Eightfold AI and Workera have proven the enterprise skills intelligence model. No African equivalent exists. This is the measurement layer for Africa's AI workforce.
28 objectively scored questions, four per domain. Every question has one correct answer, which is what makes a percentile against the database meaningful rather than a comparison of self-assessments.
How models actually work: training, inference, context, and the failure modes that follow.
Systems that take actions and use tools, and the controls that keep them inside their remit.
Reading a dataset honestly: distribution, correlation, leakage, and why accuracy alone misleads.
Specifying a task so the output is usable, structured, and repeatable rather than lucky.
Impact assessment, oversight that actually works, and the African obligations now in force.
Choosing a first use case, scaling past the pilot, and measuring adoption against an operational metric.
Working fluently with the tools themselves: retrieval, sampling settings, and where write access is unsafe.
A fluency score out of 100 and a score for each of the seven domains, so a person can see exactly which capability is missing rather than being told they are broadly behind.
Every score is placed against the accumulated database, so an organisation sees not only how its people scored but where they sit against everyone else measured.
Each sitting is objectively scored, and tab switches, copy and paste attempts and fullscreen exits are recorded so a flagged result can be reviewed rather than quietly trusted.
An individual score on its own is a number. The same score expressed as a percentile against a growing national database is a position, and a position is what an employer, a ministry or a university can act on.
That database does not exist in Africa today. There is no shared instrument, no accumulated pool of verified AI capability data, and therefore no way to say whether a cohort is ahead of or behind the national picture. Each pilot adds to the pool, which is what makes the first movers the reference point for everyone who follows.
The instrument is fixed and objectively scored. That matters: a benchmark built on self-assessment measures confidence, not capability, and confidence does not survive contact with an audit.
Academic validation partnership with Covenant University (in discussion).
Take the diagnostic as an individual. Seven domains, 28 questions, and your score placed against the database.
$1,500
One cohort, up to 25 people.
$5,000
One organisation, up to 100 people.
Custom
Multi-institution or national cohorts.
Request a pilot, or take the framework overview first.
Seven domains, 28 objectively scored questions, one correct answer each. Percentiles are computed against every prior sitting in the database at the moment of submission, and the cohort size is reported alongside the figure so the position can be read in context.
This diagnostic measures AI fluency at the moment of sitting. It is a capability measurement, not a certification, and it does not replace a professional qualification.