If you’ve researched microtask and AI data work before, you may have come across UHRS, which stands for Universal Human Relevance System. It’s Microsoft’s internal AI training data evaluation platform, and it’s accessible to workers through approved partner platforms like Appen’s CrowdGen. Tasks include search result relevance rating, image labeling, content quality evaluation, and intent classification for Microsoft’s AI systems, including Bing and Cortana. The pay is task-based and can range from a few cents per task to well above $10 per hour, depending on your qualification level and the project type. It’s one of the more consistent AI data platforms in terms of ongoing task availability.
Pay
$8 to $15 per hour for general tasks. Specialist qualifications unlock higher-tier tasks paying $15 to $25 per hour. Pay is task-based with rates visible before accepting each job
How You Make Money
You apply through Appen's CrowdGen platform, pass the UHRS qualification assessments, and complete search evaluation and data annotation tasks
Payment Method
PayPal, Payoneer, bank transfer (via Appen's payment system)
Minimum Payment
Varies based on Appen's current payment schedule
Skill Level
Beginner to Intermediate (qualification assessments required per task type)
Pros
Backed by Microsoft which guarantees task legitimacy and volume
One of the more consistent task availability platforms in the AI data space
Pay is shown before accepting tasks so there's no surprise when you submit
Accessible globally through Appen's platform in 130+ countries
Builds genuine AI literacy and industry experience over time
Cons
Access requires going through Appen's CrowdGen which has its own application process
Monthly payment cycle is slower than weekly-paying platforms
Some qualification tests are rigorous and not everyone passes on the first attempt
Task availability can shift based on Microsoft's current AI development priorities
Best For
Workers globally who want structured, Microsoft-backed AI evaluation work with consistent availability and visible pay rates before starting any task
