# How to Test Multi-User & Approval Flows with AI

> Automate multi-user and approval-flow testing with AI agents – one acts as User A, another as User B, in isolated sessions, to verify hand-offs and approvals.

Source: https://qa.tech/use-cases/multi-user-testing

---
QA.tech can run two agents as two different users in fully isolated sessions – one submits, the other approves – so workflows that span people are tested end to end. The "[Wait For](https://docs.qa.tech/core-concepts/dependencies#wait-for)" primitive spawns a second actor in a separate browser with its own cookies; "[Resume From](https://docs.qa.tech/core-concepts/dependencies#resume-from)" continues from a saved state.

[Book a demo](https://qa.tech/demo)

Sub-use-cases

**Covers** Submit → approve chains, sender/receiver messaging, collaborative editing hand-offs and multi-stakeholder workflows.

1.  01
    
    ## What a multi-actor test should cover
    
    Approval chains, sender/receiver flows, collaborative actions, and any workflow where one user must react to another's action.
    
2.  02
    
    ## How does AI run two users at once?
    
    One agent authenticates as User A and acts; a second agent ([Wait For](https://docs.qa.tech/core-concepts/dependencies#wait-for)) runs in an isolated session as User B and responds; state is coordinated across the run.
    
3.  03
    
    ## When to test multi-user flows
    
    On changes to approval logic, collaboration or any cross-user workflow.
    
4.  04
    
    ## Who needs multi-actor testing
    
    Enterprise platforms with approval chains, collaborative SaaS, messaging and marketplace products.
    
5.  05
    
    ## How QA.tech helps
    
    Coordinating state across multiple user sessions defeats most automation. QA.tech orchestrates true multi-actor tests in isolated sessions, including [cross-application](https://docs.qa.tech/core-concepts/applications-and-environments#multi-app-flows) hand-offs, so approval and hand-off flows are finally testable.
    

▒ ░▒ ▒ ▓░░ ░▒▓▓░▓░░ ░ ░░ ░░░▒░▒▒░░ ▒░▒░ ░ ░░░░ ░░▓░░▒░ ░░▒█▒ ░ ░▒█ ▓▒ ░░ ▓▒░▒█ ▒ ░░▓ ░░▓░▒ ░░░░░ ░ ░ ░▒▒ ░▒░▓░ ░▒ ▓ ▒▒░░ ▓ ░░░░░░ ▓▓░ ░ ░ ░▓ ░ ▓█ ░▒▒░ ░▒ ▒▒▓ ░ ░░ ░░░░▓ ▒ ▒ ░░▒ ▒▒░ ░▒ ▒▒ ▓░░▓ ░▒▓▓░▒░░█░░▒▒░ ▒░ ▒▓▒ ░░▓ ▓ ▒▒ ░▒▒░░ ░ ▒ ▒░░░░

## Companies running multi-user & multi-actor flow testing with QA.tech

-   [![Airpelago](https://qa.tech/assets/airpelago-1y9XaVff.svg)Tests across multiple personas: customers, partners and operators.](https://qa.tech/case-studies/airpelago---scaling-drone-operations-with-reliable-end-to-end-testing)

FAQ

## Common questions

Can two users really run in one test?

Yes – "Wait For" spawns a second agent in a fully isolated browser session.

How is state shared between actors?

The run coordinates state so User B reacts to what User A did.

What flows is this for?

Approval chains, sender/receiver, and any multi-stakeholder workflow.

Related use cases

-   [
    
    Account & Profile Settings Testing
    
    An AI agent runs every settings action a user can take – update profile fields, change email and password, upload an avatar, toggle preferences – then verifies the changes persist after a refresh. Settings pages look simple but touch storage, auth and email, so they regress quietly.
    
    Read ██▒▒░▓ ░░▒▒░░▒▒░░█▒░▓█░░░▒▒▓ ▓░▒▒░░ ▓░▓▒░░░█░░░░█▒ ▓░░░░░▒░░▒ ░░▒░░▒ █░░░ ░ ░░░░░▒▓▓▓░░░░▒ ▒░▓░▒▒▒▓▒ ░▓ ░░▓ ░░░](https://qa.tech/use-cases/account-settings-testing)
-   [
    
    Analytics & Tracking Event Testing
    
    An AI agent completes the real user flow in a browser – a checkout, a signup, a form submit – then verifies the tracking request actually fired: URL pattern, method, status code, and the payload it carried. The purchase beacon with the wrong revenue value fails the test, before it fails the month-end numbers.
    
    Read ░░ ░▓ ░░▒ ░░▒ ░▒ ░▒▓ ▒▓█ ░░▒▓░▒ ░ ▒░█ ░▒░▒█▓▓▒░░ ░▒▒░▓░░█░▓░░▒░ ░ ░░ █░▒█▒█░▒▒▒░ ▒░▒▓▒ ░▒░▒▒▒░ ░░ ░▒ ░▒▒░██▒░░](https://qa.tech/use-cases/analytics-tracking-testing)
-   [
    
    Authentication & Login Flow Testing
    
    An AI agent runs your real login journeys – including the parts traditional tools can't, like opening an email or reading an SMS code – and verifies the user ends up authenticated. QA.tech gives the agent its own inbox and phone number, so magic links, one-time passwords and confirmation emails are tested end to end, not mocked. See how Upsales replaced 320 hours of manual testing a month with this approach.
    
    Read ░▒░ █▒▓░░▓▒ ░▒░▒ ░░░ ▓█░ ░ ░ ░░▒▓▓▒░ ░░░░▓ ░ ▒▒░▓▓▒▒▒░ ▓░░▓ ▓░▒░░░▒ ▒▒▒░▓░░ ░░▒░░ ░░█░▒ ░░░ ░▒░█░▒░▒▓█▒▒░█ ░](https://qa.tech/use-cases/authentication-testing)

[← Previous

Multi-Environment Testing

](https://qa.tech/use-cases/multi-environment-testing)[Next →

Notifications & Activity Feed Testing

](https://qa.tech/use-cases/notifications-testing)

## Your team moves fast. Can your testing keep up?

QA.tech agents test your product autonomously, so moving fast never means shipping broken. See how it works in a 30-minute demo.

[Get a demo](https://qa.tech/demo)