The four letters, and why interviewers score with them
Situation gives the context, briefly. Task names the part that belonged to you. Action walks through what you did and the reasoning behind each step. Result states what changed, preferably as a number. The structure exists because a specific event can be checked and scored. A general claim about yourself cannot.
Behavioural interviews are marked against a grid, not judged on charm. Public-sector panels, banks and graduate schemes grade each answer against a named competency, and an answer that never arrives at Action stalls in the middle band regardless of how it sounded. Rehearse with Sharp AI Interview and you will be nudged forward the moment context starts to sprawl.
This page covers storytelling only. For research, logistics and the rest of the week, work through the guide to preparing for an interview alongside it.
Tip 1 - Start from a story where something was at risk
An answer inherits its strength from the event underneath it. Choose moments with a live risk: a slipping deadline, an unhappy client, a rota nobody would cover, a finding your manager did not want to hear. When nothing could have gone wrong, there is no decision for the interviewer to examine and nothing to mark.
The stakes do not need a corporate setting. Running a till alone while the card machine failed carries more evidence than a polished group project, because the pressure was real and the choices were yours. Candidates routinely bin their strongest material for sounding unglamorous. Test both options aloud in Sharp AI Interview before deciding which to keep.
Tip 2 - Spend most of the answer on Action and Result
The common failure runs two minutes of scene-setting into a rushed conclusion. Reverse the weighting. Hold the context to fifteen seconds and the task to ten, then give the next minute to the steps: what you did, in what sequence, why that route, and what you adjusted when the first attempt fell short. A timed run makes the imbalance visible within one answer.
A quick check is to count your own verbs. Mapped, rang, rewrote, escalated: a solid Action section holds five or six of them. If the answer is mostly nouns and background, you are describing an event rather than your part in it, and the panel has nothing to credit you with.
Tip 3 - Stay visible inside a team story
“We delivered on time” carries no information about you. Give the team its credit, then draw the line around your slice: the team redesigned the rollout, and your piece was persuading the client to accept the new dates in writing. When your own contribution cannot be separated out, swap the story for one where it can.
The opposite error is just as expensive. Strip every “we” from a collaborative story and you fail the teamwork competency while apparently demonstrating it. Credit the group, define your seat in it, then narrate your own actions in the first person. That boundary is easier to find in rehearsal with Sharp AI Interview than mid-panel.
Tip 4 - Put a number on the Result, then a lesson
A figure makes the outcome concrete: queue times cut by a third, £1,800 raised in ten days, returns down 9 per cent. Where no metric exists, use a consequence instead. The account was retained. A neighbouring team adopted your checklist. The escalation was dropped. An estimate you can defend outranks an adjective every time.
Close with a single sentence on what you would keep or change. It shows reflection and answers the follow-up before it is asked. One sentence, no more. In practice sessions it is the piece most often missing from answers that otherwise sounded complete.
Tip 5 - Learn stories, not scripts
A memorised answer breaks the first time the wording shifts. “Tell me about a conflict” and “describe a disagreement you handled badly” want the same story angled differently, and a script cannot make the turn. Know six to eight stories thoroughly, then practise assembling them against questions you have not seen. The follow-ups in Sharp AI Interview are unscripted for exactly this reason.
Say the stories out loud. Silent reading finishes in 20 seconds and hides every gap; speaking takes 100 and exposes them all. If delivery is where you struggle, pair this page with the notes on handling interview nerves. Ten minutes a night for a week does more than one long session the evening before.
One answer, taken apart
Below is a standard question answered in roughly 130 words. Read it aloud against a clock; most people finish in 75 to 90 seconds. Then rebuild the shape from your own material.
“Describe a time you worked under pressure.”
S: “Last summer I was the senior front-of-house assistant at a hotel when the linen delivery for a 90-guest wedding failed on the morning of the event.” T: “The events manager was off site, so having the room ready by 2pm came down to me.” A: “I rang two nearby branches and borrowed half the linen from each, sent a taxi rather than waiting for the courier, and put one colleague on pressing while the rest of us reset the room in sections. I also told the couple's coordinator exactly what had happened and when each table would be done, so nobody had to chase us.” R: “The room was finished 20 minutes early, the couple never knew, and borrowing between branches became our standard fallback. Early, honest timelines buy you the room to fix the problem.”
