The Rock and the Hill

The founder describes a Greek story.

A man pushes a rock up a hill. He gets close to the top. The rock rolls back. He starts again. That was the Farm Trust experience.

Farm Trust was his first product under Casa Management Solutions. The concept was sound.

A central nervous system for African agriculture, connecting farmers to market prices, to distribution networks, to localised AI-generated advice in their own language, delivered each morning to their phones.

The technology worked. The product was built. However, the demand was not there in the form required.

Moreso, the farmers who would benefit most from the product could not afford a subscription.

The organisations that could afford it, agricultural input companies, NGOs (Non-governmental Organisations), development bodies, were distracted by other priorities. An ongoing conflict had tightened budgets. Sales cycles stretched. The rock rolled back.

Rather than pushing again, the founder looked inward.

He asked a simpler question. What problem do we have ourselves, right now, that we could actually solve?

The answer was a coding agent. Specifically, a good, affordable one that did not require an American bank card or a $100 monthly subscription to access meaningfully.

That question, turned inward, became BWAT.

---

The Core Problem

The AI boom is real, but the access is surprisingly not.

The frontier models, the genuinely powerful ones that can build software, analyse data, and solve complex problems, are gated behind three barriers the founder names precisely.

1. Language barriers, for populations whose first language is not English and whose local context is not represented in training data.
   
2. Connectivity barriers, for populations where consistent internet access is not guaranteed.
   
3. Economic barriers, for populations where $100 a month represents more than a month’s income for the majority of people.
   

The addressable market for frontier AI tools, priced for Western purchasing power, is a fraction of Africa’s actual population.

This is not primarily a cultural problem or a digital literacy problem. It is a pricing architecture problem.

The tools were built and priced for markets where they make economic sense. Africa was not those markets.

BWAT was built specifically for the gap between the frontier models that exist and the African users who cannot reach them.

---

The Strategic Decision Layer

The most interesting product decision in this case is where BWAT chose to enter.

Developers are the fastest AI adopters globally because the technology was originally built by developers for developers.

In Uganda specifically, and across East Africa, there is a growing community of software developers, students, and technical professionals who understand what a coding agent is and who have immediate, practical use cases for one.

They also have the baseline digital literacy to adopt a VS Code extension without significant onboarding friction.

Find the users who already understand the category, prove the product with them, accumulate usage patterns, iterate based on their feedback, and then extend to the broader population who need a different form factor entirely.

The broader population the founder describes for phase two is genuinely different.

These are users who have never opened VS Code, who may not know what a browser extension is, who may not type fluently, but who could absolutely benefit from the core capability of conversational AI if it arrived through a medium they already understand and trust.

The planned voice call feature, where any person can pick up their phone, call BWAT, and speak to it directly, is the form factor designed for this population.

That two-phase architecture, developers first, general public second, is a distribution strategy as much as a product roadmap.

The developer community generates word of mouth, usage data, and subscriber revenue that funds the harder and more expensive work of building for the general public.

The founder’s seven years of ERP and enterprise software delivery for clients before BWAT launched is not background noise. It is the technical foundation that makes BWAT credible rather than aspirational.

---

Ecosystem Context

Uganda currently has almost little to none AI regulatory framework.

The founder does not present this as an opportunity. He presents it as a problem with a specific implication.

Countries develop AI governance frameworks when they have something domestically significant enough to govern.

France developed a more sophisticated AI framework than Uganda because France has Mistral AI, a company doing impressive enough work that the government cannot ignore it. Uganda has not yet produced a company at that level of visibility.

The founder’s position is direct. He wants to be the reason Uganda develops an AI governance framework.

Not because regulation is inherently good, but because being big enough to attract regulatory attention is itself a measure of significance, and being invited into the rooms where those policies are made is more valuable than being affected by policies made without your input.

The absence of significant African AI infrastructure more broadly is something the founder names with genuine frustration.

He cites Sunbird AI, a collaboration with Makerere University producing local language voice models, as the most impressive African AI contribution he is aware of.

He wants to change that.

BWAT is described explicitly as his contribution to that change. Not as a market play primarily, but as a mission.

---

Observed Patterns

The founder describes the build-before-validating-demand failure, spending a year on a project before confirming that paying users wanted it, as his biggest mistake.

The lesson he draws is precise. Evaluate demand first. Not generally, but specifically with people who will part with money.

That principle, applied consistently since, has produced BWAT, which emerged from validated internal demand.

Failure is data. The response to the failure is what matters. That reframe has shaped how Casa Management Solutions runs its product iteration process.

The founder’s recommendation for developers learning their first cross-platform language is also a specific and technically credible recommendation.

A full-stack developer who has built enterprise ERP systems, LPG tracking infrastructure, and a VS Code coding agent recommending Flutter for its universal deployment capability across Android, iOS, desktop, and smart TV is not general advice.

It comes from practical experience of what cross-platform development actually costs in terms of time and context-switching.

---

Open Variables

The founder describes a specific, time-bound milestone, raising user numbers to a defined threshold to qualify for or maintain access to a VS Code extension programme.

Even though the exact nature of this deadline is not yet specified however, the urgency is explicit.

Bootstrapped growth without a marketing budget makes hitting that threshold genuinely challenging.

Offering the most powerful and most expensive-to-serve AI agent on the free tier has driven meaningful user acquisition and daily active user growth.

It has also created a burn rate that subscriber revenue alone does not cover. When the promotional period ends and the most capable features move behind the paid tier, the revenue trajectory should improve.

Whether the user base built during the free period converts to paid in sufficient numbers is the commercial question the next six to eight weeks will begin to answer.

Building a VS Code extension for technical users is a solved distribution problem. Building a voice-based AI interface for users who have never interacted with AI, accessible via a standard phone call, is a fundamentally different product with different trust, language, and infrastructure requirements.

How that transition is resourced, sequenced, and funded will determine whether BWAT achieves its broader accessibility mission or remains primarily a developer tool.

---

Why This Matters

For founders building developer tools in emerging markets, this case makes a specific and underappreciated argument.

The fastest adopters of any new category are the people who understand the category most intimately.

In AI, that is developers. Building for developers first is not a narrow strategy. It is a staging strategy.

The developer community provides the usage, feedback, and word-of-mouth infrastructure that funds and informs the expansion to less technically literate audiences.

For investors, the BWAT case presents a founder with seven years of enterprise software delivery experience, a demonstrated ability to pivot decisively when demand fails to materialise, a live product with growing daily active users three months post-launch, and an accessibility thesis that addresses a structural market failure rather than a niche opportunity.

The bootstrapped, pre-seed positioning, combined with the time-sensitive VS Code programme milestone, creates a specific investment window where modest capital would have disproportionate impact on trajectory.

---

Final Strategic Takeaway

Most founders either have the vision without the self-awareness, or the self-awareness without the vision. The ones who have both tend to build longer and adapt better.

BWAT is three months old at the time of our conversation with the founder.

The founder is obsessed with the product, he is iterating daily, building the team slowly but correctly, and doing all of it in a country with no AI governance framework because he intends to be the reason one gets written.

Baby steps. But in the right direction. And with a clear enough view of the destination that the steps are unlikely to go sideways.

---

This article is drawn from an in-depth founder interview conducted by Afriq IQ with Simeon Mubiru, founder of BWAT. Selected insights and observations are published here.