The Process
How WellsWorkforce works
A structured quality process on both sides. Contributors are vetted before they work, and clients receive reliable output backed by a continuous grading system.
For Contributors
Your path to paid work
Every contributor goes through the same structured process. No pay-to-play, no invite-only gates. Just a clear path from application to earning.
Apply
Sign up and select from the platform's twelve work tracks — from data labeling and translation to specialized fields like finance, legal, and medical expertise. You can apply to multiple tracks.
Assess
Complete a skills assessment designed for each category. These are real task samples, not generic aptitude tests, that predict how you'll perform on actual paid work.
Get Graded
Expert reviewers evaluate your assessment against quality benchmarks. Your grade determines which tasks you can access and at what pay tier.
Get Matched
Tasks are matched to your verified skill level and category expertise. Higher grades unlock higher-paying, more complex work across all categories.
Build Your Reputation
Every completed task feeds into a single quality score that follows you across every track you work in. Your reputation is portable.
Grow Your Earnings
Consistent quality unlocks priority task access, higher pay rates, and early access to new project types. Your track record speaks for itself.
For Clients
How you get work done
Submit a project, get matched with vetted contributors, and receive quality output, with oversight at every step.
Submit a Brief
Describe your project: task type, volume, timeline, and quality requirements. Include calibration examples if you have them.
Get Matched
We match your project with contributors whose grade and category expertise fit your requirements. No unvetted workers touch your data.
Review Output
Review completed work against your specifications. Human quality checks are layered into the process, not just automated filters.
Ongoing Oversight
Track contributor quality scores in real time. Flag issues, request rework, or adjust project parameters as needed.
Quality System
What makes the output reliable
Category-specific assessments
Every contributor is tested on real task samples for their specific category, not generic screening that predicts nothing about actual performance.
Continuous grading
Quality isn't a one-time gate. Contributor scores update continuously based on task accuracy, consistency, and client feedback.
Human review in the loop
Automated checks catch obvious errors. Human reviewers handle nuance, ambiguity, and edge cases that algorithms miss.
Calibrated specifications
Every task includes clear instructions, edge-case examples, and calibration samples so contributors understand expectations before work begins.
Data-handling standards
Contributors agree to confidentiality terms. Client data is encrypted in transit and at rest. Access is scoped to the minimum needed for each task.
Unified reputation
One quality score across all task types. A contributor who excels at transcription brings that same proven reliability to AI training data work.