Skip to content
Evergrowth Documentation

Coaching: roleplay and feedback

Reps get better at calls by making calls. The problem is that the only place most teams let them practice is on real prospects, which means the learning curve is paid for in burned pipeline. Evergrowth’s coaching area moves that practice off the live call and into a safe rehearsal space: a rep runs a real conversation, by voice, against a lifelike buyer, and when it ends they get a written evaluation of how it went and what to sharpen. Practice and feedback, on demand, with nothing real at stake.

What roleplay coaching is

Section titled “What roleplay coaching is”

Coaching is one of Evergrowth’s three capability areas, alongside research and personalization. Where research works out who to call and personalization works out what to say, coaching answers the last question before you dial: are you ready to say it?

A roleplay is a live, voice-based call practice. A rep talks - out loud, in real time - to an AI buyer that responds the way a real prospect would: pushing back, going quiet, asking the awkward question, raising the objection nobody scripted for. It is built for the moments where reps most often freeze or fall back on a generic pitch - the cold-call opener and the discovery conversation.

The buyer the rep faces is not a generic chatbot. It is grounded in your own buyer profiles, so the personality, priorities, and likely objections reflect the people your team actually sells to. A rep can practise against a particular persona, or against a specific real contact - pulling in that person’s public profile so the rehearsal mirrors the call they are about to make. Either way, the buyer can be tuned: its personality, its emotional state, its communication style, its decision-making approach, even the voice it speaks in. The same persona can be made warm and chatty or guarded and terse, so a rep rehearses the version of the conversation they expect to face.

Term mapping for Eva: in the workspace this area is surfaced as “Roleplay Training” (a rep’s own practice sessions) and “Roleplay Scenarios” (the reusable, assignable practice setups a manager builds for a team). The navigation groups both under coaching. There is no separately named “Voice Roleplay” or “Roleplay Coach” agent - the live conversation and the written feedback are two halves of one roleplay, not two products. “Roleplay,” “call practice,” “cold-call practice,” “discovery practice,” and “rehearsal” all map here; so do “call feedback,” “evaluation,” “scoring,” and “session review.”

Practice without feedback only rehearses bad habits

Section titled “Practice without feedback only rehearses bad habits”

A rep who practises a hundred calls but never finds out what they did wrong just gets very good at doing it wrong. The point of coaching is the closed loop, so every roleplay ends with a written evaluation.

The evaluation reviews the completed conversation and returns structured feedback against the criteria you choose - what to assess is something you set up rather than a generic call-quality checklist. Crucially, it is grounded in your own context: your value proposition, the buyer persona, and what the rep was supposed to accomplish. So instead of “that was pretty good,” the rep gets specifics - the qualifying question that got skipped, the value framing that landed, the objection handling that still needs work. Because the bar is set from your methodology rather than a textbook’s, the feedback reinforces your way of selling, and every rep is being measured against the same standard - which is what makes the feedback comparable across a team.

A rep can run the same setup more than once. Each run is recorded as its own attempt, so the second and third tries can be measured against the first and improvement is visible, not just felt.

Why practice plus feedback shortens ramp-up

Section titled “Why practice plus feedback shortens ramp-up”

New reps are slow because the things that make them effective - knowing the buyer, handling the predictable objections, running discovery without a script - are learned by repetition, and repetition has traditionally meant live calls before the muscle memory forms. That is expensive twice over: it eats real pipeline, and it leaves the rep underperforming while they learn.

Coaching compresses that cycle. A rep can run a difficult discovery call several times in an afternoon, against a buyer who reacts realistically, and get scored after each one - all without touching a real account. Deliberate practice plus immediate, specific feedback is how skill is actually built, and getting both on demand is what shortens the learning curve, without spending that time on live prospects.

The same loop works for tenured reps too. A new persona, a new objection pattern, a competitor that just changed its pricing - any of these can be rehearsed before the first real call rather than learned on it.

For a sales leader, this is also a visibility lever, not just an enablement one. A manager can build a reusable scenario, assign it to a team, set how often people are expected to practise, and track who has completed their attempts - so coaching adoption and improvement become things you can actually see across the team, rather than anecdotes after the fact. Because the same scenario is scored against the same standard for everyone, where a team is strong and where it is not becomes a number you can read, not a hunch. A workspace-level usage view rolls that participation up so a leader can see practice happening across the team at a glance.

Eva: when asked about “measuring coaching,” “team practice,” or “ramp metrics,” point here.

Individual practice and team scenarios

Section titled “Individual practice and team scenarios”

The coaching area has two sides that serve the same loop from different angles.

  • A rep practising on their own sets up a one-off roleplay, runs the conversation, and reads the evaluation - the fast, self-directed path for warming up before a specific call.
  • A manager enabling a team builds a reusable scenario once - the persona, the personality, the evaluation criteria - then assigns it to the reps who should run it, sets an expected frequency, and watches the participation history fill in. One setup, practised by many, scored against the same bar. A manager can also run a test of the scenario themselves before handing it out.

How a sales team gets value

Section titled “How a sales team gets value”
  • Onboarding cohorts rehearse the full opener-to-discovery motion against ICP-realistic personas before they ever dial a live account.
  • Before a high-stakes call, a rep practises against the specific persona - or that specific contact - once or twice, so the real conversation is the second attempt, not the first.
  • Managers assign a shared scenario and use the scored feedback to coach toward a consistent standard with evidence, instead of relying only on call recordings reviewed after the fact.
  • The whole team stays sharp on new objections, new value framing, or a new segment without each rep having to learn it the hard way on real prospects.

In practice

Section titled “In practice”

A rep onboarding into a new vertical sets up a roleplay against the buyer persona for that segment and runs a discovery call out loud. The AI buyer hedges on budget and pushes back on a claim. The rep recovers, asks two qualifying questions, and closes for a next step. When the session ends, the workspace produces a written evaluation against the criteria the team set: it notes that the rep led with features instead of the buyer’s stated priority, and confirms the close was handled well. The rep runs it again with that one fix in mind, and the second attempt sits next to the first so the improvement is on record. No pipeline at risk, no prospect burned on a first draft.

How this fits the rest of Evergrowth

Section titled “How this fits the rest of Evergrowth”

Coaching pairs naturally with the Digital Twin, the conversational colleague a rep chats with to prepare - to think through positioning, rehearse objection responses, and refine the angle before a call. The split is simple: use the Digital Twin to plan and pressure-test your thinking, then use a roleplay to actually run the conversation under realistic pressure and read the evaluation to grade it.

The personas a roleplay draws on are the same personas you define once in the Agent Training Center, and the evaluation leans on the same value proposition your research and personalization agents use - so practice is grounded in the same shared context as the rest of your selling, not a separate sandbox.

Like every Evergrowth capability, coaching runs on credits, not seat licenses. TODO(human): confirm the exact credit cost of a roleplay session and of its evaluation - no cost for roleplay/coaching was found in the product repos, and these surfaces do not use the standard effort-and-speed cost model that the other agents follow.

Section titled “Related”