Africa Pre-Mortem · Bern
Find out
why it fails.
Before it does.
Your team loves the idea.
Hire us to argue against it.
Before money goes into an African market, we rank every way the plan can fail. CHF 1,900, in five business days.
Investors welcome. NDA on request. A limited number of engagements at a time.
Some of them sink.
The Africa Pre-Mortem
Eighteen months from now,
the project has failed.
Why?
Nobody in the room will say it. We write it down while the budget is still yours.
You get each likely failure ranked by damage, the cheapest way to test it, and our verdict.
In writing, in five business days.
A savings app
for traders in Douala.
Cameroon · Pre-Mortem · Sample
- 01 · Trust
- Traders already save with a tontine treasurer they know by name. Test: ten conversations with tontine members before any build.
- 02 · Distribution
- Mobile-money agents earn more selling airtime than signing up users. Test: compare agent commissions.
- 03 · Unit cost
- Fees on small deposits erase the margin. Test: a one-page unit-cost model.
- Verdict
- Change. Partner with the tontines.
Where we come in
Start with the problem.
The tools come later.
Most of our work starts with something small and expensive: a form typed in twice, or an address the software cannot read.
Useful is a human thing
The water came in.
The stall is still open.
Floods, power cuts, and customers who trust a person before an app. Software that ignores this fails quietly, and people pay for it.
We bring AI in from day one and keep it where it earns its place. Often the better answer is a cleaner database or a simpler rule.
How we workWhen the idea survives
Before the build,
know what is worth building.
The Africa AI Feasibility Assessment turns one workflow in one country into a decision you can defend. Building comes later, at CHF 300/hour, and only if the case holds.
In writing, in ten business days.
Can we stop retyping
every service request?
Cameroon · French and English requests · Human review
- The question
- Can a tool draft a usable record without inventing what is missing?
- The evidence
- Sample requests you may share, a reviewer, and a definition of a usable record.
- The boundary
- Draft only. A person confirms unclear locations and dispatches.
- The decision
- A narrow trial, better inputs first, a simpler system, or no build.
Before you sign, before you scale
Two more questions
worth asking early.
Adoption Fit Study
It works in the demo. Will people use it? We look at trust, language, habit and the workaround you are replacing.
Will they use it? 02 / Before you sign · CHF 1,200Proposal Second Opinion
A vendor promises AI, scale and local fit. We find the unsupported claims, the year-two costs and the lock-in.
Read it before you signBeyond the last tower
Most of the continent
is not a capital city.
Past the last paved road, most software assumptions stop holding. We design for the places the coverage map leaves out.
Question it.
Take it apart.
Build it better.
Founder, OSIH · Bern
The person behind the work
You speak to the person
who builds.
Whoever reads your brief does the work, through to any build. OSIH runs alongside other work, so the client list stays short.
Our integration depth is in Cameroon, with a network across Nigeria, Zambia, South Africa, Kenya, Côte d’Ivoire and Senegal.
Francis Osih on LinkedInWe build what we advise on
Things we have
put into the world.
Our own products, built and run by OSIH.
Custom work
What is everyone
sure will work?
Send us the plan. If it fits, you get a fixed proposal within two business days.
Send your brief