Assess high agency through evidence of how a person notices a problem, chooses an action, involves the right people, and follows through. Confidence, long hours, fast speech, or an impressive employer are not substitutes for that evidence.
The phrase is useful only if the hiring team agrees what it means for the role. An engineer resolving an unclear reliability issue and a seller investigating a new buyer segment may both show initiative through very different work.
Define the behavior you need
Choose an example of ambiguity the role will face. State the responsibility the person will have and the boundaries they must respect.
| Behavior | Evidence question | What to examine |
|---|---|---|
| Notices a meaningful problem | What did you notice before being assigned a task? | Connection to a real need |
| Chooses a useful first step | What did you investigate before acting? | Judgment under uncertainty |
| Works across boundaries | Whose input or agreement did you need? | Collaboration and decision rights |
| Follows through | What happened after your first attempt? | Revision, ownership, and learning |
Use consistent core questions and criteria across candidates. That follows the basic design of structured interviews described by OPM.
Separate initiative from unauthorized action
Moving quickly is not always useful. Ask what the candidate deliberately avoided changing, which decisions required approval, and how they handled risk.
A fictional example: an engineer discovers confusing deployment instructions. One response is to rewrite the production process alone. Another is to reproduce the problem, involve the owner, improve the instructions, and check whether the next deployment is clearer. The second account may provide better evidence of useful initiative, depending on the actual authority involved.
Do not reward rule-breaking merely because the story ends well.
Ask about the unsuccessful attempt
A polished success story can conceal weak reasoning. Ask what the candidate tried that did not work, when they recognized it, and what evidence changed their approach.
An outcome can be disappointing even when the decisions were reasonable. Evaluate the reasoning and contribution without pretending the interview reveals everything about the event.
Write an evidence-based conclusion
Replace “very high agency” with a statement such as: “Identified an unresolved customer issue, tested a small change with the responsible team, and monitored whether it helped.”
Then state what remains unknown. If the example involved a large support team, explore how the person would work with the resources your startup has.
Avoid screening for unpaid side projects or constant availability as universal proof of initiative. People can demonstrate the relevant behaviors within ordinary responsibilities.
Use the project deep-dive guide to examine personal contribution and the scorecard guide to make the evaluation consistent.