The origin of OKRgen
My interest in OKRs started a few years ago, roughly around the time I collaborated with Javier Martín as a colleague at Sngular. This interest then intensified during my tenure as Strategy Director at Neock and, today, has become a tool that often emerges in the transformation processes I assist with: In a world of constant change and growing competitiveness, establishing clear and measurable goals has become essential.
Inspired by my own interest in OKRs and by a publication from Javier Martín, I decided to delve into creating a simple tool that utilized generative artificial intelligence to assist individuals in taking initial steps, exploring, and finding inspiration in working with OKRs.
ChatGPT has been incorporated into my daily workflow ever since it helped me create an escape room for my daughter’s ninth birthday . It was my “Aha” moment, and we’ve been inseparable since.
As a trainer, facilitator, and consultant, I wanted to investigate:
- To what extent someone like me, who hasn’t programmed in over 15 years, could use a no-code tool to implement an app integrated with ChatGPT that would make it accessible to those who might not necessarily want or know how to use OpenAI’s interface or pay a subscription.
- To what degree a straightforward tool could support a learning process about OKRs.
- Could it even generate ideas about sectors and business areas unknown to an experienced consultant or trainer?
Determined to try, I chose a no-code tool (Bubble.io) and set to work to implement OKRgen ((http://okrgen.hugolopes.es/)), an OKR generator powered by ChatGPT. Altogether, I dedicated about two or three days of work, which includes time spent learning the no-code tool and OpenAI’s API.
I’m pleased with the outcome relative to the time invested. 😊 I believe this tool has the potential to broaden horizons for both professionals and newcomers in the OKR field, allowing exploration across a myriad of sectors and examples that can inspire and guide in the formulation of objectives and key results.
I’ve funded the OpenAI account of OKRGen with what I believe can cover the costs for about 20,000 OKRs, so I encourage you all to try it out. I’d love to receive your feedback on my LinkedIn (https://www.linkedin.com/in/hmlopes/).
The evolution of OKRGen’s Prompts
A prompt is a suggestion or idea that you provide to artificial intelligence to initiate or guide a conversation. It’s a cue that helps AI understand what you want to discuss and generate appropriate responses.
The creation of OKRgen led me to contemplate the significance of optimizing and simplifying the “prompts” used to guide AI. One notable example is how the original prompt, which was lengthy and detailed, evolved into a much simpler one. This optimization becomes particularly relevant considering the cost of using the API, measured in tokens (similar to the number of words).
The initial prompt of OKRgen was inspired by a publication from Javier Martín and extensively outlined some key principles when drafting OKRs. (The principles listed here describe the importance of making OKRs ambitious but attainable, qualitative, aligned with company mission and vision, among others, as well as criteria for defining both Objectives and Key Results.:
For a proper definition of the Objectives, it’s important to consider the following criteria:
- The Objective statement contains a verb, ideally indicating an achievement, change, or transformation. For instance: convert, transform, achieve, attain, etc.
- The Objective should not have a metric; it should be qualitative since we leave the numbers for the key results, which is the quantitative part of the OKR.
- Also, Objectives should be stated qualitatively since this is how we work to achieve the purpose, which will similarly be qualitative and not quantitative.
- The Objective should be Ambitious (transcendent and inspiring) and Practical (specific and action-oriented).
- The Objective should focus on the impact on the organization and not on a specific activity to be carried out.
- The timeframe for the Objective and its level of application is determined by the OKR, whether it’s quarterly or yearly, company-wide or team-specific.
- Lastly, something very important: an Objective is never a task, activity, or initiative to be carried out but rather a reflection of the impact that all of these have on an organization.
- On the other hand, when correctly stating the Key Results, you should follow these criteria:
- Key Results should be specific, realistic, verifiable, and measurable.
- A Key Result must always be quantitative because it should be a measure of the progress we’re making with our work towards achieving the Objective.
- Some people believe that a “Yes or No” Key Result is also valid, whether I achieved it or not, but we think it’s always better to be numeric for more precise progress assessment.
- If we state the Key Result in the past tense, we will be placing ourselves in a situation of having achieved the Objective, and this will help us better internalize it. For instance: we achieved, we accomplished, etc.
- The Key Result should also include a verb indicating advancement or progress, like increase, improve, grow, etc.
- There should always be multiple Key Results per Objective because in this way, we have a clearer vision of whether we’re truly achieving what we set out to do.
- A Key Result should never be a task, as we must avoid the complacency of doing the work without worrying about the impact it has on our strategy.
- Could you start by giving me an example for a company that specializes in “<activity sector>”?”
During a trial-and-error process, I simplified the prompt, given that ChatGPT already “knows” how to generate OKRs, and we only need to provide minimal context and underscore specific aspects. As such, the original prompt has evolved into its current form:
You are an OKR expert, an employee of a company, and you generate 1 Objective and 3 results. The objective should not contain numbers. Key Results must always be: quantified to a time horizon of X months, indicating a way to measure progress. A company in the <activity sector>
The potential of tools like OKRGen in learning processes
I believe this tool can be an ally in understanding and learning the OKR methodology. It encourages users to learn by playing and experimenting, facilitating the development of practical skills and knowledge that could potentially be applied in real companies.
By offering OKR examples, users can experiment with various sectors and areas, gaining deeper insight into how objectives and key results can be established in different contexts. This allows users to practically apply their knowledge and adapt it to their needs, thus completing an experiential learning cycle.
Given its extensive knowledge across diverse sectors and areas, ChatGPT can offer users a rich learning experience, even if they’re working independently. By exploring OKR examples in different contexts, users can access a wide variety of examples, akin to those they might acquire by participating in a community of practice with professionals from other sectors.
Moreover, OKR consultants and professionals can also use it to generate examples, enabling them to connect and empathize with the potential needs and goals of clients from industries they might not be familiar with in detail.
What do you think? Are you up for giving it a try?
