Cold call AI roleplay simulator (L2-3)
Reps practice cold calls with an AI persona before live dials. New-hire ramp drops 30%.
The steps
- 01
Select and configure the platform
To begin, select a dedicated AI roleplay platform like Second Nature, Hyperbound, or Quantified. While it is tempting to try and build this using a raw ChatGPT prompt, purpose-built tools provide the necessary voice latency, sentiment analysis, and structured scoring rubrics required for professional sales training. • Ownership: Sales Enablement Manager or RevOps. • Time Estimate: 1-2 weeks for procurement and initial setup. • Process: Start by creating a trial account. Navigate to the 'Persona' or 'Scenarios' menu. For beginner setups, ensure the tool integrates with your CRM (Salesforce/HubSpot) so roleplay data can eventually sync to rep profiles. • Common Pitfall: Avoid building a custom internal tool; the voice processing delay in DIY solutions will frustrate reps and kill the realism. • Definition of Done: A platform is selected, seats are provisioned for a pilot group, and the base 'Sales Methodology' (e.g., MEDDICC or Sandler) is configured in the global settings.
- 02
Build three specific persona scenarios
Instead of overwhelming new hires with every possible customer type, focus on your 'Power 3' scenarios. For a typical B2B SaaS company, these are: 1) The 'Not Interested' Gatekeeper, 2) The 'Busy Executive' (The Pivot), and 3) The 'Happy with Current Provider' (The Displacement). • Ownership: Sales Manager. • Time Estimate: 4 hours. • Execution: Inside the platform, go to 'Scenario Builder.' Input a specific ICP (Ideal Customer Profile) description. For example, 'Target: VP of HR at a 500-person tech company. Pain point: High turnover. Personality: Pragmatic and time-poor.' • Example Prompt/Config: Set the 'Resistance Level' to 7/10. Define the 'Winning Criteria' as: a) Stating the value prop in <15 seconds, b) Asking one open-ended discovery question, and c) Securing a specific follow-up time. • Definition of Done: Three distinct scenarios are live and tested by a veteran rep to ensure the AI's objections feel realistic.
- 03
Define the AI scoring rubric
Map out the specific metrics the AI should track during each simulated call. Beginners often make the mistake of just looking for 'completion,' but you need qualitative data. • Settings to adjust: Locate the 'Rubric' or 'Scorecard' section. • Specific fields: Set thresholds for 'Talk-to-Listen Ratio' (aim for 45/55), 'Filler Word Count' (less than 5 per minute), and 'Sentiment Score' (Neutral to Positive). • Custom Keywords: Input your 'Must-Say' phrases (e.g., 'Efficiency,' 'Integration,' 'ROI') and 'Never-Say' phrases (e.g., 'Is now a good time?', 'I just wanted to...'). • Common Pitfall: Setting the bar too high initially. Start with basic hurdles and increase difficulty as the rep progresses through the ramp. • Definition of Done: Each of the 3 scenarios has a quantitative scorecard that triggers an automated 'Pass/Fail' grade based on these parameters.
- 04
Deploy mandatory daily practice blocks
Incorporate AI roleplay as a mandatory daily activity during the first 14 days of a new hire's tenure. This replaces passive 'shadowing.' • Ownership: Sales Enablement. • Time: 30 minutes per day per rep. • Implementation: Add a calendar block for all new hires titled 'AI Roleplay - Scenario A.' Instruct them to complete 5 sessions or until they hit a target score of 85%. • Manager Action: Navigate to the 'Leaderboard' or 'Insights' tab in the tool. Look for 'Area of Improvement' trends across the cohort. If everyone is failing on 'Handling the Price Objection,' run a group coaching session on that specific topic. • Pitfall: Letting reps skip this because they 'feel ready.' Live calls are more expensive than AI minutes. • Definition of Done: All new hires have completed at least 50 simulated calls before their first live dial.
- 05
Conduct weekly data-driven coaching
Transform the manager's role from a micro-manager to a data-driven coach. • Ownership: Front-line Sales Manager. • Time: 1 hour weekly. • Process: Every Friday morning, open the 'Manager Dashboard.' Filter for 'Low Scores' and 'High Sentiment Volatility.' Click into a specific session recording. Most tools will highlight the exact transcript moment where the rep stumbled. • Feedback Loop: Leave a comment in the tool or a Slack message like: 'I noticed the AI gave you a 40% on Objection Handling in Scenario 2. You're jumping to the pitch too fast. Listen to this 30-second clip and try again focusing on the Empathy-Pivot-Ask framework.' • Definition of Done: Every new rep receives at least one targeted coaching snippet per week based on their AI performance data.
- 06
Tie graduation to simulation scores
The biggest shift in this playbook is moving from 'Time-Based Graduation' to 'Skill-Based Graduation.' A rep does not start live dialing because they have been there for 3 weeks; they start because they have cleared the AI hurdle. • Ownership: VP of Sales / Enablement. • Criteria: Define 'Graduation' as achieving a consistent 90% score across all 3 scenarios over 3 consecutive days. • CRM Workflow: Once the rep hits the target, update their status in the HRIS or CRM to 'Active Caller.' Document the 'Score at Graduation' to correlate it with future revenue performance. • Pitfall: Promoting a rep who has 10 years of experience but can't pass the AI simulation. Trust the data; if they can't beat the AI, they won't beat the market. • Definition of Done: A formal checklist where the 'AI Certified' badge is a hard prerequisite for receiving a live lead list.
Next playbooks
Unfamiliar terms are defined in the AI and Revenue Dictionary. Related frameworks live in the framework library.