A Nairobi founder went months without posting. Then he built an AI tool, used it on himself, and published three times in two weeks.
We paid attention.
Then we read the spec. Now we’re building something for you.
Let’s be honest about something.
You know what your business does. You know why it matters.
You could explain it at a pitch, at a funeral, at 2am when nobody asked.
The story is there.
But somewhere between the idea and the published post — between the thing you know and the thing you actually say out loud, on the internet, where people can find it — something dies.
Usually it’s your energy. Sometimes it’s your weekend. Always it’s the post.
Sound familiar? Good. Please stay!
The Problem Has a Name. It’s Called Friction.
Samuel Ralak is a founder who went ages without publishing anything. Not because he had nothing to say. Because the process of saying it was exhausting and half the audience will not understand what he is saying if he had to put it out raw, to the world.
Editing...
So he built Hadithi — an AI-powered writing tool built around his own workflow — and started using it on everything he published. Three posts in two weeks after months of silence.
He calls it dogfooding. Using your own product before you ask anyone else to. Snoop doggy would call it getting high on own supply.
We call it the only honest way to launch anything.
But Here’s What Most AI Writing Tools Won’t Tell You
AI is fast. Genuinely fast.
The problem is that fast in the wrong direction is just a more efficient way to end up somewhere you didn’t want to be.
Most AI writing tools — the generic ones, the ChatGPT wrappers, the tools built for no market in particular — will confidently produce the wrong thing.
Not broken things.
Competent, fluent, publishable-looking things that don’t sound like you, don’t speak to your audience, and don’t carry the weight of what your business actually does.
The 2025 Stack Overflow Developer Survey found that more developers actively distrust the accuracy of AI tools (46%) than trust it (33%).
That gap isn’t cynicism.
It’s pattern recognition.
It’s what happens when people use a powerful tool without the structure to make it work for them specifically.
That structure is what Hadithi was built to solve.
Why This Is a Kenyan Business Problem, Not a Personal One
This isn’t about being lazy or unbothered or too busy. Kenyan founders are none of those things.
This is about infrastructure.
Specifically, the absence of it.
The businesses we cover on Nnalolile are not short of material. They are short of process.
A founder who can pitch their company in a room full of investors — fluently, compellingly, with receipts — often can’t get that same clarity onto a page that Google can find.
Not because they don’t know what to say. Because the path from knowing to publishing is full of friction, and friction, left unaddressed, becomes silence.
Silence is expensive. In a market like Kenya — where trust is still being built between businesses and their audiences, where your digital presence is often the first handshake — silence is a business decision.
Just not one most people make consciously.
The Attention Economy Doesn’t Care About Your Story.
But Your Customers Do.
We’ve written about this before on Nnalolile — the commercial attention economy and what it costs Kenyan businesses who try to play by its rules. Algorithms want speed. Virality. The thing that performs today and means nothing tomorrow.
Your customers want something different.
They want to understand you.
They want to know what you actually believe, how you actually work, and whether trusting you with their money, time, or problem is a decision they’ll regret.
That’s not built in a viral moment. It’s built in the accumulation of content that is specific, honest, and consistent. It’s built through storytelling that doesn’t talk down to the reader or dress up nothing in expensive adjectives.
It’s built through Hadithi.
What Is Hadithi — and Why Is It Built Differently?
Hadithi is Swahili for story.
Hadithi njoo, uongo njoo, utam kolea! — story come, even if it’s a lie, as long as it’s as good as grandma’s.
You already know this.
But Hadithi the tool is not another generic AI writer. Not ChatGPT with a Kenyan flag on it.
Here’s what makes it different: before a single line of code was written, Samuel built Smith — an open-source AI orchestration library that powers everything Hadithi does.
And before Smith was written, he wrote 851 lines of specification for it.
Not as documentation.
As thinking.
As a contract for what the tool would do, what it would refuse to do, and how it would stay honest when the code and the original intent started to drift apart.
That is an unusual amount of discipline for a tool that, on the surface, just helps you publish a post. But that discipline is exactly why it works.
Most AI tools are built conversationally. Describe what you want, iterate, describe again. It’s fast at the start — and then it degrades. Each prompt is a lossy compression of what you actually meant.
By the time you’re three drafts in, the AI has a blurry version of your intent and is filling the gaps with whatever sounds right. The result is plausible.
Not yours.
Hadithi was built to resist that.
Its architecture was designed to keep your voice at the centre — not as a style setting you toggle, but as the actual foundation of how it generates and refines content.
Ralak’s workflow:
- vision before architecture,
- architecture before spec,
- spec before code, and
- a live feedback loop that catches drift before it compounds.
He calls the spec a thinking tool — “a way to catch the places where I thought I understood what I wanted but didn’t.”
That same rigour is what Hadithi brings to your story.
We’ve Already Tested This Logic.
When we covered Peter Mayeku Waste Paper Ltd — a Kenyan SME that turned waste paper into tissue products — the story was never just the factory opening.
It was the circular economy, the NEMA compliance, the #FagiaNairobi cleanup movement that fed the supply chain.
The full context.
The whole human picture.
When we put ChatGPT and Gemini through an East African creative brief — testing both tools not for global benchmarks but for local performance — we weren’t reviewing technology. We were asking: does this work for us?
That is the question Hadithi is built to answer for Kenyan business storytelling.
Not: can AI write?
But: can it write this, for this market, in this voice, for this audience?
We think yes.
The Short Version
You have the story.
You have the business.
You have the thing worth saying.
What you’ve been missing is the infrastructure to say it consistently, clearly, and in a way that compounds over time rather than disappearing into the timeline.
That’s what Hadithi is for.
This piece is the introduction. The next few are the proof. And if you’re a young Kenyan founder, consultant, or business lead who’s ready to stop being the best-kept secret in your industry — the early access list is open.
You know where to join the movement, don’t you?
Hadithi is being built by Samuel Ralak. Nnalolile For Business is looking to bridge as the editorial partner on the product’s public rollout. Smith, the open-source AI orchestration library powering Hadithi, is available on GitHub.
Leave a Reply