Intro to Roleplay Persona Customization

You can tune:

• Information strategy — How much the agent shares, when, and how directly.
• Engagement progression — How the conversation moves from rapport to credibility to pain acknowledgment.
• Difficulty — How challenging the agent is (objections, skepticism, patience, complexity).
• Conversation style — Verbosity, storytelling, humor, and question frequency.

Why use it?

• Consistent challenge — Make agents consistently challenging (and consistently realistic) so reps get predictable practice.
• Standardized difficulty — Align practice difficulty across teams and regions.
• Buyer-type tuning — Match agent behaviour to different buyer types: skeptical exec, curious evaluator, price-focused prospect, and more.
• Beyond prompts — Control behaviour through structured settings instead of relying only on free-form prompt text.

Important product details

• Separate scorecards: Roleplay persona behaviour applies to the main roleplay agent (buyer/decision-maker). The platform uses two distinct scorecards when a gatekeeper is involved:
  • One for the gatekeeper phase (how well the rep navigated the gatekeeper).
  • One for the agent/buyer phase (how well the rep performed with the decision-maker). Behaviour customization affects the agent/buyer persona and its scorecard only; gatekeeper behaviour and scoring are separate.
• Background noise: The Background Environment (background noise) setting applies to the gatekeeper when a gatekeeper is configured on the roleplay. If there is no gatekeeper, it applies to the main agent call. See Gatekeeper Feature Support for details.

2. Behaviour Controls (Structured Settings)

Behaviour customization is organized into four groups: Information Strategy, Engagement Progression, Difficulty Settings, and Conversation Style.

2.1 Information Strategy

Controls how the agent shares information and when it surfaces pain points.

Example use: For a “skeptical exec” persona, you might set Deflection Tendency higher and Proactive Sharing lower so the rep must earn information. For a “curious evaluator,” Proactive Sharing and Detail Level can be higher.

2.2 Engagement Progression

Controls how the conversation moves through stages of engagement.

Example use: For “price-focused prospect” training, set Credibility Establishment high and Pain Acknowledgment to unlock only after value is clear. Use Regression Triggers so that premature discount talk pushes the agent back.

2.3 Difficulty Settings

Controls how challenging the agent is and whether difficulty adapts during the call.

Example use: For “consistently challenging” practice, set Overall Difficulty Level to Hard and Objection Sophistication high. For standardized team training, use the same difficulty profile across agents. Enable Adaptive Difficulty for reps who need the bot to get easier or harder as they improve or struggle.

2.4 Conversation Style

Controls how the agent speaks and interacts at a stylistic level.

Example use: For exec-style practice, lower Verbosity and Response Length; for a collaborative evaluator, increase Question Asking Frequency and Storytelling Tendency.

3. Use Cases

Consistently challenging practice

• Set Difficulty Settings (e.g., Overall Difficulty Level, Objection Sophistication, Skepticism Level) so every rep faces the same level of challenge.
• Use Engagement Progression so the agent doesn’t “give in” too early; Credibility Establishment and Pain Acknowledgment ensure reps must earn advancement.

Standardized difficulty across teams and regions

• Create shared behaviour presets (e.g., “Enterprise discovery – hard”) and apply them across agents used by different teams or regions.
• Keeps calibration consistent for coaching and benchmarking.

Tuning for different buyer types

• Skeptical exec — High Deflection Tendency, low Proactive Sharing, high Skepticism Level, low Verbosity, high Credibility Establishment.
• Curious evaluator — Higher Proactive Sharing, Detail Level, Question Asking Frequency, Storytelling Tendency; moderate Objection Sophistication.
• Price-focused prospect — High Information Withholding on budget, Regression Triggers on premature discounting, strong Credibility Establishment before Pain Acknowledgment.
• Friendly but guarded — Warm Initial State, high Rapport Building, but high Credibility Establishment and Deflection Tendency until trust is built.

Adaptive difficulty for varied skill levels

• Enable Adaptive Difficulty with Adjustment Speed and Performance Window so the agent gets easier or harder based on how the rep is doing.
• Useful for self-serve practice where one agent serves multiple skill levels.

Realism and pacing

• Use Pain Point Timing and Pain Point Depth so pain emerges in a realistic sequence.
• Use Engagement Progression and Regression Triggers so the conversation feels like a real buyer journey, not a script.

4. Examples

Example 1: Skeptical executive (hard)

• Information Strategy: Proactive Sharing low, Direct Answer low, Deflection Tendency high, Detail Level medium, Pain Point Timing late, Pain Point Depth high (strategic).
• Engagement Progression: Initial State guarded, Credibility Establishment high, Pain Acknowledgment only after credibility, Regression Triggers pushy pitch, assumptions, jargon.
• Difficulty: Overall hard, Objection Sophistication high, Skepticism Level high, Patience Threshold medium, Decision Complexity high.
• Conversation Style: Verbosity low, Response Length short, Humor low, Question Asking high (challenging questions).

Example 2: Curious evaluator (medium)

• Information Strategy: Proactive Sharing high, Direct Answer high, Deflection Tendency low, Detail Level high, Pain Point Timing mid, Pain Point Depth medium.
• Engagement Progression: Initial State open, Rapport Building high, Verbal Demonstration high, Pain Acknowledgment moderate.
• Difficulty: Overall medium, Objection Sophistication medium, Skepticism Level low–medium, Patience Threshold high.
• Conversation Style: Verbosity medium, Storytelling high, Question Asking high (exploratory).

