For example, I recently tasked ChatGPT with generating different business cases for my OKR workshop, with company-level OKRs and departmental OKR examples. (we couldn't use real examples from our clients due to confidentiality). ChatGPT's initial results were far from ideal. It made some common mistakes, like formulating cumbersome objectives and key results in the form of tasks. It should be understood that he was trained on the general abundance of information on the web, where there are also bad examples. That's where my expertise came in handy.

Here's an example of how I trained ChatGPT:

1. I provided a complete example that had been manually prepared by our team.
2. I gave ChatGPT our OKR quality checklist: Objective: Laconic, easily remembered; Emotionally motivating; Specific, easy to understand, what is to be achieved; Crucial for company strategy execution in the chosen period; Change (aimed at changes), and not RUN (routine). Key Results: Outcome, not a TASK; Quantitative; Measurable on a regular basis (weekly); Vital but sufficient set for reaching the Objective; Ambitious, with a clear value for the business
3. I instructed ChatGPT to adjust the case to align with the company's OKRs and with each department, and to make the objectives more inspirational and qualitative, with no more than 7 words. The key results should be outcome-based, quantitative, measurable on a regular basis, and ambitious.

By following these steps, we were able to improve ChatGPT's output and generate more effective OKRs for our workshop. With some guidance and training, ChatGPT can be a useful tool for businesses looking to optimize their OKR strategy. (As you maybe recognized, the last paragraph was written by ChatGPT themself after I asking him to improve my post)