AI Agent Tester - Real-World Tasks

Everyday AI agent testing

Test two AI agents side by side on real errands using your own accounts and smartphone. This US contract pays $25 to $45 per hour, with approximately 1.5 hours per run including evidence collection and scoring.

The project is already active, and work can be completed entirely from your phone on your own schedule. No coding or professional AI background is needed.

Responsibilities

• Run scenarios involving bills, returns, email follow-ups, bookings, shopping, groceries, travel, family logistics, outings, and forms.
• Approve or stop an agent at key steps and screen-record every attempt.
• Score completion, quality, control issues, satisfaction, and time.
• Provide clear written feedback on what worked and what did not.

Accounts and equipment

• An active personal AI assistant account at the Plus tier.
• Active access to Muse and Dots.
• A smartphone capable of recording its screen.
• A payment card for test purchases, which are reimbursed afterward.

Participant background

Participants must be based in the United States, hold a bachelor's degree, and already use AI in daily life. The listing calls for two to three years of experience, preferably in engineering, business management, or office and operations work.

Careful instruction-following, documentation, and clear written communication are important. Generalists are welcome, while domain experts are preferred.

Selection and onboarding

Selection includes screening questions, an interview lasting 15 to 30 minutes, and profile review. Selected participants receive an offer and complete onboarding within 24 to 48 hours, with work beginning as tasks are assigned.