Example 3: Price-focused prospect (medium–hard)

• Information Strategy: Proactive Sharing low on budget, Deflection Tendency high on pricing, Pain Point Timing mid, Pain Point Depth medium (efficiency/cost).
• Engagement Progression: Credibility Establishment high, Pain Acknowledgment after value clarity, Regression Triggers early discount ask, feature dump without ROI.
• Difficulty: Information Withholding high on budget, Objection Sophistication medium, Patience Threshold medium.
• Conversation Style: Response Length medium, Question Asking medium (often about cost and ROI).

Example 4: Adaptive difficulty (self-serve)

• Difficulty: Adaptive Difficulty enabled, Overall Difficulty medium (starting point), Adjustment Speed medium, Performance Window last 3–5 exchanges.
• Other behaviour controls set for a “typical discovery” profile; difficulty then moves up or down based on rep performance within the session.

5. How-To Guides

5.1 Where to find Behaviour Customization

1. Go to Roleplay → Agents and create a new agent or open an existing one for edit.
2. In the agent configuration, find the Behaviour or Behaviour Customization section (may be under Advanced or Agent settings).
3. You’ll see the four groups: Information Strategy, Engagement Progression, Difficulty Settings, Conversation Style. Expand or open each to adjust settings.

5.2 How to set Information Strategy

1. Open the agent → Behaviour → Information Strategy.
2. Adjust each slider or dropdown:
   • Proactive Sharing — Low = agent holds back; High = agent volunteers more.
   • Direct Answer — Low = more deflection; High = more direct answers.
   • Deflection Tendency — Higher = more “Why do you ask?” / topic shifts.
   • Detail Level — Low = high-level only; High = specific details, numbers, examples.
   • Pain Point Timing — Early / Mid / Late (when pain is revealed).
   • Pain Point Depth — Surface / Medium / Deep (how much emotional or strategic depth).
3. Save or save as draft. Changes apply to new roleplay sessions.

5.3 How to set Engagement Progression

1. Open the agent → Behaviour → Engagement Progression.
2. Set Initial State (e.g., open, neutral, guarded, cold).
3. Adjust Rapport Building, Credibility Establishment, Pain Acknowledgment, Personalization, and Verbal Demonstration to match how you want the conversation to flow.
4. Set Progression Stage if the agent should start at a specific stage (e.g., early discovery vs. evaluation).
5. Configure Regression Triggers (e.g., pushy pitch, wrong assumption, jargon, premature discount) so the agent steps back when the rep makes these mistakes.
6. Save. Use Instructions to rep or scenario text to align rep expectations with these settings.

5.4 How to set Difficulty Settings

1. Open the agent → Behaviour → Difficulty Settings.
2. Set Overall Difficulty Level (Easy / Medium / Hard) as the base.
3. Tune Objection Sophistication, Information Withholding, Skepticism Level, Patience Threshold, and Decision Complexity as needed.
4. If using Adaptive Difficulty:
   • Turn Adaptive Difficulty Enabled on.
   • Set Adjustment Speed (how fast difficulty changes).
   • Set Performance Window (e.g., last N turns or N seconds) used to evaluate performance.
5. Save. Difficulty applies to the agent/buyer phase only; gatekeeper difficulty (if any) is configured separately.

5.5 How to set Conversation Style

1. Open the agent → Behaviour → Conversation Style.
2. Set Verbosity and Response Length Preference (short / medium / long).
3. Adjust Storytelling Tendency, Humor Usage, and Question Asking Frequency to match the buyer type.
4. Save. Style works together with the agent’s system prompt; behaviour settings take precedence for structured behaviour.

5.6 How behaviour relates to scorecards and gatekeeper

1. Agent scorecard: Behaviour customization affects the main roleplay agent (buyer/decision-maker). The agent scorecard evaluates the rep’s performance in that conversation. All behaviour settings apply to this phase.
2. Gatekeeper scorecard (separate): If the roleplay has a gatekeeper, the gatekeeper phase uses a separate scorecard. Behaviour customization does not apply to the gatekeeper; gatekeeper persona and difficulty are configured in the Gatekeeper section. See Gatekeeper Feature Support.
3. Background noise: When a gatekeeper is present, Background Environment applies to the gatekeeper call. When there is no gatekeeper, it applies to the main agent. See Gatekeeper Feature Support.

5.7 Saving and reusing behaviour presets (if available)

1. After configuring behaviour, check for Save as preset or Save as template in the Behaviour section.
2. Name the preset (e.g., “Skeptical exec – hard”) and save.
3. When creating or editing another agent, choose Apply preset and select the saved preset to copy behaviour settings.
4. Adjust any settings as needed and save the agent.

If your workspace does not support presets, duplicate an existing agent that has the desired behaviour and then change name, scenario, and scorecard as needed.

6. Quick Reference

For gatekeeper setup, scorecards, and background noise, see Gatekeeper Feature Support. For multi-persona roleplays, see Multi-Persona Roleplay Support.