08. The Sparkling Water Problem

There was a machine at my hotel in Paris.

It was at the breakfast station, just sitting there among the pastries and the fruit, and it dispensed sparkling water.

Not from a bottle.

From itself.

You pressed something and out came cold, carbonated, lightly flavoured water that tasted like someone had finally figured out what water was supposed to be.

I stood there probably longer than was reasonable for someone who was supposed to be getting breakfast.

By the time I got back to Lagos I had already started researching the machine.

Found it.

Ran on cartridges, CO2 I think, each one lasting a certain number of uses before needing to be replaced and imported again. I did the maths on sustainability and didn't like what I found. The idea went into a drawer.

But it was never really about sparkling water.

Sparkling water barely exists in Nigeria outside of San Pellegrino and Voss, both imported, both priced accordingly. If you wanted to make it locally and affordably you would need either the machine and its cartridges, or glass bottles and a production process that actually made sense here.

I had already started thinking about branding before I remembered I had three businesses and a website developer waiting on my feedback.

The drawer it is.

But that process, the seeing and the asking and the mapping and the almost, that is just how my brain works.

I do not need to be in another country for it.

I just need to see something working somewhere and I immediately start asking whether that same problem exists somewhere else, and what the solution would need to look like to actually fit there.

I tried this once properly. It did not go the way I planned.

Before I relocated back to Nigeria I had been living in Spain, and I had a plan B.

A logistics app.

I had seen how aggregation worked elsewhere, bring the dispatch riders onto one platform, create visibility, create accountability, stop the chaos.

I understood the problem. I found a developer. We worked through milestones together and at the end of it he handed me a file I could not use.

Completely unusable.

Everything we had agreed, everything I had paid toward, gone.

I came back to Nigeria with no plan B and no app and a very specific kind of anger that sits in your chest for a long time.

So I started anyway.

That was Sendieco.

And what I found was that the problem I had identified was real!

The pain point was everywhere.

Back then if you needed something picked up and delivered you had to personally know a dispatch rider, and even then you could not trust the timeline. The riders were not exactly lying. They were pooling.

Your package got picked up and then merged with other packages and redistributed among riders and the person who collected from you was not necessarily the person who delivered to someone else.

You had no visibility into any of it. You could not follow up because the rider did not belong to you, belonged to whoever ran the operation, and asking "where are you?" felt presumptuous and got you nowhere anyway.

I knew this was a problem worth solving.

What I did not know was whether the conditions for the solution existed yet.

And?

No. They did not. Nobody wanted to aggregate. Every business with its own riders wanted to keep their own riders.

The trust infrastructure that makes a platform work was not there. I was trying to build a bridge before either bank had decided it wanted to connect.

So I put up a post on LinkedIn.

"Help me find my partner."

I spoke to two people.

One was connected to Mano, which exists now and is doing reasonably well, and only took about eight years to get there. The other one presented himself very well, said all the right things, and then when it was time to actually sit down and work, he disappeared. First the excuses, then nothing. MIA. Something happened, something came up, and then silence.

That experience did something to my confidence in the available talent that took a long time to repair.

Not because of the app, but because of the pattern.

People presenting as capable and then evaporating the moment the real work begins. I have seen it enough times now that I check for it early.

But arriving too early to a market is its own kind of research.

Now I watch InDrive and Bolt and the hire purchase schemes that gave individual riders their own bikes and technically their own independence and I see something assembling itself slowly in its own Lagos way. The problem I saw in 2017 is being solved. Just not by me. Just not yet.

Electric vehicles are popular in Nigeria now. Someone was promoting them on a page I came across, importing them from China, and from the quality of the content you could tell he genuinely wanted people to understand what EVs were and eventually buy one.

I am eyeing one myself so I was reading through the comments.

Someone posted that electric vehicles emit radiation.

The page owner responded gracefully. He said the radiation levels are very low, nothing to be concerned about. The commenter came back. "What about the battery fires," he or she said. "Those fires are incredibly difficult to control." The page owner responded again. These fires are rare, he said. They should not be a focus.

I watched this exchange and I was not angry at the commenter. I was curious about him. He had not invented that fear from nothing.

He had encountered something, a headline or a video or a half-heard conversation, and he had translated it into a conclusion that felt logical from where he was standing.

This happens every time something new lands somewhere that was not part of its original design. The thing arrives. The understanding of the thing lags behind. Someone fills the gap with what they have, which is usually not enough.

This is what I keep thinking about when people ask me what I think about AI in Africa.

It's not whether the tools work. They work.

It's not whether people are using them. Everyone is using them. For therapy, for photo editing, for generating prompts to generate other prompts.

What I do not yet see enough of is people asking what problem we are actually solving. Not what the tool does.

What the underlying need is, whether that need exists here in the same form, and what it would have to look like to work in this context rather than the one it was designed for.

I think I have a feel for that question.

I have been in enough rooms in enough countries to know when something is being copied at the surface and when someone is asking what is underneath it.

I know what it feels like to arrive too early.

I know what it costs when the ground is not ready.

I know what it looks like when a platform that took eight years to find its footing finally starts making sense.

Whether that makes me qualified to speak about AI in Africa specifically or just someone who has been paying attention, I am genuinely still working out.

But I know what I am looking for. And I know how to wait.