Candidates prepare for a panel as though it were a harder version of a one-to-one interview. It is a different exercise. In most structured processes each panellist is assigned a competency area and fills in their own scorecard, and the hiring decision is made afterwards in a discussion you are not in.

That changes what you are optimising for. You are not trying to win the room. You are trying to give four people enough material each to defend a yes when they are asked to.

Who is in the room and what they are scoring

Composition varies, but panels usually contain some combination of these roles. Work out which is which in the first five minutes.

  • The hiring manager. Scoring whether you can do the job and whether they want to manage you. Usually asks about scope, judgement and how you handle being wrong. This is the vote that carries the most weight.
  • The peer or future teammate. Scoring whether working with you would be easy. Often asks the most practical, day-to-day questions. Their veto is quiet but real — teams that dislike a candidate usually get their way.
  • The technical or functional assessor. Scoring depth. Will go further than the others into detail and is not being unfriendly by doing so.
  • The skip-level or senior stakeholder. Scoring trajectory and whether you would survive contact with the wider organisation. Asks broader, vaguer questions and listens for how you structure an answer rather than its content.
  • The HR or talent partner. Scoring consistency, motivation and any risk in the process — notice period, expectations, references. Often chairs and keeps time.

If nobody introduces their role, ask at the start: "Could I ask what each of you works on?" It takes ninety seconds, it is entirely normal, and it tells you how to pitch every subsequent answer.

Eye contact and addressing the group

The mechanic is simple and most people get it wrong in one of two directions: answering only the person who asked, or spraying eye contact around the table like a public speaker.

The pattern that works is to begin with the questioner, move through the others during the middle of the answer, and return to the questioner for the conclusion. Roughly two-thirds of the answer to the person who asked it, the rest distributed.

Use names. Panels are long and panellists notice when you have remembered who they are. Write the names down as they introduce themselves, in seating order.

On a video panel, the mechanics change: look at the camera rather than the gallery for the main points, and use names more deliberately, because the visual cues that carry a group answer in a room do not survive the grid.

The silent panellist

There is usually one. They ask nothing for forty minutes and it is unnerving.

Almost always it is benign: they are taking notes, they are there to observe, they are the second reviewer, or their assigned competency comes later. Occasionally they are the most senior person and have chosen to watch.

Do not treat silence as hostility and do not over-correct by directing everything at them. Include them in the sweep of the answer as you would anyone else, and give them one clear opening near the end: "Is there anything from your side I have not covered?" That question converts a silent observer into someone with something to write down.

The technical questioner who seems hostile

The most common misread in panel interviews. Someone starts pressing on detail, follows your answer with another harder question, and does not soften it with warmth.

In the large majority of cases this is not hostility. It is how technical assessment works: they push until they find the edge of your knowledge, because the edge is the information they need. A candidate who is never pushed to the boundary cannot be scored.

So the correct response to "how would you handle it if the data was incomplete?" is not defensiveness, and certainly not bluffing. It is:

  1. Answer directly, at the level of detail asked for.
  2. When you reach the limit of what you know, say so in one sentence and then say what you would do about it. "I have not implemented that myself. What I would do is X, and I would check Y with someone who has."
  3. Do not pad. Panels are timed, and a long answer to a narrow question costs a different panellist their question.

Admitting the boundary clearly is scored well by almost every technical assessor. Bluffing is scored badly by all of them, and they can tell.

What the panel is actually doing

Each person fills in a separate scorecard on a separate competency, then argues it in a debrief you are not present for. Your job is to hand every panellist one concrete, memorable example in their own area — the thing they can quote in that room. Four specific stories beat one impressive performance.

A few mechanics that pay for themselves

  • Ask each panellist a different question at the end. One question each, tailored to their role, beats three questions aimed at the chair. It also tells you things.
  • Reuse examples deliberately, not accidentally. If two panellists ask about the same project, cut it differently: the delivery angle for one, the stakeholder angle for the other. Saying "as I mentioned earlier" and stopping wastes your best material.
  • Watch the clock for them. If the chair is visibly managing time, shorten your answers. Panellists remember candidates who made their schedule work.
  • Note who cared about what. Two words per panellist, written down immediately afterwards.
A panel decision is made in a debrief. Give each person a sentence they can repeat in it.

Follow-up afterwards

Send one email, to the chair or the recruiter, within 24 hours. Ask them to pass on thanks to the panel by name. Individually emailing four people you have just met reads as heavy-handed in most markets, and in the UK especially; in parts of the US, individual notes are more conventional, so take the local steer if you have one.

Keep it to four or five sentences. Reference one specific thing discussed, and if there was a question you answered poorly, add two sentences of a better answer. That is the only legitimate use of a follow-up email to change a score, and it occasionally works.