See what's new

Testlify
AI in recruitment
Last updated on: 11 November 202618 min read

Role simulations in hiring: how to design, run and score them

Role simulations put candidates in real job situations. See the formats, AI interviewer prompts, scoring rubrics and where they fit best.

Role simulations in hiring: how to design, run and score them

A candidate can describe how they would calm an angry client and still freeze when one shows up on the call. Role simulations in hiring remove that guesswork by putting candidates inside a real job situation and scoring what they actually do.

A role simulation is a short, scored exercise where a candidate handles a realistic situation, such as an upset customer, a discovery call or a full inbox. An assessor or an AI interviewer plays the other side.

Every candidate faces the same scenario and is judged against a rubric written before anyone applies. This guide shows how to pick the format, write the prompts, score the results and keep the process fair.

TL;DR:

  • Role simulations show what a candidate can do on a real task from the role, which a resume or a "tell me about a time" question cannot.
  • Use them as one strong signal next to a skills test and a structured interview, never as the whole decision.
  • Match the format to the role: a live role-play or chat simulation for sales and support, an in-basket for managers, a work sample for technical roles.
  • Run the simulation after a cheap screen and before the panel, then use the output as the material the interview probes.
  • Write the rubric first: three to five observable behaviors per prompt, scored on a fixed scale by two reviewers working independently.
  • Keep it short. Most role simulations should take 20 to 45 minutes, and a take-home should stay under 90 minutes.
  • AI-led simulations need candidate notice, human review and a bias audit where the law asks for one, such as NYC Local Law 144.
Summarise this post with:ChatGPTGeminiClaudeGrokPerplexity

What are role simulations in hiring?

Role simulations in hiring are structured exercises that place a candidate inside a situation the job really contains and score how they handle it. Instead of asking how someone would calm an angry client, the simulation puts an angry client in front of them.

The output is observed behavior and work product, not a claim about behavior.

A role simulation can be live, recorded on video, or run by a conversational AI interviewer that plays a customer, a colleague or a stakeholder and asks follow-up questions based on what the candidate says. The prompt is the scenario or task that starts the exchange.

Two clarifications help. A role simulation is not a mock interview: a mock interview rehearses the conversation, while a simulation replaces part of the conversation with evidence.

A role simulation is also one member of a wider family of work-based assessments. For the full set of formats, from coding tasks to take-home projects, see our guide to the job simulation.

This page focuses on interactive, scenario-based exercises: role-plays, chat simulations and AI-led video prompts.

Build your dream team — Book a product demo

Why use role simulations in hiring instead of interview questions alone?

Role simulations in hiring add evidence that interviews alone cannot produce. Traditional interviews ask candidates to describe past behavior, and polished speakers do well.

A simulation asks candidates to show current capability under realistic conditions, so the strongest talker and the strongest performer are no longer the same person by default.

Hiring teams are moving this way. A 2025 report by SHRM and the SHRM Foundation, The Skills-First Movement, found that 70% of HR professionals expect skills-first practices to play a larger role in hiring decisions over the next three years, and 29% name a lack of assessment tools as a barrier.

Role simulations bring four practical benefits:

  • Evidence over claims: The hiring manager sees how a candidate handles the work, not how they describe it.
  • A fair comparison: Every candidate gets the same scenario and the same rubric, which reduces the room for hiring bias.
  • A realistic job preview: Candidates see what the job involves and some self-select out early, which saves both sides time.
  • Better interviews: The simulation output gives the panel something concrete to probe, so later rounds go deeper.

Which role simulation formats can you use?

Role simulation formats differ in what they measure, how long they take and how well they scale. There is no single best format, only a best format for a given role, stage and amount of candidate patience.

These six formats cover most hiring needs.

Format

Measures

Best for

Candidate time

Main risk

Live or recorded role-play

Communication, composure, persuasion

Sales, support, people managers

20 to 30 minutes

Assessors play the part differently

AI-led role-play (video, voice or chat)

Communication and judgment, at scale

High-volume customer-facing roles

10 to 20 minutes

Weak prompts or vague criteria

Chat simulation

Written tone, speed, de-escalation

Support, chat sales, remote operations

15 to 30 minutes

Measures writing more than product knowledge

In-basket exercise

Prioritization, judgment, delegation

Operations, project managers, team leads

30 to 45 minutes

Feels artificial if the inbox is thin

Situational judgment test

Judgment about workplace scenarios

Early, high-volume screening

15 to 20 minutes

Shows opinion about work, not the work

Work sample or take-home

Core role skill, quality of output

Technical, analytical and creative roles

45 to 90 minutes

Drop-off and scope creep into unpaid work

A few notes on each:

