Key Result Validator: checks that a key result is measurable
Key Result Validator is the agent that reads a key result and tells you whether it is measurable and meaningful. It takes the guesswork out of reviewing OKRs. Instead of debating whether "improve onboarding" is a key result, you get a short, specific answer. You run it on a key result, from a record action or as a workflow step, like any reusable agent in Worknano.
What it does
- Checks the key result against the usual SMART criteria: is it specific, measurable, achievable, relevant and time-bound.
- Looks closely at the title, the initial value and the target value, because those three decide whether progress can be measured at all.
- When the key result belongs to an objective, checks that it supports that objective by comparing it with the objective title.
- Marks the key result as valid when nothing needs to change, so you are not buried in noise.
- Otherwise recommends only the fields that should change and explains each recommendation in a sentence or two, using the field labels you see on screen.
When it helps
- At the start of a quarter, a team lead runs it on every draft key result before the planning meeting, so the debate is about priorities and not wording.
- A new manager inherits a set of OKRs. A quick pass shows which key results have a target but no starting value, so nobody can tell what "done" means.
- You add a validation step to the workflow that publishes OKRs, so every key result gets the same review before it goes live.
How it works with your data
The agent has no tools. It only reads the key result record that is handed to it, plus the objective title when there is one, and it replies with a verdict and suggested changes. It does not edit anything. You decide what to change and make the edit yourself. It runs on a fast model, which suits a short review that you may want to repeat many times. Everything stays inside your workspace permissions.
Make it yours
Every agent in Worknano is a record in the Agents module, and this one is no different. You can change its instructions to match how your company writes OKRs, switch the model, adjust the temperature, or duplicate it for a stricter version. Teams often add their own rules, such as "every key result needs a number and a date". The same interface and the same database serve every module, so what you learn here applies to all other agents. See Build your own agent and how to edit an agent.
Related: Objective Validator reviews the objective above the key result, and Chat Agent can answer OKR questions inside a thread. For the bigger picture, read about AI agents and AI assistants.