Communication interview questions
Communication interview questions test whether a candidate can be understood — by an executive, a new teammate, and someone outside their field, in writing and out loud. The 13 below cover audience adaptation, writing, listening, and status reporting. Fluency in conversation is the easiest communication skill to fake; these questions make candidates perform the rest.
All 13 questions
Communication skills interview questions
- 01Tell me about a piece of work you had to explain to two very different audiences. What changed between the versions?
- 02Describe the most complicated thing you've had to make understandable for someone with no background in it. How did you check it landed?
- 03Walk me through the most useful document you've ever written — one people used when you weren't in the room. What made it work?
- 04Tell me about a time you realized too late that you'd misunderstood what someone was asking you for. Where did the listening break down?
- 05How do you keep people who depend on your work informed while it's still in progress? Walk me through your reporting on the last long project you ran.
Communication questions about writing with AI
- 06What do you do to an AI-drafted message before you send it, beyond checking the facts? Show me with the last one you sent.
- 07Have you used an AI tool as an editor on something you'd already written — to make it shorter or clearer? What did it improve, and where did it miss your point?
- 08Your updates and documents are increasingly read by AI summarizers before any human sees them. Has that changed how you write? Should it?
- 09Tell me about a message an AI helped you write that landed badly with the reader. How did you find out, and what changed?
Situational questions for written and verbal communication
- 10An hour before a colleague presents to leadership, you spot a real error in their deck. They're busy rehearsing. What do you do — channel, words, timing?
- 11You get five minutes in a leadership meeting to report on a project that's half on track and half in trouble. What do you say first, and what do you cut?
- 12A colleague keeps reading your written comments on their work as harsher than you intend, and their replies are getting defensive. What do you change?
- 13A non-technical executive asks why your project needs three more weeks — in one short email, today. Talk me through what you'd write.
Questions 1–5
Communication skills interview questions
Tell me about a piece of work you had to explain to two very different audiences. What changed between the versions?
A strong answer changes what each audience hears, not just the wording — the executive gets the decision and the cost, the specialist gets the mechanism — while both versions stay true to the same facts.
What to look for
- Starts from what each audience needs to decide, not from the content itself.
- Cut real detail for one audience and can say what was cut and why.
- Both versions carry the same facts — simplifying never became spinning.
Example answerRed flags
Example answer
I ran a pricing migration. Finance got a two-page memo: revenue risk, rollback plan, dates. Support got a one-screen cheat sheet of what customers would see and the three sentences to say about it. Same migration, same facts. The version I never sent anyone was the full project doc — nobody needed it.
Red flags
- One explanation for everyone; the audience is expected to adapt to them.
- Simplifying slides into overselling — the easy version quietly drops the risks.
Describe the most complicated thing you've had to make understandable for someone with no background in it. How did you check it landed?
Strong answers pick a genuinely hard topic, build the explanation on something the listener already knows, and verify understanding with a question or a task — never by asking 'does that make sense?'
What to look for
- Anchored the explanation in something the listener already understood.
- Verified understanding by what the listener did or asked next, not by the nod.
- Knows what they deliberately left out and can defend the cut.
Example answerRed flags
Example answer
Our lawyer needed to understand consent flows in our data pipeline for a regulator response. I drew it as mail forwarding — who opens what, who copies whom — then asked her to walk one edge case back to me. She caught a real gap in my analogy, which told me she genuinely had it.
Red flags
- The 'complex thing' turns out to be simple, which says something about their day-to-day.
- Checked understanding by asking 'make sense?' and accepting the nod.
Walk me through the most useful document you've ever written — one people used when you weren't in the room. What made it work?
Look for a document with a named audience and a job to do: it answered the reader's questions in the reader's order, worked without the author present, and got maintained or copied by others.
What to look for
- Wrote for the reader's questions in the reader's order, not the author's process.
- Evidence it worked alone: fewer interruptions, new hires using it, teams copying its shape.
- Rewrote or cut it based on how people actually read it.
Example answerRed flags
Example answer
A runbook for billing incidents. Version one was chronological — how the system works. Nobody used it. I rewrote it around the three questions people ask at 2 a.m.: what broke, who do I page, what do I tell the account teams. Engineers I've never met still follow it, and support copied the format.
Red flags
- Their proudest document is long — length offered as evidence of quality.
- No idea whether anyone used it after it shipped.
Tell me about a time you realized too late that you'd misunderstood what someone was asking you for. Where did the listening break down?
The strong answer locates the breakdown in their own listening — an assumption filled the gap where a question belonged — and ends with a checking habit that exists today, like restating the request before starting.
What to look for
- Owns the miss instead of blaming the other person's phrasing.
- Can name the exact assumption that substituted for a question.
- A concrete habit now: restates requests, confirms deadline and purpose up front.
Example answerRed flags
Example answer
A department head asked for 'a quick look at the churn numbers'. I delivered a full analysis a week later; she'd needed one number for a board meeting the next morning and got it elsewhere. I'd heard the topic and skipped the deadline. Now I ask when it's needed and what decision it feeds, every time.
Red flags
- Every misunderstanding in their stories is the other person's fault for being unclear.
