The problem before the model
How to sit with a client, find the process that is actually costing them, and scope something that can be delivered.
Unemployed graduates are trained over twelve weeks to design, build and deploy commercial AI agents, and keep the majority share of the revenue their agents earn.
Interns work a structured build cycle alongside a senior engineer, on problems brought to us by paying clients. The output is not a portfolio project. It is an agent that goes into production and gets licensed to a business.
Seventy-five percent of the licence revenue from a completed agent goes to the intern who built it. The remainder covers hosting, support and the next cohort.
How to sit with a client, find the process that is actually costing them, and scope something that can be delivered.
Prompt and tool design, retrieval, evaluation, guardrails, monitoring, and the operational discipline of running something a client depends on.
POPIA, data handling, audit trails and the documentation a client's compliance officer will ask for.
On your systems and your data, so the graduates are trained on the environment they would actually work in.
The twelve-week agentic AI course runs against your processes, your tooling and your governance requirements, alongside a soft-skills course covering client communication, meeting discipline and written handover.
After the course you take the graduates on a three-month contract and see who delivers in your team. Hire the ones who do, with no placement fee for the rest. You assess people on real work rather than on an interview.
Hosting a cohort is a hiring pipeline and an enterprise and supplier development contribution at the same time, and every graduate placed is a first real job in tech for a South African.
75% of candidates were placed with a business inside the twelve weeks of the course over the past year.
Send your CV, or tell us the process you would like an agent to take off your team's hands.
cv@rahn.co.za