SMART objectives are widely associated with George T. Doran's 1981 management article. Later evaluation literature has also warned that context matters when SMART criteria are applied. For employee onboarding, the wording should help a person understand good work, not turn every day into a numerical target.
Begin with the work, not the acronym
Start by asking what a credible result would look like in the real role. Then identify the quality standard, the evidence a manager can reasonably check, and the point at which the result should be reviewed. A useful outcome should be concise enough to guide action.
Separate the result from the evidence
The result describes what should be achieved. The pass standard describes the quality or condition that makes the result acceptable. The evidence explains what the new hire should provide or what an authorised manager should observe. Keeping these elements separate makes the expectation easier to understand and review.
- Broad instruction
- Improve our social media and get recruiters more involved.
- Outcome
- By Day 30, maintain a four-week approved content calendar across the agreed social channels, with named publishing responsibilities and a weekly performance review.
- Pass standard
- Each planned item has a channel, owner, and due date; required approval is recorded; missed activity has a named recovery action.
- Evidence
- Approved calendar, published links, approval record, and weekly performance summary.
Avoid four common mistakes
- Do not ask the new hire to promise a lagging result they cannot control, such as a guaranteed follower count.
- Do not combine several independently assessable outcomes into one sentence.
- Do not ask for evidence that is excessive, irrelevant, or difficult to verify.
- Do not treat polished writing as proof that the underlying work was completed.
Use AI as an authoring assistant
In Primi Centum, AI may draft an outcome, pass standard, and evidence requirement from manager-provided authoring context. The suggestion remains editable and requires human approval. It does not judge employee performance or make an employment decision.
The Information Commissioner's Office explains that meaningful human oversight requires reviewers to stay engaged, interpret the output, and retain the authority to challenge it. That is a useful standard for any AI-supported HR authoring workflow.
Is the result specific to the role? Is the standard proportionate? Is the evidence relevant and fair? Can the person control the expected work? Does the wording respect company policy, reasonable adjustments, and lawful process?