Chief of Staff Interview Questions

I earn it through preparation, not title. In my first weeks in any Chief of Staff role, I sit down with each executive one on one and ask what keeps them up at night, not what their function does on paper. That gives me real problems to solve early, which is worth more than any org chart position. I also make sure the first few things I touch are done well and quietly: a board deck that goes out with zero last-minute fire drills, a meeting that actually ends with decisions instead of a re-run next week. Executives start routing harder problems to you once they have seen you handle the small ones without drama. Credibility here is compounding: it is built one reliable delivery at a time, and it can be lost in a single mishandled confidential conversation.

I use a simple filter: does this decision need the CEO's judgement specifically, or just their awareness? Judgement items are things with irreversible consequences, meaningful ambiguity, or a need for external relationships only the CEO holds, like a key investor call or a leadership hire. Awareness items get a line in a weekly digest instead of a meeting slot. I also weigh urgency against reversibility: a reversible decision that is urgent can usually be delegated with a clear boundary, while an irreversible one gets escalated even if it feels minor. In practice this means I intercept a lot of requests that would otherwise land directly in the CEO's inbox, triage them, and only pass through what actually needs their signature. I check in monthly on whether my filter is calibrated correctly, because it drifts as the business changes.

I design the review around decisions, not status updates. Before the meeting I send a one-page pre-read with the three or four metrics that matter, flagged red, amber, or green against target, plus a short list of the decisions we need the room to make. That pre-read has to go out at least 24 hours ahead, or the meeting degrades into people reading slides live. In the room, I keep status reporting to five minutes total and spend the rest on the flagged items and open decisions. I close every review with a recap of what was decided, who owns the next action, and the date it is due, sent out within the hour while it is still fresh. When I ran this format at a 200-person company, we cut our monthly business review from 90 minutes to 45 and the follow-through rate on action items roughly doubled because ownership was unambiguous.

AI has become part of how I handle the volume of information that moves through this role. I use it to summarise long email threads and board materials before meetings, which saves me real time when I am prepping three executives for the same session with different levels of context. I also use it to draft first passes of internal memos and meeting recaps, which I then edit heavily for tone and accuracy, since a memo that sounds wrong coming from the CEO's office causes more damage than it saves in time. For synthesising themes across many pieces of qualitative feedback, like exit interviews or a round of skip-level conversations, it is noticeably faster than reading everything manually and still catches patterns I might miss. What I never hand to AI is judgement calls involving confidential personnel matters or anything where the wording itself carries political weight: those need my full attention and cannot be delegated to a tool that does not know the relationships in the room.

During a restructuring at my previous company, I was read into a planned leadership change two weeks before it was announced, while continuing to run normal weekly meetings with the exact people who would be affected. I kept a strict mental separation between what I knew and what I could act on: I did not change my behaviour toward anyone, did not hint at what was coming, and declined a colleague's direct question about whether something was happening by saying plainly that I could not discuss it rather than lying outright. In parallel I worked with HR and the CEO to prepare the communication plan, timeline, and a Q&A document for managers so the announcement itself would go smoothly. When it landed, there were no leaks and no one felt blindsided by how it was handled, even those who were unhappy with the decision itself. The lesson I carry forward is that discretion means being clear that some things are not yours to share yet, not deception.

I was asked to lead the consolidation of three separate customer onboarding processes into one, across sales, product, and support, none of whom reported to me and all of whom had reasons to prefer their own version. I started by mapping each team's actual constraints rather than assuming they were just being territorial: support cared about ticket volume, sales cared about time to first value, product cared about data consistency. I built a proposal that addressed each concern explicitly and got each functional lead to co-author a section of the final plan, so it was never presented as my project imposed on them. I ran a fortnightly working session to keep momentum and used the CEO sponsorship I had, sparingly, only twice when a team lead pushed back beyond the point of legitimate disagreement. We launched the unified process in nine weeks, and onboarding time dropped by 30 percent in the following quarter.

I once deprioritised a request from our Head of Sales to build a pricing exception tracker, judging it as a nice-to-have against a heavier quarter of board prep and a hiring push. Six weeks later it turned out three separate deals had stalled because reps had no visibility into which exceptions were already approved, costing us roughly two weeks of extra negotiation time on each. I owned the miss directly with the Head of Sales rather than explaining it away, walked through the reasoning that led me there, and asked what signal I had missed. It turned out I had not asked enough people outside my usual check-ins before making the call. I rebuilt the tracker within a week using a simple shared spreadsheet as a stopgap, and I added a standing monthly touchpoint with functional leads specifically to surface requests I might otherwise deprioritise on limited information.

I would start by mapping the decisions the business actually needs to make on a regular cadence, then design meetings around those decisions rather than around organisational habit. Typically that means a weekly leadership sync focused on the two or three metrics that matter most that week, a monthly business review that looks at trends and makes resourcing calls, and a quarterly strategy session tied to OKRs or whatever planning framework the company uses. Each layer needs a clear owner, a standing agenda template, and a documented decision log so nothing gets re-litigated from scratch each time. I would also build in a lighter-weight async layer, a written weekly update from each function, so the in-person time is reserved for actual discussion rather than status reads. I roll this out gradually, piloting with one or two teams before company-wide adoption, because an operating rhythm imposed all at once usually gets ignored within a month.

I start at least two weeks out by aligning with the CEO on the three messages the board absolutely needs to leave the room believing, then work backward from there to build the deck and supporting data. I run at least one full dry run where I play a difficult board member and push on the numbers, because the first time an exec hears a hard question should never be in the actual room. I pre-brief individual board members ahead of contentious items where possible, so there are no unpleasant surprises live, and I prepare a one-page appendix of backup data for anything I expect to be challenged. On the day, I sit in the room to track action items and unresolved questions in real time so the follow-up email goes out within hours, not days. The measure of a good board prep is that the meeting itself feels almost uneventful because everything difficult was already surfaced beforehand.

I keep a single source of truth, usually a shared tracker in Notion, Asana, or a simple spreadsheet depending on what the company already uses, that lists every active initiative with an owner, a target date, and a current status. I resist the temptation to build something elaborate: the tool matters far less than the discipline of updating it weekly and using it as the actual reference in meetings rather than as documentation nobody opens. For recurring commitments made in meetings, I maintain a decision log tied to each session so we can trace back why something was decided, which becomes invaluable when priorities get questioned months later. I also use calendar analysis periodically, literally auditing where the CEO's time actually goes over a month, because stated priorities and calendar reality often diverge and that gap is one of the clearest signals I have that something needs to change.