What’s the news?
In May and June 2026, Cresta introduced three products that help companies with their AI agent lifecycle. AI and CX leaders want fast development of AI agents to maintain a competitive edge. But this accelerating speed of innovation has resulted in companies deploying numerous AI agents without the appropriate testing or operational discipline. Cresta’s announcements address building, validating, deploying, and continuously improving AI agents at scale.
First, Cresta launched Synthetic Customers, which creates realistic customer personas based on historical enterprise conversation data. These synthetic personas simulate customer interactions for AI agent testing and human agent training while also generating deeper customer behavior insights. Unlike traditional, often static personas built from surveys or CRM records, Synthetic Customers continuously evolve based on actual customer conversations.
Second, Cresta introduced AI Agent Testing 2.0, a major expansion of its automated testing suite whose goal is to help companies validate AI agents before deployment and maintain confidence as those agents move into production. Cresta drafts testing requirements from company documentation, turning them into human-calibrated LLM evaluators. Then, they generate test cases rooted in customer behavior, questions, feedback, and more.
Third, Cresta unveiled Conductor, an agentic development platform designed to accelerate AI agent creation and track success post-production. Conductor gives developers and technical teams a natural-language interface where they can explain, in plain language, what they want to build—then move through discovery, blueprint creation, agent development, testing, deployment, and optimization. The company says Conductor enables organizations to build production-ready AI agents twice as fast while maintaining enterprise-grade controls and oversight. Conductor integrates tightly with both AI Agent Testing 2.0 and Synthetic Customers to create a continuous development and optimization cycle.
Taken together, the announcements address growing demands from enterprises for a comprehensive AI agent lifecycle management platform—and most importantly, one that changes with customer and agent behavior patterns.
What do our metrics say?
n Metrigy’s Customer Experience Optimization: 2025-26 global study, more than 77% of the 656 participating companies say AI testing and optimization tools are very or moderately important, yet only 34.7% were using the tools as of fall 2025. Another 46.3% are planning to implement, so we expect growth in this product category. New risks introduced by AI and automation are the top pain points leading companies to this tool category, as Metrigy discusses in its CX Assurance for AI-Powered Customer Experience market overview report. CX leaders also say they need more rigorous compliance monitoring and must address system failures, while legacy solutions have limited scope. Operational cost reductions and improved customer retention are the top benefits realized among those companies already using advanced CX assurance tools.
What’s Metrigy’s take?
Cresta’s moves to strengthen its CX assurance portfolio underscore the importance of addressing broad operational issues of AI agent development, deployment, and continuous improvement. The first wave of generative AI focused on creating assistants and chatbots. The second wave focused on deploying AI agents capable of performing tasks. The emerging third wave focuses on swiftly building and managing AI agents as enterprise systems that require governance, testing, observability, optimization, and lifecycle management.
Cresta is positioning itself squarely in that third wave.
With the increased complexity and innovation of AI agents, it’s vital to ensure they’re performing as designed, not only at deployment but also at every moment post-deployment. Companies can use these AI-based products to operationalize AI agents, as well as improve human agent performance.
The most differentiated aspect of the announcements may be the integration between the three offerings. Many vendors offer AI agents. Some provide testing. Others offer simulation environments. Few are attempting to connect
development, testing, deployment, optimization, and customer simulation into a unified workflow.
The Synthetic Customers announcement is particularly noteworthy. Most enterprises continue to build customer personas using surveys, demographics, and CRM attributes. Leveraging actual conversation data to create dynamic customer representations could significantly improve testing coverage and agent readiness. This becomes increasingly important as organizations deploy AI agents into customer-facing environments where edge cases and unexpected behaviors often create the biggest risks.
Similarly, Conductor reflects a growing trend toward “AI building AI.” As enterprises struggle to scale agent development across hundreds or thousands of use cases, natural-language-driven development environments could substantially reduce implementation effort. However, the value will ultimately depend on whether organizations can maintain governance and quality as development accelerates. That brings us back to AI Agent Testing 2.0, which may be the most strategically important announcement of the three. Enterprises and consumers consistently tell Metrigy that trust remains one of the biggest barriers to broader AI adoption and acceptance. Testing, validation, and compliance capabilities are rapidly becoming foundational requirements for building that trust.
Viewed together, these announcements indicate Cresta is pursuing a platform strategy that spans the full AI agent lifecycle rather than competing solely in the increasingly crowded AI agent market.
What does Metrigy recommend?
Organizations evaluating AI agents should expand their focus beyond agent capabilities, costs, and integrations.
When assessing AI platforms, enterprises should evaluate:
- How agents are tested before deployment
- How ongoing performance is monitored
- How compliance requirements are validated
- How customer simulations are created and maintained
- How quickly new agents can be developed and optimized
- How AI-generated changes are governed and approved
We recommend that organizations ask vendors to demonstrate not only the AI agent itself, but also the complete lifecycle surrounding that agent—from development through testing, production monitoring, and optimization.
For enterprises already deploying AI agents, Cresta's announcements highlight an important lesson: success increasingly depends on operational discipline as much as model performance. The organizations that achieve the greatest AI ROI over the next few years likely will be those that establish repeatable frameworks for testing, simulation, governance, and continuous improvement.
Cresta's latest announcements suggest the company is betting that AI agent management will become just as important as AI agent creation. Based on current enterprise adoption trends, that's a wager worth watching.
Robin Gareiss
CEO & Principal Analyst for AI CX
