20 Minutes to an AI-Powered Year-End Review That Hits All the Right Notes with Your Boss

Every year, the year-end review you present to senior leadership gets roasted:

• “This is garbage”
• “Nothing but fluff”
• “Where’s the data? Where’s the proof?”
• “How come I never heard about this?”
• “What is this? What even is this?”

This year, you don’t have to worry anymore. In just 20 minutes, AI can help you write a year-end summary that truly resonates with your boss.

Here’s what one of the report documents looks like: (Besides this one, there are several other deliverables: a PPT outline, a quantitative report, and a summary report. This one looked the most visually appealing, so I picked it for the demo.)

The summary report is also way better than what I could write myself — thorough and well-structured. Here’s a look at the outline:

Prerequisites

Work logs, such as:

• Weekly reports

An AI IDE or CLI (any of these will do), such as:

• Claude Code
• Cursors
• Qoder
• Even a VSCode AI plugin like: Tongyi Lingma

Just get it installed, configured, and verify that AI chat works.

Everything else that’s non-essential (or unfamiliar to non-developers) like version management and toolchain management — don’t worry about it for now.

Hands-On Walkthrough

My Weekly Report Overview

I’ll use my own case (IT Monitoring Team) as an example. My weekly reports roughly follow this format (honestly, they’re pretty messy):

The weekly report content looks like this (also quite messy — full of typos, no consistent structure. Just a rough logic of: monitoring → monitoring sub-items or projects → details):

Import

Move all your weekly reports into a new folder. This step keeps the project folder clean and prevents unrelated files from confusing the AI.

AI Prompt

Then open the folder with your AI IDE or AI CLI, using it as the workspace. Start typing your prompt. My prompt was pretty basic:

You are the monitoring lead of an IT operations team. The year-end review needs to showcase our work content and workload through data. Please create a quantitative summary based on the existing "weekly reviews."

(You might need to install Python first.)

Then just let the AI run on its own. When it asks to install tools, libraries, or run commands, you can mostly just click “allow.”

What the AI Did (Fully Automated)

Here’s what the AI did — all fully automated. I didn’t intervene or provide any additional input:

▽ Checked the project structure

▽ Reviewed several typical weekly review files to understand their structure and content format

▽ Created a script to analyze all weekly review files, extract key data, and perform quantitative analysis (a Python script — we don’t need to understand it, the AI does)

This script handled the following tasks:

1. Loading data
2. Analyzing data
3. Printing summary results
4. Generating a detailed report

▽ The first run threw an error (we don’t need to understand it either) — the AI will fix it on its own:

After the fix, it ran successfully and generated two outputs:

• 年终总结_量化报告.md
• weekly_analysis_report.json

▽ It also generated a PPT outline

▽ Next, it created a Python data visualization script to generate charts for better data presentation (we don’t need to understand this either)

Then it ran the script, installed dependencies…

▽ Finally, it produced a comprehensive summary document that consolidated all the analysis results:

And wrapped up with a final summary.

Done 🎉🎉🎉

The whole process took roughly 10–20 minutes.

Takeaways

This use of AI was honestly pretty rough around the edges, but even so, the AI spent far less time than I would have and produced reports far better than what I could write.

Here are some lessons learned:

1. As a data source, my weekly reports were pretty thin. If your weekly reports contain richer data, the results should be much better than mine. You can also feed in other important deliverables or documentation.
2. The AI’s general approach was to use Python to summarize the data, generate charts with NumPy and Matplotlib, and then leverage its own knowledge to write a standard year-end review.
3. The prompt I used was still quite vague. If your company has more specific requirements for year-end reviews, you can absolutely refine the prompt to better match your needs.
   1. Of course, this was a one-shot generation. You can also continue the conversation with the AI to iteratively adjust and refine your year-end review.
4. Some AI tools can generate PPTs directly. If the one you’re using can’t, you can ask it to generate a PPT-style webpage instead.

That’s it.