Skip to main content
The Sandbox allows you to test your AI agent’s conversation skills before launching real campaigns. You act as a prospect while your AI agent tries to sell your offering, helping you identify and fix issues with your setup.

​
How Sandbox Works

Conversation Simulation:
  • Your AI agent attempts to sell your offering
  • You role-play as different types of prospects
  • Test various scenarios and responses
  • Evaluate AI behavior and conversation quality
What Gets Tested:
  • Prompt quality: How well your Context and First Message prompts work
  • Offering effectiveness: How well the AI understands and presents your solution
  • Conversation flow: Natural dialogue progression and responses
  • Objection handling: AI’s ability to address prospect concerns

​
Testing Strategies

​
Good Prospect vs Bad Prospect

Good Prospect Testing:
  • Act interested and engaged
  • Ask relevant questions about your solution
  • Show buying intent signals
  • Test if AI can close effectively
Bad Prospect Testing:
  • Be skeptical and challenging
  • Raise objections (timing, relevance…)
  • Act disinterested or distracted
  • Test AI’s persistence and objection handling

​
Challenge Your AI

  • Push boundaries: Test edge cases and difficult scenarios
  • Ask tough questions: See if AI can handle complex inquiries
  • Be uncooperative: Test how AI handles difficult prospects
  • Change topics: See if AI can redirect conversations effectively

​
Requirements

Campaign Dependency: You must create a campaign before using the Sandbox. The Sandbox tests specific campaign configurations, so you need:
  • A configured campaign
  • Selected offering
  • Selected prompts
Important: Your LinkedIn doesn’t need to be connected to use the Sandbox.

​
What to Look For

​
Red Flags

  • Unnatural tone: AI sounds robotic or scripted
  • Generic responses: Lack of personalization or context
  • Poor objection handling: Can’t address prospect concerns
  • Weak closing: Fails to secure meetings or next steps

​
Quality Indicators

  • Natural conversation flow: Responses feel human and contextual
  • Strong personalization: AI uses prospect information effectively
  • Confident objection handling: Addresses concerns professionally
  • Clear value communication: Effectively explains your solution’s benefits

​
Optimization Process

  1. Test thoroughly: Run multiple scenarios and prospect types
  2. Identify weaknesses: Note where conversations break down
  3. Refine components: Update prompts or offering based on findings
  4. Retest: Validate improvements in new Sandbox sessions
  5. Repeat: Continue until conversation quality meets standards

​
Best Practices

​
Testing Approach

  • Be thorough: Don’t rush the testing process
  • Vary scenarios: Test different prospect personalities and situations
  • Document issues: Keep notes on what needs improvement
  • Test after changes: Always retest after modifying prompts or offerings

​
Before Launch

  • Complete testing: Ensure AI handles all common scenarios
  • Validate improvements: Confirm all identified issues are resolved
  • Team review: Have others test if working in a team environment
The Sandbox is your safety net. Invest time in thorough testing to ensure your AI agent performs excellently before engaging real prospects.
Last modified on February 15, 2026