What are TestGenie Agents?

TestGenie Agents are the agentic layer of TestPlus: AI agents that watch your requirements, defects and test executions, work out what should happen next — a locator fix, an impact-based regression scope, a defect retest — and raise it as a proposal in the Agent Center. Agents propose; you decide. Nothing changes without an approval, and every decision is recorded.

Can TestPlus agents change my tests without approval?

No. Every agent action runs through a per-project autonomy policy with three levels per action type: Auto, Needs approval, or Forbidden. New projects start conservative — agents only propose, humans approve. As trust grows, project admins can promote individual actions to Auto. Destructive actions such as deleting tests can never be automated.

What is self-healing write-back?

Classic self-healing fixes a broken locator for one run and forgets — the same heal is re-paid on every execution. TestPlus agents detect the heal, match it to the page-object element, and propose promoting the healed locator permanently. On approval the element is updated, the old locator is kept as a fallback, and the next run heals zero steps.

Home / Product / TestGenie Agents

🤖 Agentic QE

TestGenie Agents —
Agents Propose. You Decide.

AI agents watch your requirements, defects and test runs — and bring you the next action as a ready-to-approve proposal. Locator fixes made permanent, impact-based regression scopes, defect retests queued the second a fix lands. Never a change without your approval.

See Agents Live What They Do

app.qynte.com/agent-center

OverviewRequirementsTest DesignTest Plan & ExecutionAutoPlus

🩹Locator update — LoginPage.userName Pending

[name="user"] → [name="username"] · healed 1x — each heal costs an AI call

📋REQ-42 changed — impact assessed Pending

✏️ 4 linked tests need modification · ▶️ 6 automated run on approve · 2 manual flagged

Approve

Reject

Agent loops

Four Loops That Close Themselves

Self-Healing Write-Back

A heal fixes one run — the write-back makes it permanent. Agents detect the heal, match the element, and propose the locator promotion. Old locator kept as a fallback; next run heals zero.

Requirement Impact

A requirement changes and you get exactly two answers: which linked tests need modification, and which tests — manual and automated — need execution, scoped by the Knowledge Graph.

Defect Retest

The moment a defect is marked fixed, linked tests are flagged ready-to-retest and a proposal lands: automated tests run on approve, manual tests are named for your testers.

Test Design Changes

When a requirement's contract drifts, agents flag the test cases whose steps no longer match — proposed as reviewable edits, never silent rewrites.

Agent Center

One dedicated home: the proposal inbox across every loop, decision history, and live agent activity — with contextual cards right where you work in Monitor and test cases.

You Stay in Control

Per-action autonomy policy: Auto, Needs approval, or Forbidden. Conservative by default, promoted as trust is earned. Deleting tests can never be automated.

How it works

Signal to Action in Five Steps

1 · Signal. A run heals, a requirement changes, a defect is fixed — in TestPlus or synced from Jira / Azure DevOps.
2 · Analyse. Agents assess impact using traceability links, run history and the Knowledge Graph.
3 · Propose. A card lands in the Agent Center with the evidence — what, why, and exactly what will change.
4 · You decide. Approve or reject in one click — recorded with your name, forever auditable.
5 · Execute & learn. Runs fire, locators update, tests get flagged — and the next signal starts the loop again.

Watch a Loop Close

app.qynte.com/agent-center · autonomy policy

🩹Heal write-back Needs approval

▶️Regression runs Needs approval

🔍Impact analysis Auto

🗑️Delete tests Forbidden

Autonomy is a dial, not a switch — promote each action to Auto only when it has earned it.

Your QA Findings Shouldn't Sit Idle

Heals forgotten. Fixed defects waiting weeks for a retest. Requirement changes nobody re-scoped. Watch the agents close those loops — with you in charge.

Request a Demo

TestGenie Agents —Agents Propose. You Decide.