Cresta, the leading contact center AI platform for human and AI agents, announced the launch of Automated AI Agent Testing, a major enhancement to the Cresta AI Agent platform that gives enterprises the confidence to deploy safely, reliably, and at scale. The comprehensive suite of tools uses AI to test AI, so businesses can deploy AI Agents with the confidence that they will perform as expected.

"Accuracy and trust are crucial in any AI deployment, but it's difficult for people performing manual testing to catch potential errors at scale," said Ping Wu, CEO of Cresta. "Our new Automated AI Agent Testing suite uses AI to test AI with capabilities like visitor simulations and expert-aligned LLM judges, we can catch potential problems across even the most extreme edge cases before AI Agents ever reach a human customer. With Cresta's Automated Testing, businesses can deploy AI Agents with confidence."

Trust is key for businesses implementing AI agents. According to PwC's 2025 AI Agent Survey, only 25% of business leaders say they would trust an AI agent to act autonomously in customer interactions. Cresta's Automated AI Agent Testing suite provides organizations with evidence-based assurance that every agent release is validated, compliant, and production-ready.

Cresta's Automated Testing suite runs 15x more tests than traditional human testing methods, leading to 35% faster release cycles and improving accuracy by 20%.

Cresta's enhanced testing suite introduces several breakthroughs that streamline testing, improve reliability, and accelerate safe deployments:

• Expert-aligned LLM Judges: LLM-powered tools assess both what an AI agent says and how it follows critical workflows, ensuring security steps are followed, responses are true to approved answers, and more. LLM judges prevent subtle compliance and accuracy issues at scale.
  
  
• More dynamic Simulated Visitors: Realistic, AI-driven "virtual customers" stress-test agents by mimicking how a real customer would behave, based on millions of real conversations within Cresta. Simulated visitors can generate hundreds of variations and personas, surfacing edge cases and failure points that manual testing often misses.
  
  
• Enhancements to AI Agent Evaluators: Dynamic evaluators give enterprises flexible ways to measure AI Agent performance. Cresta's flexible, modular evaluation tools can be applied across workflows, conversations, or single turns—ensuring consistency and compliance while reducing repetitive work for testers.
  
  
• In-Product Feedback Loop: A new in-product feedback loop allows businesses to review past customer conversations, instantly label or categorize any mistakes, and create a test case so the AI Agent will avoid the error in the future.

Wu added, "Cresta helps you move fast and scale confidently with the best-performing and most secure AI Agents. With Automated AI Agent Testing at their fingertips, businesses have the tools they need to trust that every AI Agent release is reliable, scalable, and ready to ship."