- No habit changed — the identical miss could happen tomorrow.
How do you keep people who depend on your work informed while it's still in progress? Walk me through your reporting on the last long project you ran.
Strong candidates report on a rhythm the audience can rely on, lead with status against the promise — on track, at risk, slipped — and send bad news at the same cadence as good news.
What to look for
- A regular rhythm that survived the busy weeks, not updates only when asked.
- Leads with status against the commitment, then detail for whoever wants it.
- At-risk items appeared in updates before they became misses.
Example answerRed flags
Example answer
For a six-month data migration I sent the same four-line note every Friday: green or amber, what changed, what I needed, next milestone. The format never changed, so people read it in ten seconds. When week nine went amber, nobody panicked — they'd watched that risk appear in the note two weeks earlier.
Red flags
- Reports only when asked, or only when the news is good.
- Updates are activity logs — long lists of busyness with no status against the promise.
Questions 6–9
Communication questions about writing with AI
2026 · AIWhat do you do to an AI-drafted message before you send it, beyond checking the facts? Show me with the last one you sent.
Strong answers go past fact-checking to fit: cutting generic padding, restoring their own word choices, and rewriting whatever this specific reader would stumble on. The edit for tone and audience is the real work.
What to look for
- Edits for a specific reader, not just for correctness.
- Names the tells they strip — filler openers, hedging, uniform paragraphs.
- Some messages skip AI entirely, and they can say which ones and why.
Example answerRed flags
Example answer
Last week: a note asking the platform team to take over a flaky integration. The draft was polite and endless. I cut four paragraphs to five sentences, moved the ask into line one, and deleted 'I hope this finds you well'. The facts were the model's; the bluntness was mine. That team says yes to blunt.
Red flags
- Sends AI drafts after a fact check only — tone is whatever the model chose.
- Claims colleagues can't tell the difference, without ever having asked one.
Have you used an AI tool as an editor on something you'd already written — to make it shorter or clearer? What did it improve, and where did it miss your point?
A strong answer treats the model as a critic, not a ghostwriter: it names a real improvement the tool found and a specific spot where 'clearer' changed the meaning — and the candidate kept editorial control of both.
What to look for
- Fed the tool their own draft and kept the argument theirs.
- Names a specific suggestion they accepted and why it was right.
- Caught a place where the rewrite altered the meaning, and rejected it.
Example answerRed flags
Example answer
I asked a model to tighten a proposal that had crept to three pages. It found two paragraphs saying the same thing — fair, I cut them. But it also 'simplified' my recommendation from 'pilot with one team' to 'roll out gradually', which is a different proposal. I took the cuts and kept my sentence.
Red flags
- Accepts the tool's rewrite wholesale — shorter is assumed to be better.
- Has never fed the tool their own writing and can't say why not.
Your updates and documents are increasingly read by AI summarizers before any human sees them. Has that changed how you write? Should it?
There's no settled answer; you're testing judgment. Strong candidates have noticed the shift and write so the load-bearing point survives summarization — conclusion first, one point per paragraph — without writing for the machine at the reader's expense.
What to look for
- Has actually noticed summarizers sitting between their writing and their readers.
- Puts conclusions and asks where a summary can't drop them.
- Still writes for the human — no bullet soup aimed at the machine.
Example answerRed flags
Example answer
It has. Our director reads AI summaries of project updates, and one summary dropped my only caveat — the launch date depended on a vendor. Since then the first line of every update is the thing I can't afford to lose, and each risk gets its own paragraph. Summarizers keep first lines; they eat buried clauses.
Red flags
- Hasn't thought about it and shows no curiosity when asked.
- Optimizes for the summarizer at the reader's expense — fragments with no argument.
Tell me about a message an AI helped you write that landed badly with the reader. How did you find out, and what changed?
Strong answers name a real misfire — wrong register, generic warmth the reader noticed — show how the feedback reached them, and end with a boundary: certain kinds of messages the model no longer touches.
What to look for
- A specific misfire with a specific reader, honestly told.
- Found out by asking or observing, not by accident months later.
- A boundary exists now — they can name the message types they hand-write.
Example answerRed flags
Example answer
I let a model polish a note thanking a teammate who'd covered my on-call week. He said, half joking, that it read like an HR template — and he was right, the warmth was generic. The fix wasn't a better prompt. Any message whose whole point is that I mean it, I write myself.
Red flags
- Nothing has ever landed badly — which usually means feedback never reached them.
- The fix was purely technical: better prompts, same outsourced judgment.
Questions 10–13
Situational questions for written and verbal communication
An hour before a colleague presents to leadership, you spot a real error in their deck. They're busy rehearsing. What do you do — channel, words, timing?
This tests delivery mechanics under time pressure: interrupt now, privately, lead with the error and a fix, and leave the colleague in control of the correction. Waiting to be kind fails them.
What to look for
- Interrupts immediately and privately — an hour is a gift, not a buffer.
- Opens with the error and a suggested fix, not with softening preamble.
- Leaves the presenter in charge of the correction; offers help, doesn't take over.
