Writing annual performance reviews is often a daunting task of reconstructing an entire year from memory. I recently experimented with using Microsoft Copilot to help draft my review, and the results were a massive time-saver.
While Copilot did the heavy lifting of gathering data, it’s important to clarify: this tool isn’t doing the review for you. It is acting as a highly efficient archivist and drafting assistant. Here is how I approached the process and why a hybrid human-AI workflow is essential.
Here is the original prompt:
Please review my email, meetings and Teams notes to help me prepare my annual review. It is a bit challenging in that my role changed from Assistant Director in OIT Academic Media Production and Engineering to OIT Media Architect in January 2026. I also need to note at some point I was out on sick leave for December 2025. As assistant director, please focus on team management and administration for 3 engineers and 5 production staff, Panopto, Warpwire, Media Beacon, Interactive Video Application (STQRY), Captioning and AI Media Engineering and Training. My focus as Media Architect as acting focused on engineering and training and lecturing on AI video and other AI work such as general AI communications and work like my automation for things like my receipt workflow tool.There are specific reports that need to be responded. I will send them to you at one at a time. I will need for you to be brief but complete with the information you have in each area discussed. Question 1: List your goals, highlights, and accomplishments for the current evaluation year.
The Workflow: Data Mining at Scale
The real power of Copilot in this context is its integration with the Microsoft 365 ecosystem. Because it has secure access to my email, Teams chats, and calendars, it can “see” the year in a way that is difficult for a human to quantify quickly.
Key Strategies for Success:
-
Date Specificity: I constrained the search to the specific fiscal/calendar year. This prevents the AI from pulling in outdated achievements. You can see that I provided specific areas and services of interest for my review.
-
Role Transitions: Since my role changed during the year, I explicitly informed the prompt of that shift. This helped Copilot categorize my “before and after” responsibilities correctly.
-
Targeted Areas: I identified the specific projects I managed and technical systems I developed. By giving Copilot these “anchors,” it could ignore the noise and focus on high-impact contributions.
Why the “Word Buffer” is Essential
Even though Copilot provided a fantastic structure, I didn’t paste them directly into our official system (Kuali Build). Instead, I used Microsoft Word as an intermediate staging area for two critical reasons:
-
The Human Editing: You must edit for tone and accuracy. Copilot provides the “what,” but you provide the “why” and the professional nuance.
-
Data Security & Reliability: During my process, I noticed that some changes saved directly in Kuali didn’t stick due to session timeouts or errors (it’s a web app). Having a master Word document ensured that none of my polished work was lost.
A Note for Managers and Directors
To be clear: I am not officially recommending this as a departmental standard. Decisions regarding the use of AI tools like Copilot for performance management rest entirely with department leadership and HR policy. However, as a proof of concept, this experiment showed that AI is uniquely suited for the “data gathering” phase of a review—allowing us to spend less time digging through old emails and more time reflecting on our professional growth.
The Takeaway: Copilot is for drafting, Word is for refining, and you are for the final review. Use the tools to save time, but keep your hands on the wheel.
One more thing…
As awesome was Copilot, I asked them to draft a blog for this process in the blog. I don’t know if that was a feature or a bug, but I typed a prompt a day after I had done the draft for the review and got this error:
I used Snagit because I couldn’t find any way to export the chat. I imported it into Gemini and it helped create the blog draft that you are now reading. Another very important is having options when you’re blocked.
