CS Interview Coach
Start Free

Behaviours

Working Together: Civil Service Examples (Why Harmony Isn't Evidence)

Published

"We all got on really well and pulled together" is the sound of a Working Together answer scoring a 3. Not because the team wasn't lovely, but because harmony is unattributable — a panel scoring you against descriptors cannot award marks to an atmosphere. The behaviour's name misleads candidates into telling stories about teams that worked; the descriptors reward something narrower and more demanding: relationships you built, across some kind of gap — a team that didn't trust yours, a stakeholder you'd never met, a colleague you clashed with — plus one element almost nobody prepares: wellbeing, yours and others', which sits in the official descriptors from EO upward and goes unclaimed in nearly every interview.

What panels are actually scoring

A relationship with a "before" and "after". The strongest Working Together evidence starts with distance or friction — teams that worked around each other, a partner org that saw yours as a nuisance, a colleague relationship that had gone cold — and shows the specific things you did that changed it. No gap crossed, no evidence.

Reciprocity. The descriptors say sharing information, resources and support — two-way. Candidates reliably tell one-directional stories (I helped them, I kept them informed). The mark is in the exchange: what you gave, what you asked for, what came back.

Perspective genuinely taken. Not "I listened to their concerns" — a moment where the other side's view changed what you did. If their perspective appears in your story but your approach is identical afterwards, the listening was decorative.

Wellbeing as an act. Noticing a colleague drowning and doing something proportionate about it; protecting your own capacity honestly rather than heroically. It's explicit descriptor language and it's nearly free marks, because so few candidates claim it.

The questions panels actually ask

  • "Tell me about a time you built a relationship with another team or organisation."
  • "Describe a time you worked with someone whose working style was very different from yours."
  • "Give an example of overcoming difficulties in a working relationship."
  • "Tell me about a time you supported a colleague who was struggling."
  • "How have you created an inclusive environment in a team you've worked in?"

Probing follow-ups: "What did you specifically do?", "How did the other person see it?", "What did you get back from the relationship?" — the first because harmony stories are unattributable, the second because one-sided stories are unverifiable, the third hunting for reciprocity.

Fail vs pass: the same relationship at 3 and at 6

A realistic answer about cross-team working. Same events, told twice.

FailVersion 1 — scores around 3/7

"Our team's work regularly depended on a technical team based in another office, and the relationship between the two teams wasn't great — requests would sit in a queue for weeks and there was a fair amount of frustration on both sides. When I took over as the main point of contact, I made a real effort to improve things. I was always polite and professional in my emails, I made sure our requests were clear and complete, and I built up a good rapport with my counterpart there over time. I also made a point of thanking them when work came back quickly. Gradually the relationship improved — our requests started being turned around faster, and the general tone between the teams became much friendlier. By the end of the year the two teams were working together far more smoothly, which benefited everyone."

Read it as a panel member. Perfectly pleasant, and almost entirely weather. "Made a real effort", "built up a good rapport", "the tone became friendlier" — none of it is an action a scorer can point to, and the improvement "gradually" happens with no mechanism visible. The other team exists only as a queue with moods; we never learn why requests sat for weeks, what their side of the frustration was, or whether anything the candidate did addressed it. Politeness and clear emails are baseline professional conduct, not behaviour evidence. This is the harmony trap in full: a real improvement the candidate probably did contribute to, told in a way that gives the panel nothing to score. A 3.

PassVersion 2 — the same two teams, around 6/7

"Our team depended on a technical team in another office, and the relationship was poor — our requests sat for weeks, and both sides had settled into blaming the other. When I took over as point of contact, I started by asking to spend half a day with them rather than sending a better email. That half-day changed my understanding of the problem: from their side, our requests were the issue — arriving without the reference data they needed, so each one meant a chase-and-wait loop before they could even start, and they'd learned to deprioritise us because our tickets were the high-effort ones. So the fix I'd assumed we needed — them working faster — was wrong; we needed to change what we sent. I drafted a one-page request standard with my counterpart rather than for him, so it reflected what his team actually needed, and I took it back and sold it to my own team, which was the harder half — colleagues initially saw it as extra admin imposed to make another team's life easier. I made the case with our own numbers: the chase-and-wait loop was where our three-week delays actually lived. In return, my counterpart agreed his team would flag incomplete requests within two days instead of letting them sit — a commitment I could take back to my team as their side of the bargain. A few months in, I noticed my counterpart had gone quiet and turnaround was slipping again; instead of assuming the arrangement had failed, I called him — his team had lost two people and he was drowning. I got agreement from my manager for us to temporarily handle the initial triage step ourselves, which kept the pipeline moving and, more than anything else I did, convinced his team the relationship ran both ways. Average turnaround went from over three weeks to under one, and the request standard is still in use by both teams."

What changed:

  1. The other side becomes real. The half-day visit produces the pivotal discovery: we were the problem. Perspective-taking that reverses the candidate's own diagnosis is the strongest form of "listening" there is.
  2. Actions replace atmosphere. A visit, a co-drafted standard, an internal selling job, a negotiated two-day commitment — four attributable acts where Version 1 had rapport.
  3. The harder half is owned. Persuading your own team is the part candidates always skip. Including it shows the relationship work went in both directions.
  4. Reciprocity is explicit. We change our requests; they commit to two-day flagging; later, we take on triage while they're short-staffed. Exchange, not service.
  5. The wellbeing moment. Noticing the counterpart drowning and responding practically converts the descriptor almost nobody claims into the story's turning point.
  6. A durable, numbered result. Three weeks to one, standard still in use — mechanism visible, outcome measured.

How Working Together changes by grade

  • EO: contacts developed beyond your own team, genuine interest in others, help asked for and offered — plus the wellbeing awareness point. A well-told version of a smaller-scale story scores highly. Calibration in the EO guide.
  • HEO: professional relationships with a range of stakeholders, collaboration that shares real resources, and team spirit you actively invested in. Version 2 above sits here. See the HEO guide.
  • SEO: the same descriptor band, calibrated up — partnerships across organisational boundaries, wellbeing support put in place rather than noticed, diverse input actively sought. Friction at this grade should involve stakes: delivery, money, seniors. The SEO guide and SEO hub cover it.
  • G7: networks built and maintained to achieve shared objectives, assumptions challenged while compromising where it helps progress, and care for wellbeing across a whole team. One good bilateral relationship is HEO evidence at G7 — the panel wants a web. See the G7 guide.

Common mistakes that cap your score

  1. The harmony trap. A team that got on well, described warmly. Unattributable, unscoreable.
  2. No gap crossed. If the relationship was fine at the start and fine at the end, there's no evidence in the middle.
  3. One-way traffic. All giving or all receiving. The descriptors say share — show the exchange.
  4. Decorative listening. Their concerns "taken on board" while your plan stays identical.
  5. Trashing the other side. Friction stories where the other team are simply unreasonable make panels wonder what the other team's version sounds like. Give their position its full weight — it makes crossing it more impressive.
  6. Wellbeing left unclaimed. If your story contains a struggling colleague you actually helped, say so specifically. It's in the descriptors and almost nobody scores it.

Every Working Together answer you prepare should pass one test: could the panel score you, specifically, from what you've said? If you're not sure yours does, run it through a scored mock — the AI coach marks against the same grade descriptors panels use, first three sessions free, with the full 250+ question bank on membership.

Common questions

Practise before the real thing

Reading how panels score is step one. The candidates who pass are the ones who practised saying their answers out loud — and got scored feedback against the official Civil Service Behaviours framework before interview day.

Start Practising Free

3 free practice questions · No card needed · Cancel anytime