There has never been an easier time to write a good OKR. Feed any half-decent language model a description of your team and a paragraph on what you want to achieve, and within seconds you have an objective and three key results that read like they came out of a textbook. The wording is tight, the metrics sound credible, the structure ticks every box. The craft of writing goals, something that used to take experienced facilitators years to teach, has been almost entirely commoditised.
So why are so many of the teams I work with struggling more than ever to make their goals stick?
For readers unfamiliar with the framework, OKRs (Objectives and Key Results) are a way for teams and organisations to set ambitious goals and define a small number of measurable outcomes that show whether they are getting there. It is a simple idea with a few core formulation principles, but the value has always come less from the format itself and more from the conversation it forces a team to have.
That, I think, is where AI is quietly changing the picture, and not always for the better.
The perfect-goal trap
The first and most obvious thing AI does in a goal-setting process is produce goals that look, on the surface, faultless. The structure is clean, the language is sharp, and the metrics sound like the right ones to track. However, perfectly formulated does not mean yours.
I have seen this play out repeatedly. A team accepts an AI-drafted set of OKRs, nods along, and then a few weeks later realises they have no idea how to act on them. The metrics rely on data the team cannot actually collect. The objective uses language no one on the team would naturally choose. The goals are technically excellent and emotionally detached. They do not feel owned, because they were not built.
A rule of thumb worth keeping in mind: a goal you did not argue about is rarely a goal you will fight for.
AI as a sparring partner, not an author
The more interesting use of AI in goal-setting is, I would advocate, the opposite of what most teams reach for first. Instead of asking it to write the goals, use it to challenge them.
An AI that knows the OKR craft is a remarkably good coach. It can ask the uncomfortable questions a team might avoid asking itself. Is this key result actually measurable, or just numeric? Does this objective describe an outcome, or merely an activity? Is this ambition real, or just safe wording dressed up as a stretch?
In this mode, the team still writes the words. The team still owns the discussion. AI becomes a sparring partner that sharpens the thinking, rather than a ghostwriter that replaces it. That distinction sounds small, but in practice it determines whether the goals will live in the team’s daily work or sit unused in a document.
The bird’s-eye view, the real frontier
Looking further ahead, there is a third shift I find genuinely exciting, and it has very little to do with formulation. It has to do with perspective.
Imagine an AI that does not just help one team write better goals, but understands the goals of every team in the organisation. It can spot where two teams are quietly working on the same problem with different words. It can flag a sales team’s ambition that has no marketing support behind it. It can show a leader that what looks like a coherent strategy on paper is, in practice, a collection of disconnected priorities.
This is where AI moves from being a writing assistant to being something closer to a strategic mirror. The value is not in writing the goals at all. It is in seeing the whole picture and feeding that insight back into how each team’s goals are shaped. Done well, this is how AI can finally help with the alignment problem that organisations have wrestled with for decades.
Where I would draw the line
For all the upside, there is a caution that I think gets lost in the rush to automate. Goal-setting is one of the most inherently human things a team does together. The discussion about what matters, what to leave behind, and what to commit to is the very thing that creates engagement. The output of that conversation is a document. The value of it is the conversation itself.
If we let AI do the thinking, we get the document, but we lose the conversation. And without the conversation, the document is just words.
A personal note
This thinking is what is driving the work I am currently doing on OKRnest, a goal-setting platform built around the idea that AI should amplify the human conversation rather than replace it. It integrates AI into both goal creation and an organisation-wide overview of where goals stand, but the design starting point is that the team still leads the thinking.
The reason I keep coming back to this principle is a moment I witnessed in a coaching session. The team had been in a tremendously strong conversation about priorities, focus, and what really mattered for the coming quarter. There was real energy in the room. Then, when it came time to formulate the OKRs, one participant turned to an AI tool and read out what it suggested. The OKRs were, on the surface, perfect. And the conversation died on the spot.
You can’t challenge perfectly formulated goals. That, more than anything else, is what I think we need to be careful about as AI takes its place in the goal-setting process. The question is not whether AI belongs in how we set goals. It is whether we let it do the thinking that should be ours.