Example answerRed flags
Example answer
I'd walk over or call — not message — and say it in one line: 'Slide 9 doubles the renewal number; here's the correct figure.' Then stop talking. Rehearsal time is theirs. If they want, I'll fix the slide while they keep rehearsing. What I wouldn't do is sit on it to spare their nerves.
Red flags
- Softens so much the error gets lost — the colleague walks in unwarned.
- Goes to the colleague's manager first, or edits the deck without telling them.
You get five minutes in a leadership meeting to report on a project that's half on track and half in trouble. What do you say first, and what do you cut?
Strong answers open with the overall verdict and the one decision they need from the room, keep the trouble specific, and cut the methodology and the journey. Five minutes is a headline plus one risk, not a tour.
What to look for
- The first sentence carries the verdict — nobody waits four minutes to learn if things are fine.
- Names what gets cut without being prompted: process, history, the parts going well.
- Turns the trouble into a decision or an ask, not just a confession.
Example answerRed flags
Example answer
First sentence: 'Launch holds for March, but the integration workstream will miss unless we drop one of two features — I need that call today.' Then two minutes on the options and their costs. What I cut: everything on track gets one sentence, and nobody hears how the sausage was made unless they ask.
Red flags
- Builds up chronologically and reaches the bad news at minute four.
- Spends the five minutes looking good — the trouble surfaces later, from someone else.
A colleague keeps reading your written comments on their work as harsher than you intend, and their replies are getting defensive. What do you change?
The strong answer changes their own writing first — intent stated up front, questions instead of verdicts — and moves the recurring tension to a live conversation rather than typing ever more carefully at each other.
What to look for
- Owns the readability of their own tone instead of diagnosing the colleague as oversensitive.
- Concrete changes in the writing: intent up front, questions over verdicts, credit where it's real.
- Takes the pattern to a live conversation instead of another, softer comment.
Example answerRed flags
Example answer
I'd talk to them once, live: 'My comments keep landing harder than I mean them — what would make them easier to work with?' Then change two things in writing: open with what the work gets right, and phrase judgment calls as questions. Written words carry no face or voice; the reader supplies the tone.
Red flags
- Concludes the colleague needs thicker skin and changes nothing.
- Stops giving real feedback at all — harmony wins over honesty.
A non-technical executive asks why your project needs three more weeks — in one short email, today. Talk me through what you'd write.
One artifact tests translation and candor together: strong candidates give the reason as plain cause and effect, say whether it was estimation error or new information, and end with what the delay buys.
What to look for
- The reason survives without jargon — a mechanism the executive could repeat to a peer.
- States plainly whether this was an estimation miss or genuinely new information.
- Ends with what the three weeks buy, and any decision still open to the executive.
Example answerRed flags
Example answer
Three sentences. 'The customer-data import handles far more formats than we scoped, and fixing it properly takes three weeks. That's my estimation miss, not new requirements. Shipping on the original date means hand-processing imports for our first 200 customers; the delay removes that.' Executives forward emails like this, so I write it to be forwarded.
Red flags
- The explanation is jargon fog the executive can neither act on nor repeat.
- Blames upstream teams, or buries the actual reason mid-paragraph.
Scoring rubric
| Score | Evidence anchor |
|---|---|
| 1 | Fluent in conversation and nothing else checks out: explanations stay abstract, no document or story survives a follow-up, and every miscommunication in their examples was someone else's fault. |
| 2 | Communicates adequately one-to-one but shows no adaptation — the same register for every audience, updates only when asked, writing described by the effort it took rather than what readers did with it. |
| 3 | Clear speaker and competent writer with at least one real story of adapting to an audience. Listening and reporting habits exist but are recent or inconsistent; written tone still misfires now and then. |
| 4 | Adapts deliberately by audience, reports on a rhythm where bad news arrives early, and edits AI-drafted text for reader and tone. Owns past misunderstandings and names the checking habit each one produced. |
| 5 | Communication is engineered: documents that work without them, updates people quote, complex topics that come out simple without becoming wrong. You end the interview able to repeat their explanations yourself — the surest sign. |
Frequently asked questions
What are good interview questions about communication?
The reliable ones make the candidate perform the skill: explain something complex on the spot, describe a document other people used, deliver a five-minute status report. Questions that let candidates narrate their communication style mostly measure confidence.
What do communication skills interview questions actually predict?
Asked well, they predict whether stakeholders hear about risk early, whether documents work without the author in the room, and whether messages land as intended. Asked badly, they predict interview polish and nothing else.
What are good written communication interview questions?
Ask for artifacts and habits: the most useful document they've written, how they open a status update, what they change in an AI draft before sending it. A short pre-interview writing sample beats any spoken description of writing.
How do verbal communication interview questions differ from written ones?
The interview itself is a verbal sample — time-boxed explanations and live status briefings test speaking directly. Written skill can't be spoken; probe it through artifacts the candidate produced, or request a short writing sample beforehand.
Can you assess communication skills in a first-round interview?
The verbal half, yes — a candidate explaining their own work to a stranger is the test running live. Kira runs this subset by voice, giving you a scored recording before any human interview. Written skill still needs an artifact.