  • Live role-play: The candidate plays their role and an assessor plays a customer, a report or a stakeholder. Fix the script and the rubric in advance, or two assessors playing the same difficult customer will produce two scores you cannot compare.
  • AI-led role-play: An AI interviewer plays the other side and adapts its follow-ups to each answer. It removes scheduling and keeps the character consistent for every candidate.
  • In-basket: The candidate gets a realistic inbox of competing requests and a fixed window to triage, delegate and explain the order. See in-basket training for how the exercise is built.
  • Situational judgment test: A scenario with several possible responses, scored automatically. Treat situational judgment tests as a filter rather than proof of skill.
  • Work sample or take-home: A scaled-down piece of the real job. Cap it at 90 minutes of genuine effort, say so in the brief, and never ask for work the company could ship.

For senior or leadership roles, several of these are often combined into an assessment centre day.

Which roles benefit most from role simulations?

Customer-facing roles benefit most from role simulations, because communication and judgment drive their daily performance and both show up within minutes of a well-built prompt. Technical and leadership roles gain too when the scenario asks candidates to explain their reasoning out loud.

Role

Example scenario

What to watch

Sales rep or account executive

A discovery call with a prospect who pushes back on price

Questions asked, objection handling, a clear next step

Customer success manager

A client finds out a promised feature is delayed

Ownership, a concrete plan, checking the client is satisfied

Support or contact center agent

A live chat with a customer whose order went missing

Empathy, accuracy, policy use, resolution time

Team lead or people manager

A one-to-one with a strong performer who has started missing deadlines

Listening, directness, agreed actions

Operations or project manager

An inbox with a supplier delay, an escalation and a budget request

Priority order and the reasoning behind it

Software engineer or analyst

Walking an interviewer through how they would debug a failing report

Problem framing, trade-offs, clear explanation

Where do role simulations fit in the hiring process?

Role simulations fit after a cheap screen and before the expensive panel. A simulation costs the candidate real time and the company real reviewer hours, so it should not be the first gate.

A simulation is also wasted at the end, when the panel has already formed a view it will defend.

Where role simulations in hiring fit: screen, skills test, role simulation, interview, decision

A workable order for a mid-volume role looks like this:

  1. Screen on hard requirements: Eligibility, location and the two or three non-negotiable skills.
  2. Run a short skills test or SJT: Size the field down to a workable shortlist.
  3. Run the role simulation on that shortlist: Score it blind against a rubric written before anyone saw a candidate.
  4. Interview the people whose work stood up: Use the simulation output as the thing you probe in a structured interview.
  5. Decide with all the evidence side by side: The test score, the simulation result and the interview notes.

Step four is where most of the value sits, and most teams skip it. Ask why the candidate chose that trade-off, what they would do with another day, and what they would cut.

Candidates who did strong work can explain it. Candidates who got lucky cannot.

For high-volume roles, an AI-led simulation can move up to step two, because it scores every candidate the same way without any scheduling.

How do you design a role simulation prompt for an AI interviewer?

A role simulation prompt starts with the three to five highest-impact tasks in the job. Turn each task into a scenario a candidate can handle in two to four minutes of conversation, then give the AI interviewer a character, a context and rules for how to follow up.

Inage showing how recruiters can design effective video interview prompts for role simulations

Align each prompt with a competency

Map every prompt to a competency in the job description. A sales role needs prompts for discovery, objection handling and qualification, not "tell me about yourself".

A simple grid with skills on one axis and prompts on the other shows coverage and stops two prompts testing the same thing.

Ground the scenario in the real job

Before writing a prompt, ask three questions:

  • What does this person deal with on a typical workday?
  • Which situations need the most judgment, communication or skill?
  • Which problems separate strong performers from average ones?

Use real customer language, real internal terms and the edge cases your team sees every week. A customer success prompt might start from a real escalation email.

Candidates do better, and your candidate experience scores improve, when the scenario feels like an ordinary Tuesday afternoon in the role.

Give the AI interviewer a character and a context

Instead of a list of questions, tell the AI who to play. That makes the exchange feel real and pushes the candidate to engage rather than lecture into a camera, as in this example:

"You are a client who has just found out that a promised product feature will not ship on time. Nobody from the candidate's company told you; you are hearing it for the first time on this call. You are disappointed but professional. Open by saying you read the release notes and cannot find the feature."

