
Mavu matches workers to the right task from day one.
One permission step at onboarding gives each new worker a real profile, on their own terms.
The data
Work history, behavioural patterns and skills shown through real activity.
Mavu matches global workers to AI training tasks. There are over 400 million gig workers globally, with experience that rarely reaches a platform profile. Mavu integrated the consent layer so new workers could share context they already had.
What people agreed to
One permission step at onboarding, with each worker in control of what they share.
Mavu has a real profile for each worker before a single task is assigned.
The record
Each worker’s consent for analytics to be run on their data, to match them to tasks.
Each match Mavu makes from that data is consented inference.
What it made possible
First-task completion rose 34%. Early churn in the first 30 days fell 28%. Workers were matched to suitable tasks three times faster.
The figures compare new workers who onboarded through the consent layer with those who went through standard onboarding. The quality of AI training output improved measurably.
Getting the right task to the right worker from day one changes everything about how a platform performs. [OpenDataLabs] gave us the signal we needed to do that.
Angelo Paolo Kalaw, CEO, Mavu