All stories

Seven tips for writing performance reviews with generative AI


  1. Check your company's AI policy first: You might be limited to approved tools, specific data for generative AI, or no generative AI at all. Get all the details before you start.
  2. Avoid sharing sensitive or confidential info: Unless you're using an approved tool vetted for data privacy and security and are sure about including employee and organizational details, stick to general and anonymous content.
  3. Avoid leaving in bias: Your words or the LLM's words are likely to contain bias. Writing a bias-free review without performance review software pointing out unfairness, harm, or perpetuating stereotypes is tough. Don't forget to review for personality traits-based comments, instead of work-related ones, unclear phrases, feedback the person can't act on, and needlessly gendered terms. *Tip* Consider taking Textio’s free equitable feedback course to train your eye for these issues. You can also explore our equity-focused performance review templates.
  4. Use it to consolidate data from various sources: Gather all the info and feedback on your employee's performance, then let  AI identify trends and summarize critical themes. This can help you shape the content of the review. Remember to always double-check to ensure accuracy, as the AI might have misinterpreted some details.
  5. Use it to edit: Develop a rough performance review draft or jot down an outline of the points and phrases you want to use, then ask AI for a hand. Get it to trim or expand sections, suggest different ways to word tricky parts, and fix any spelling or grammar slip-ups. Tools like Textio Lift can even give you feedback on how clear, actionable, relevant, and biased your writing is, giving tips for improvement. 
  6. Take the time to nail down your prompts:  Instruct AI to base recommendations solely on your specific input and skip telling it to avoid bias—you might get strange results. Share the performance review examples, frameworks, and models you want it to follow. However, if you're using a purpose-built tool for performance feedback, you won't need to fine-tune your prompts like this because they're all pre-programmed.
  7. Fact check: General LLMs might sound spot on, but don't fall for it. They can eloquently spew out nonsense. So, verify every single word.
Using generative AI for writing performance reviews can level up your feedback game and save you time and stress. It's all about knowing how to use it correctly and picking the right tool to get your desired results.
Plus, it is always a good idea to review and edit the AI-generated content before finalizing the performance review, as human input and perspective are still essential in providing effective feedback.
All stories
Get stories like these delivered right to your inbox.