. Hiding it can backfire if colleagues know; overstating it undersells your own judgement.
Pankit Kumar has 10 years in Data Science & AI, building and shipping production systems in regulated pharma and clinical environments. He is a freelance trainer at Boston Institute of Analytics, AnalytixLabs and Scaler, and has taught this material to thousands of working professionals.
This article is part of our Data Analytics with Generative AI programme — 3–4 months. The full analyst stack — Excel, SQL, Power BI and Python pipelines — then a generative-AI layer you can prove is right.
Explore Data Analytics with Generative AI