The interview presentation is the UK final round's favourite instrument, and most candidates prepare the wrong half of it. The slides are the half you control at home; the assessment mostly happens in the other half — the brief you decoded (or didn't), the slot discipline you showed, and above all the Q&A afterwards, where the panel finds out whether the person matches the deck. Panels routinely score a middling presentation with a strong defence above a beautiful one that collapsed under the first hard question.
This page covers the whole task in assessment order: what the brief is really asking, a structure that fits a ten-minute slot without sprinting, delivery mechanics that survive nerves — and four fully worked Q&A moments, marked against the four criteria aurate uses in live sessions, because the Q&A is where the decision actually gets made.
One scope note: if your presentation sits inside a longer multi-exercise day, our assessment centre guide covers that format's own rules — this page owns the standalone presentation task and its aftermath.
Practise these questions out loud with a live AI interviewer — free. Start with Try Me
How the marking guidance works
Each model answer below is marked against the four criteria a live aurate session scores:
See how a full session is scored
aurate is a practice tool. Marking guidance describes what strong practice answers show — it isn't an employability assessment.
Why it's asked: Some panels ask it aloud; every panel asks it silently. The question tests whether you treated the brief as a task or as a judgement exercise — and saying your interpretation out loud, assumptions included, converts a guessing game into a demonstration of how you'd scope real work.
Gladly — because the brief made me choose, and I'd rather show you the choice than hide it. The task was 'present how you would improve operational efficiency across the network' — ten minutes, no other constraints. That could honestly fill a week, so I scoped it three ways.
First, I took 'efficiency' to mean cost-per-mile and lost-mileage, not headcount — because in bus operations those are the levers an operations manager actually owns, and headcount belongs to a different conversation with different people in the room.
Second, I anchored on what I could evidence: your published punctuality data and the two depot visits I could make as a member of the public. Where I'm guessing, the slide says 'assumption' on it — three of them, and I'd genuinely welcome being corrected on any of them today.
Third, I set aside anything needing capital sign-off. A first-six-months plan built on new vehicles is a plan for someone else's authority.
So: ten minutes on the three operational levers I believe you can pull this year, from the seat you're hiring for. If I've scoped it wrongly, that conversation is more useful to both of us than the slides — shall I go on?
Marking guide
Why it's asked: The classic post-presentation stress test. The panel isn't renegotiating your plan; they're watching whether you prioritise on the spot with visible reasoning, or defend everything and thereby defend nothing. It also exposes whether your original plan had a spine — a best version knows what it would cut.
Then I'd keep the pricing work and the range review, and I'd cut the store-refit pilot — and I can tell you why in that order.
The pricing work keeps its place because it's the only part of the plan that's self-funding inside the year: the elasticity review on our top 200 lines costs analyst time we already have, and in my current business the equivalent exercise paid for itself in fourteen weeks. Cutting it to save money would be cutting the part that makes money.
The range review stays because half its value is stopping spend — delisting the tail that ties up warehouse space and working capital. Halving my budget makes that MORE urgent, not less.
The refit pilot goes, and I want to be honest that it hurts: it was the part of the plan the branch teams would have felt first, and morale is a real currency. But it's the only element where the return arrives in year two, and a halved budget is the business telling me it needs returns in year one. I'd bank the design work we'd already done and bring it back in the next planning round.
What I wouldn't do is shave everything by half — three underfunded initiatives deliver less than two funded ones. If the budget moves again, the same logic reorders the same list.
Marking guide
The deck is rehearsable at home — the Q&A needs an opponent
You've just read two defence moments that decide presentations. Rehearsing them against a live interviewer who pushes back — budget halved, figures challenged, time cut — is what an aurate session does, and it marks the defence on the same four criteria used across this page. Two free sessions. No credit card.
Try it freeWhy it's asked: Sometimes genuine, sometimes engineered; either way it's a prioritisation exam with an audience. Panels watch whether you visibly triage — headline, evidence, recommendation — or try to compress everything and deliver nothing. A candidate who handles this gracefully proves the presentation was understood, not memorised.
Of course — headline first, then the one number that matters, then my recommendation, and the detail can live in your packs.
The headline: of the three depot-consolidation options in the paper, only option two survives contact with the constraints you've set — option one breaches the franchise commitments on the eastern routes, and option three's savings evaporate once you price the staff relocation properly.
The number: option two releases £2.3 million a year from year two, and — the part I'd flag hardest — it's the only option that doesn't touch the apprentice training centre, which the committee has publicly committed to twice.
