Guide
The questions a first round actually has to answer
Most first-round question lists are built to sound thorough. A first round is not an interview. It exists to answer a small number of questions that decide whether a real interview is worth anyone booking.
Agentic BluePublished
A first round is not a small interview
An interview is trying to work out whether somebody is the right person. A first round is trying to work out whether an interview is worth booking. Those are different jobs, and most published question lists are written for the first one and then used for the second, which is why they take forty minutes and settle nothing.
Nearly every first round comes down to four questions. Have they done the thing. What do they expect to be paid. When could they start. Are they still available. A first round that answers those four reliably, for every applicant, has done its job, and one that explores motivation and culture without answering them has not.
The four that always apply
“Tell me about the part of your work that is closest to this role.”
What it decides. Whether the experience on the CV is the experience the role needs, in their words rather than a bullet point.
Where it goes wrong. Asked as 'walk me through your CV', it invites a chronology and burns five minutes. Naming the role's actual work in the question is what makes the answer comparable between candidates.
“What are you expecting for base salary, and is there a fixed allowance on top?”
What it decides. Whether there is any point continuing. More processes end here than on capability.
Where it goes wrong. Asking for 'expected salary' as one number gets a total, and a total hides the structure. Someone quoting a package that includes a variable bonus and someone quoting base can be thousands apart while appearing to say the same figure. Ask for base and fixed allowance separately or you have not actually asked.
“What notice do you have to give, and when could you realistically start?”
What it decides. Whether they can start inside the window the role needs.
Where it goes wrong. Notice period and start date are not the same question. People answer the first and you assume the second. Ask both in one breath, because a two-month notice and an agreed early release are a very different answer to a two-month notice and a project to finish.
“Are you still looking, and where else are you in a process?”
What it decides. Whether the candidate is real and how much time you have.
Where it goes wrong. Often skipped as intrusive. It is the single best predictor of whether a shortlist survives to offer, and candidates answer it readily when it is asked plainly rather than as a trap.
Writing the role-specific ones
Beyond the four, a first round can usually carry three or four questions particular to the role before it stops being a first round. The test for each one is simple and unforgiving: if every plausible candidate would answer it the same way, it is not a question, it is a formality.
Ask about a decision, not a responsibility. “Were you responsible for the rollout?” gets a yes. “What did you have to change when the rollout slipped?” gets an answer only somebody who was there can give.
Ask for a number where a number exists. Team size, ticket volume, deal size, headcount supported. Numbers are hard to inflate convincingly under a follow-up, and they make two hundred answers sortable in a way that prose is not.
Ask one question about the thing that most often goes wrong in the role. For a support hire, the angry customer. For a finance hire, the month that did not reconcile. The answers separate people faster than any question about strengths.
The questions that tell you nothing
“Why do you want to work here?” At first-round stage the honest answer is usually that the role looked relevant and the pay looked plausible. Everyone knows this, so the question tests preparation and nothing else, and it rewards the candidate who has done it thirty times.
“What is your greatest weakness?” Answered from a script by anyone who has interviewed before, and answered honestly by people who are then penalised for it. It is a filter on interview experience.
“Where do you see yourself in five years?” Rarely predictive and never comparable. If retention is the real concern, ask what they want their next role after this one to look like, which people answer straightforwardly.
Anything you will not act on. If an answer would not change whether you book the interview, it does not belong in a first round. That is the whole filter, and applying it usually removes half the list.
The follow-up is most of the value
A fixed list of questions collects fixed answers. “I managed a team” and “I managed a team of eleven across two sites” arrive in the same box on a form, and the difference between them is the entire reason you asked.
This is the part that makes a first round expensive to automate honestly and the part worth protecting. Whatever runs your first round, whether it is a person, a form or software, the question to ask of it is whether it can notice a vague answer and ask once more. If it cannot, it is a questionnaire wearing an interview's clothes.
It is also why the transcript matters more than the score. Read how to screen two hundred applicants for one role for the arithmetic that decides how many of these conversations you can actually afford.