Note the split: about 26 words of Situation, 17 of Task, 68 of Action, 40 of Result with the lesson attached. Each Action sentence contains a choice - two branches instead of one courier, a taxi instead of waiting, sections instead of a single sweep, telling the coordinator instead of hoping. The choices are what get marked. The word count of your own draft tells you immediately whether context is eating the answer.
The weak version of the same story: “The linen never turned up, which was stressful, but everyone mucked in and the wedding went ahead fine.” Nineteen words, no owned action, no figure, nothing an assessor can score above the bottom band. Hearing your own answer played back through Sharp AI Interview shows which of the two versions you are actually giving.
A story bank beats a question list
Prepare material, not answers. Set aside an hour, write down six to eight things that genuinely happened to you in the past three years, and give each four bullets: the context, your responsibility, three actions, one measurable outcome. Half a page per story is enough. Load the bank into Sharp AI Interview and its suggestions draw on your real experience instead of generic filler.
Then tag each story with two competencies rather than one. A good story usually stretches from teamwork into influencing, or from planning into resilience. Those overlaps are what let you cover a long panel without telling the same story twice.
Five ways good answers go wrong
All five appear constantly in mock interviews, and all five can be fixed in a single practice run.
- Drifting into the hypothetical. “What I would normally do” cannot be scored. What you did on one specific occasion can.
- Skipping the Task. Without a stated responsibility, your contribution reads smaller than it was.
- Burying the Result. “It went well” hides the 30 per cent improvement that would have carried the answer.
- Reaching too far back. Prefer the last two years. An old example needs to be clearly stronger to earn its place.
- Trailing off. “So, yeah…” undoes ninety good seconds. Deliver the final line, then stop talking.
Delivery carries its own weight, particularly on video, where posture registers before words do. The guide to confident body language covers that side of the answer.
STAR across formats
A phone screen wants the compressed cut: Situation and Task in one breath, two actions, the number. A competency panel wants the full version with the reasoning left in. One-way video punishes rambling hardest of all, so rehearse against a visible clock. The rundown of interview types shows what each format rewards, and a session in Sharp AI Interview can be set to match the one you are facing.
At assessment centres the structure surfaces differently - in how you summarise a group's progress - though the interview slot within the day is pure STAR. The transferable skill is compression: the same story told at 60, 90 and 120 seconds. Practising all three lengths takes about fifteen minutes.
When the interviewer interrupts
Interruptions signal interest. “Why that route?”, “how did your manager react?” and “what would you do differently?” are where scripted candidates unravel. Answer the question fully, then bridge back with a set phrase: “so, coming back to the event itself.” Rehearse the bridge until it needs no thought. Unannounced interruptions are part of every mock in Sharp AI Interview so that the real ones stop registering as threats.
The hardest probe is the negative one: “what went wrong along the way?” Keep one honest flaw per story, framed as something you noticed and corrected. The nightly checks ran long at first; you cut them to fifteen minutes after two evenings. Practice surfaces those soft spots before a panel finds them for you.
How much rehearsal is enough
Three or four 20-minute sessions spread across the week outperform one long evening. First session: each story aloud, timed. Second: questions shuffled, so the choosing happens under pressure. Third: interruptions on. On the day itself, run your two best stories once and leave it there. Track only two things - the length of each answer and whether the Result carried a number.
Questions candidates ask about STAR
What does STAR stand for in an interview?
Situation, Task, Action, Result. Two sentences of context, one sentence on what you personally owned, the steps you took with the reasoning behind them, then the outcome with a figure attached. Panels rely on the structure because it turns a claim into evidence they can score.
How long should a STAR answer be?
Between 90 seconds and two minutes of speech, weighted toward Action: brief context, one line of task, a full minute on what you did, and a closing result. Anything under a minute rarely contains enough to mark. Anything past three minutes invites an interruption.
What should I do if I have no example for the question asked?
Do not make one up. Take a moment, then offer the closest real parallel from study, volunteering, part-time work or a side project, and name the parallel openly. Interviewers accept a transferred example without fuss. An invented one collapses under the second follow-up.
How many STAR stories should I prepare before an interview?
Prepare six to eight. Between them they should cover teamwork, conflict, deadlines, a mistake you recovered, leading without authority, problem solving, change and difficult customers. Below six you leave gaps. Above ten you will not know any of them well.