That level of detail lets the AI hold a realistic conversation without telling the candidate exactly what is coming.

Parts of an AI interviewer prompt for role simulations in hiring: character, context, opening line, follow-up triggers, rubric

Build in follow-up triggers

Strong prompts tell the AI how to react to different answers. If the candidate is vague, press for specifics.

If the candidate handles the first turn well, add a complication, such as a second stakeholder or a tighter deadline. Follow-up triggers are what separate a simulation from a scripted questionnaire.

Brief the candidate

Tell candidates what the exercise is, how long it takes, who they will be talking to and what good looks like at a high level. A clear brief improves completion and makes the result about skill, not about decoding the instructions.

Pro tip: Write the scoring rubric and the pass mark before you send the first invitation. A rubric written after the team has seen a favorite candidate's work tends to describe that candidate.

How do you score role simulation responses?

Role simulation responses are scored on behavior patterns, not on whether a candidate hit specific talking points, because the conversation changes with every answer. Define three to five observable behaviors per prompt, anchor each score level with an example, and have two reviewers score independently before they compare.

Scoring role simulations in hiring: strong, acceptable and weak anchors and four scoring rules

Level

What it looks like in the delayed-feature scenario

Strong

Acknowledges the issue without getting defensive, offers a specific next step with a date, and checks the client is satisfied before ending

Acceptable

Addresses the issue and offers a fix, but does not confirm it solves the client's concern

Weak

Gets defensive, apologizes without a plan, or shifts the blame

Four rules keep scores consistent:

  • Score independently first: When reviewers discuss before scoring, everyone anchors to the first opinion and the second reviewer adds nothing.
  • Weight by role: A sales simulation might weight objection handling above empathy. A leadership simulation might weight transparency above tactical problem-solving. Set the weights before scoring starts.
  • Read the trajectory, not only the peak: Candidates often stumble and recover, or start strong and fade under pressure. Someone who recovers well from an early slip may be the better hire.
  • Calibrate before launch: Have two reviewers score the same five responses. Where they disagree, the rubric is usually vague, so fix the wording before real candidates take it.

What can role simulations not fix?

Role simulations cannot fix a role nobody has defined. If the hiring manager cannot say what good looks like in the first ninety days, the exercise will measure something, and it will not be the thing that matters. They cannot fix a slow process either. Adding a 45-minute exercise to a five-week loop makes it five weeks and 45 minutes.

Every new stage should replace something, usually a screening call that was not producing much. They do not remove bias on their own. Blind scoring and a fixed rubric reduce it.

A live role-play scored on gut feel by a rotating cast of assessors can hide bias behind a look of objectivity. They also cost candidates time. In August 2026, the US Bureau of Labor Statistics JOLTS release recorded 5.2 million hires and 3.1 million quits.

Strong candidates have options, and every extra unpaid hour is a reason to take another offer. That argues for short, well-built exercises, not for skipping them.

Finally, some roles do not need one. For a junior role with a two-week ramp and a low cost of error, a structured interview and a short skills test are enough.

Spend the effort where a mis-hire is expensive.

How do you keep AI-led role simulations fair and compliant?

AI-led role simulations stay fair when the rubric is fixed, every candidate gets the same scenario, a person reviews the AI's scores before any rejection, and candidates are told how they are assessed. Where AI scores candidates, local law may add duties on top.

  • New York City: Under Local Law 144, an employer cannot use an automated employment decision tool unless it has had a bias audit within one year of use, a summary of the results is public, and candidates get notice 10 business days before use.
  • Candidate trust: A Gartner survey of 2,918 candidates, published in July 2025, found that only 26% trust AI to evaluate them fairly, and 25% trust an employer less when AI evaluates them. Explaining what the AI scores, and that a person makes the decision, matters.
  • Accessibility: Offer an alternative format or extra time when a candidate asks for an accommodation.
  • Records: Keep the transcript or recording, the rubric and the reviewer's final score for every candidate.

How do you prevent cheating in role simulations?

Cheating in role simulations is prevented with proctoring, short time limits and follow-up questions that a script cannot answer. Interactive formats already help: an AI interviewer that adds a complication mid-conversation exposes a prepared answer quickly.

Fraud is a real concern. Gartner reported in July 2025 that 6% of job seekers admitted to interview fraud, and it predicts one in four candidate profiles worldwide will be fake by 2028.

