mabl Announces Next-Generation Agent-Based Testing Platform for the AI Development Era

This release introduces mabl's "Active Coverage" feature, a quality verification capability designed to maximize the speed of AI coding agent development without hindering it, addressing challenges highlighted in the company's "2026 State of Quality Engineering Report."

May 12, 2026 — As an AI-native company that has provided enterprise testing platforms since 2017, mabl today released a new set of features designed to maximize the speed of "agent-based development," which has garnered significant attention recently, from the perspective of continuous quality assurance. This new feature set, called "Active Coverage," provides essential capabilities for enterprise quality assurance as the speed and scale of AI-driven code generation continue to rapidly expand within organizations across all industries.

This feature release coincides with the announcement of the findings from mabl's "2026 State of Quality Engineering Report," a survey of approximately 1,000 software professionals. The report confirms a widening gap between the speed of code generation and quality verification. The report reveals that opinions are almost evenly split among teams currently using AI coding agents: 41% reported that "AI improved code quality," while 37% stated that "code generation speed improved, but quality declined," highlighting that the biggest factor determining whether AI improves or degrades quality lies in a team's quality foundation.

Dan Belcher, co-founder of mabl, states:

"Coding agents can write and release code faster than any development team in the past, but it's also true that agents evaluating their own work have a bias towards prioritizing release over quality. Our enterprise customers told us that the quality management layer must be independent of development and must not impede its speed at all. We have rebuilt mabl as that foundational layer. This is an 'Agent-based Quality Platform' that tests both human-created and AI-generated code with the same rigor as an independent reviewer, providing the observability, data lineage, and data retention that enterprise companies have always needed."

New mabl features are designed to enable "Active Coverage" to support a new era of agent-based software development.

Agent Instructions: Define unique development and implementation policies and testing directions for enterprises and teams directly in mabl as "rules and constraints (context)" for the entire application. This eliminates the need for detailed instructions in prompts every time an individual test is created, automatically and permanently applying consistent quality standards for the application to all AI-created tests, root cause analysis of failed tests, and self-healing processes during test execution.

Cloud Test Generation: Create tests entirely in the cloud without requiring tool setup in a local environment. Tests can be triggered via MCP from a browser, CLI, or IDE. Multiple test generation sessions can run in parallel, enabling fast test generation and coverage expansion that matches development pace without hindering the speed of new feature implementation.

Runtime Recovery: Agents resolve unexpected application behavior that occurs during test execution. Even if unstable events occur that would normally interrupt a series of tests, they are automatically avoided, and subsequent tests are continued as much as possible. This provides teams with information that, in the event of a pipeline stoppage, "a true defect that needs investigation has definitely occurred."

Interactive Results Analysis: Engineers can interactively investigate test execution results in detail (e.g., asking questions to confirm details, suggestions from a human perspective) using natural language for individual tests, deployments, or the entire workspace, reducing the time required for manual log investigation from hours to minutes. They can delve deeper into investigations with AI.

Atlassian Rovo Integration: mabl's test intelligence with Jira and Co