So my recommendation is option two, phased over eighteen months, with the detailed phasing on pages six to nine when you have time for them.
And one honest caveat I'd normally have given five minutes to: the £2.3 million assumes the lease surrender lands as negotiated — if it doesn't, the figure drops materially, and page eleven shows both cases. I'd rather you hear the caveat from me in three minutes than find it on page eleven next week.
That's the three-minute version. I'm happy to take questions now or give you the fuller picture another time — your call.
Marking guide
Why it's asked: Whether or not the panel's claim is even true, the moment tests composure and epistemics: do you defend stale data, capitulate instantly, or handle the correction like a professional who works with imperfect information every week? The recovery matters more than the figure.
Thank you — then I'd genuinely like to see them afterwards, and let me tell you what moves and what doesn't if you're right.
The figure on slide three is the 82 sites' average occupancy from your last published annual report — the most recent number available from outside the business, which is the constraint every external candidate works under. If your management accounts show it higher, that's good news and it changes my numbers in one direction: the revenue-per-unit workstream gets MORE valuable, because yield management matters most when you're nearly full.
What doesn't change is the shape of the plan. The recommendation was never built on the level of the number — it's built on the spread: your best quartile of sites running well ahead of your worst on the same rates, which is a pattern newer data almost never erases, only re-sizes.
So the honest position: my figure is the best public number, yours is better, and the plan was designed to survive exactly that upgrade. If the spread HAS closed since the report — that would genuinely change my priorities, and I'd want to understand what closed it, because that's the case study I'd be learning from in week one.
What I'd take from this exchange for the role itself: I'd rather be corrected fast than defended long. That trade is the job.
Marking guide
Why it's asked: The reliable skeleton: one minute of map ('the brief, how I read it, three parts'), seven minutes of content in descending importance — recommendation early, not saved for a finale a time squeeze can steal — and a ninety-second close that repeats the recommendation and hands the panel their first question. Nerves eat transitions first, so script the signposts verbatim even if nothing else is scripted.
Why it's asked: One slide per minute is the outer limit; six to eight for a ten-minute slot is the working norm. Slides carry what speech can't — one chart, one number, five words of headline — and never paragraphs, because the panel reads instead of listening the moment text appears. If the tech fails, the presentation should survive: panels quietly rate the candidate who presents cleanly off a printed page.
Why it's asked: Bringing notes is fine everywhere that matters; reading them aloud is what fails. The working method: cue cards with signposts and numbers only — the things nerves actually steal — glanced at during transitions, not mid-sentence. Practising aloud four times beats a fifth rewrite of the slides; fluency lives in the mouth, not the deck.
Why it's asked: Pitch the content at the most senior decision-maker's altitude — decisions and trade-offs, not task detail — while distributing eye contact across everyone, because panels compare notes on exactly this. The specialist in the room gets one deliberate moment of depth ('the phasing detail matters here, and I'm happy to go deeper in questions') so expertise is proven without hijacking the slot.
Why it's asked: The strongest available move is the bounded honest one: what you do know, what you'd need to check, and how you'd find out — delivered without apology spiral. Panels plant at least one unanswerable question deliberately; they're buying your behaviour at the edge of your knowledge, because that edge is where the real job lives.
For a ten-minute final-round task: roughly a day of thinking and research, half a day of building, and — the part most candidates skip — four full spoken run-throughs, at least one to another human. Time in the mouth beats time in the slides; the fifth rewrite of the deck buys almost nothing.
Six to eight, treating one per minute as a ceiling rather than a target. Each slide carries a headline, one exhibit and nothing the panel could read faster than you can say. If a slide exists to remind you what to say, it belongs on your cue card instead.
Yes — no reasonable panel penalises notes; they penalise reading. Use cue cards carrying signposts and numbers, glance during transitions, and keep your eyes on the room mid-argument. A candidate visibly in command of notes reads as prepared; a candidate buried in them reads as absent.
Bound it honestly: state what you do know, name what you'd need to check, and say how you'd find out — then stop, without apologising twice. Panels often plant one unanswerable question on purpose; the assessment is your composure at the boundary, not the missing fact.
Only if asked. Unrequested decks invite the panel to read ahead and pre-form the Q&A without you in the room. If they do request slides in advance, design for it: fuller notes pages, and keep one insight deliberately for the room so the presentation still adds something the PDF didn't.
Two free sessions. No credit card.