  • Proctoring: Use webcam snapshots, tab and screen monitoring, face detection and identity checks. See anti-cheating and proctoring for the controls available.
  • Time-boxed prompts: A two to four minute response window leaves little room for outside help.
  • Randomized order: Rotate scenarios and prompt order so no two applicants see the same sequence, which limits sharing on forums.
  • Follow-ups: Dynamic follow-up questions catch scripted or AI-generated answers.

What mistakes should you avoid with role simulations?

The most common mistake with role simulations is testing confidence instead of competence. A prompt like "tell us why you are a great fit" rewards speechcraft, not skill, and other mistakes come up again and again:

  • No calibration: Uncalibrated rubrics produce large gaps between reviewers and unreliable decisions.
  • Scoring on feel: Reviewers react to energy and charm instead of the behaviors in the rubric.
  • Prompts that run long: Responses longer than about four minutes per prompt raise drop-off and add little.
  • Overbuilt scenarios: Too many characters and twists confuse candidates and blur what is being measured.
  • Unrealistic situations: A scenario the person would never face in the role tells you nothing about the role.
  • Poor instructions: A confusing brief hurts completion and the employer brand.
  • Filing the result away: A simulation nobody discusses in the interview has cost everyone an hour for a number on a spreadsheet.

How do you measure whether role simulations work?

Role simulations work when they separate good hires from struggling ones, so measure them against outcomes. Before launch, test the exercise on two current employees, one strong performer and one still finding their feet.

If the exercise cannot separate those two people, it will not separate candidates.

After launch, track these for each cohort and compare them with your baseline from before simulations:

  • Quality of hire: Hiring manager ratings at 90 days. The quality of hire calculator gives you a consistent way to score it.
  • 90-day retention: Early leavers often point to a gap between the exercise and the real job.
  • Time to productivity: How long new hires take to reach the expected output.
  • Candidate feedback: Completion rate, drop-off by stage and a short post-exercise survey.
  • Recruiter time saved: Screening hours removed, which turns into a clear ROI line for leadership.

Six months after a cohort starts, compare simulation scores with performance reviews. If the exercise predicts well, give it more weight. If it predicts nothing, change it or drop it rather than keep running it because it looks rigorous.

Key takeaway: Most of the value of a role simulation comes from consistency: the same scenario, a fixed rubric, independent reviewers and a check against real outcomes.

How does Testlify run role simulations?

Testlify runs role simulations through two-way AI interviews by video, voice or chat, async chat simulations and one-way video interviews, all scored against the competencies you set for the role. The AI interviewer asks the opening scenario, listens to the answer and probes with follow-up questions.

  • Formats: Video AI, Voice AI and Chat AI interviews, plus chat simulations that recreate live customer chats, Slack threads and help desks, and one-way video interviews.
  • Design help: The AI test builder turns a job description into an assessment and suggests competencies and scoring criteria. Testlify's published C.R.A.F.T. framework on the what's new page explains how to design chat simulations.
  • More than one signal: Simulations sit next to 3,500+ tests, including situational judgment, cognitive, language and coding tests, and 25+ question types.
  • Scoring: Automated scoring with AI candidate insights, a per-skill breakdown, AI-written pros and cons and suggested follow-up questions, with human review for priority roles.
  • Integrity: 20+ proctoring checks, including AI assistance detection and speech authenticity detection for Voice AI and Video AI answers.
  • Fit with your stack: 100+ ATS and HRIS integrations with two-way sync, and 15+ languages for global teams.
  • Audited: Testlify's matching model completed a third-party NYC Local Law 144 bias audit on June 16, 2023.

Ready to run your first role simulation?

Pick one customer-facing role, write three prompts from its hardest real situations, and set the rubric before the first invite goes out. Browse the Testlify test library to pair the simulation with a skills test.

Start a 7-day free trial with no credit card, or book a demo and bring a live role to it.

Frequently asked questions (FAQs)

Snehi Parmar
Snehi Parmar

Human Resources Lead

Snehi Parmar leads People at Testlify, owning hiring, culture, performance, and retention for a 90-person team. She writes on practical HR strategy and building people processes that scale.

Get started.

Hire on proof, not resumes.

Run your first skills-based assessment free — no credit card required.

We use cookies to enhance your browsing experience, serve personalised ads or content, and analyse our traffic. By clicking "Accept All", you consent to our use of cookies.