One continent, three very different buying markets
A lot of procurement teams treat "Africa" as one line in a vendor spreadsheet. That's the first mistake. Buying AI training for a workforce split across Lagos, Nairobi, and Johannesburg is buying in three markets with different currencies, connectivity realities, funding mechanisms, and talent pressures. A vendor that's excellent in one can be wrong for another.
The pan-African upskilling market sits inside an e-learning sector growing from roughly USD 3.68 billion in 2025 toward USD 7.77 billion by 2034, with corporate learning the fastest-growing slice. That growth pulled in a crowd of vendors, from continental players like ALX and Andela to local specialists like Utiva, Gomycode, Moringa School, and Decagon. Sorting the real ones from the slideware is what this guide is for.
Related reading: Nigeria's Corporate Training Boom in 2026 · South Africa's Enterprise AI Upskilling Reset · Nairobi's Silicon Savannah in 2026.
The currency question comes first
Before pedagogy, before content, ask how the vendor prices and in what currency. This has sunk more African training contracts than bad content ever did. A dollar-denominated contract with rigid terms looks fine at signing and can become punishing within months if the naira or cedi moves. In Nigeria especially, vendors who priced in dollars mid-2024 lost accounts to local players who never carried the currency mismatch.
What to nail down in the contract:
- Currency of the quote and whether it's held fixed for the term or floats.
- Whether local-entity invoicing is available, which matters for tax and forex controls in all three markets.
- How renewals reprice. A vendor who quietly resets to spot rates at renewal is a vendor you'll be replacing.
- In South Africa specifically, whether the spend qualifies for SETA offset and feeds the B-BBEE scorecard.
None of that is glamorous. All of it decides whether the contract is still viable in month nine.
Delivery: build for the phone that's actually in the room
With over 500 million smartphone users across Africa and most enterprise learners doing their training on mid-range Android devices over metered data, delivery is not a footnote. It's a pass-fail gate. I've watched shortlisted vendors get cut for a single technical detail, an app that re-downloads video every time a half-finished module reopens, which on a field team's data plan costs more than the training.
Put these in your RFP as hard requirements, not nice-to-haves:
| Requirement | Why it matters | How to test it |
|---|---|---|
| Low module weight (MB) | Learners pay for their own data | Ask for the average; if unknown, they never measured |
| Real offline mode | Connections drop mid-module constantly | Test on a Tecno/Infinix, not the demo iPhone |
| Reliable progress sync | Lost progress kills engagement | Kill the connection mid-module, reopen, check |
| Localisation | Swahili, isiZulu, Nigerian English land differently | Ask what's actually localised vs. auto-translated |
Content vs. outcomes: the pivot every buyer should make
The vendors still leading with "4,000 courses" on slide one are selling the last decade's product. A course count is an input. Your CFO buys outcomes. The pivot in your evaluation is to stop scoring library size and start scoring whether the vendor can tie training to a metric you already track.
The single best procurement question, and the one that thins a shortlist fast: "Show me a client who looks like us, and the operational number that moved after your programme." A real vendor has that story with a name and a figure attached. A slideware vendor has testimonials about "engagement" and "empowerment." One of those is evidence. The other is decoration.
Insist on a paid pilot with a baseline
Never sign the enterprise licence first. Run a paid pilot, 60 to 90 days, on one department, with a number agreed before it starts. The discipline that makes this work is the same in all three markets:
- Pick one team of 25 to 40 with a weekly metric you already measure.
- Baseline that metric for four weeks before training begins.
- Hold a control group of comparable staff who train later.
- Measure at 30, 60, and 90 days, and expect real behaviour change around week eight, not week two.
- Convert the trained-versus-control gap into money. That's your rollout case.
A vendor confident in their product will take a paid pilot on those terms happily. A vendor who pushes hard to skip straight to the annual licence is telling you they don't back their own outcomes. Listen to that.
A scoring sheet you can actually use
Boil it down to a weighted sheet so procurement and L&D score the same things. My rough weighting for a 2026 African enterprise buy:
- Outcome evidence (a real client, a real moved metric): 30%
- Delivery fit for mobile and low bandwidth: 20%
- Commercial terms and currency safety: 20%
- Localisation and regional presence across your actual sites: 15%
- Pilot willingness and measurement rigour: 15%
Notice content library isn't a line. If you want it as a tiebreaker, fine, but it shouldn't carry weight against outcome evidence. And weight the last item more heavily than it looks, because a vendor's willingness to be measured is the best single predictor of whether they'll deliver.
Buy for the workforce you have, on the devices they carry, in the currency that won't ambush you, with a number you agreed up front. Do that and the crowded African upskilling market gets a lot easier to work with. Skip it, and you'll be running this procurement again in a year.