Can AI agents become part of the testing team?

I’m curious how teams are adapting their testing workflows as AI agents become more capable.

Instead of using AI only to generate test cases, could agents take ownership of specific testing tasks while humans review the results?

I’ve been exploring this idea with tools like Sharkly, where AI agents and humans can work within the same project workflow.

How are you currently using AI agents in your testing process?

“Sharkly” — I haven’t come across this tool before. Do you have a link or more details? Curious to know

Whether this is a newly launched tool?

Katalon’s True Platform carries six AI agents that share context across the testing lifecycle — test generation, autonomous execution, defect management, root cause analysis, and release reporting all connect, so when a test fails the bug reporter already has evidence, and when root cause is identified the report generator factors it in.

Read this blog for more information
How AI Agents Work in Testing: True Platform’s Honest Assessment

There are lot more, will ping in sometime

can you shed some lights on Sharkly?

I want to add some more points as I mentioned yesterday.
There are some more recent and very valuable upgradations have come in the Katalon like as below

  1. Run with AI (added recently in May 2026) is a prompt-to-script agentic loop: you write a test in plain language, and it dispatches an autonomous agent to execute it against your live app — clicking, filling fields, verifying states, and recovering from unexpected UI changes on its own — then hands you back a ready-to-run Katalon Studio script.

for more information please check
Katalon Product Roundup May 2026

2) TrueTest generates and maintains test cases from real production user journeys, and can now regenerate existing cases when a flow changes rather than starting over

  1. StudioAssist’s Agent Mode (built on MCP)- it acts within your project to interpret documentation, troubleshoot errors, and generate scripts or data and much more.