Call Simulation

How a call simulation works

A simulation recreates a slice of the real job. The participant gets a short brief (customer account details, a policy or two, sometimes a simplified system screen) and then takes a call. The customer may be:

• A live roleplayer, usually a trainer or recruiter following a script with set escalation points
• A recorded customer with pauses for the candidate to respond, scored afterward
• An AI voice customer that reacts to what the candidate says, used for AI-run call simulations at volume

Calls last 5 to 12 minutes in hiring and up to 20 minutes in training. In hiring, it is a specific type of job simulation, close to a situational judgment test but spoken and real-time rather than multiple choice.

Using simulations in hiring

A short simulation is one of the best predictors of how a new agent will sound in week one, because it tests the actual work rather than what the candidate says about it. It also works as a realistic job preview: some candidates realize the job is not for them, which is cheaper than finding out in training.

Practical design rules:

1. Base scenarios on real top call drivers, simplified. A first-time candidate cannot learn your refund policy, so give them a one-paragraph version.
2. Include one moment of friction. The customer interrupts, gets annoyed or asks for something the policy does not allow.
3. Score with a rubric, not overall impressions. Typical areas: greeting and control of the call, listening and accurate information, empathy, problem-solving within the rules, and closing.
4. Keep the same scenario and rubric for every candidate in a role so results are comparable and defensible.

Pair it with a voice and accent assessment earlier in the funnel; the voice test checks clarity, the simulation checks judgment and service.

Using simulations in training

In new-hire training, simulations let agents practice before real customers hear their mistakes. Good programs build a ladder: simple calls in week one, then multi-issue calls, then angry or vulnerable customers, then calls requiring system navigation at speed. Trainers score each one and agents must pass a final set before moving to nesting.

Tenured agents benefit too. Simulations work well for launching a new product, rehearsing a policy change and coaching specific skills, the same way roleplay training is used in sales. AI-voiced customers make it possible to practice dozens of calls without tying up a trainer for each one.

Example scenario and scoring

Scenario for an internet provider: "Maria Lopez calls because her bill went up by $18. Her 12-month promotional rate ended last week. Policy: you may offer a $10 monthly credit for 6 months once per account. She mentions a competitor's offer halfway through."

Scoring out of 20, four points each for: verifies identity and takes control, explains the price change clearly without jargon, acknowledges frustration without over-apologizing, offers the credit correctly and handles the competitor mention without promising what policy forbids, and confirms next steps and closes. A strong candidate scores 15 or more. Below 11 is a no-hire for this role; 11 to 14 goes to interview with notes on what to probe.

Common mistakes

Making the scenario too hard is the most frequent. Hiring simulations should test aptitude, not product knowledge the candidate cannot have yet. Other issues: roleplayers who improvise differently for each candidate, which makes scores unfair; scoring only friendliness and ignoring accuracy; running it at the very end of a long process when candidates are tired; and never checking whether simulation scores line up with quality and attrition after hire. Review that link every few months and adjust the scenario and pass mark.