Question brief

Agentic QA in 2026: what is real and what is marketing

"Agentic QA" in 2026 means an AI agent that proposes new tests from user-traffic patterns, or updates existing tests in response to UI drift. Useful as a coverage extender. Not a replacement for engineering judgement on what to test.

Direct answer

Our take: agentic features earn their keep on coverage gaps (tests no engineer would have written) and on self-heal for high-churn surfaces. They are not yet a replacement for a thoughtful test plan.

Section 1

mabl Active Coverage

Active Coverage proposes new tests from observed user behaviour. The most useful output is the "you have no test for this critical path" alert; the auto-generated test scripts still need editorial review.

Section 2

Tricentis Virtuoso (in Tosca portfolio)

Tricentis acquired Virtuoso QA in 2024. The agentic codeless authoring is now folded into the Tosca enterprise conversation; standalone Virtuoso is a less common path in 2026.

Section 3

Devin and general coding agents

Coding agents can author and run test code end-to-end in narrow scenarios. They are not yet the right primary tool for production test maintenance, but they are increasingly useful for spike work.

Section 4

Pricing posture

Most agentic features are positioned as add-ons. Expect a 15-40 percent licence uplift for Active-Coverage-class features. Negotiate the uplift independently from base licence renewal.