There is a particular kind of clarity that comes from having thought about a problem for four decades. Simon Bransfield-Garth is not a convert to artificial intelligence. He is one of its original practitioners — a Cambridge-trained engineer who wrote his doctoral research on artificial neural networks and their application to perception in 1985, at a time when most of the world did not yet have a word for what he was doing. I sat down with him earlier this week and wanted to understand not just what Akili AI is building, but how forty years of thinking about this technology shapes the way we see what is happening right now all over the workd where AI is concerned.

The Long Road To AI As We Know It Today

Simon’s career is one of the more unusual ones I have ever encountered. From Cambridge, he founded Myriad Solutions — a vector processor company that built custom chips to power neural network simulations at a time when no standard hardware was up to the job. Vector processors are what we now call GPUs. He was building the compute layer for AI in 1989. The company was sold in the mid-1990s, and Simon moved on to Symbian — the smartphone operating system that, at its peak, ran on the majority of the world’s mobile phones — where he led marketing through the years that defined the pre-iPhone era.

The Symbian story is one he tells with clarity. Symbian was built on the assumption that the world in the future would look like the world in the past. Apple’s iPhone arrived and demonstrated that it would not. Steve Jobs did not ask customers what they wanted from a phone. He envisioned what the experience could be and built the iPhone on this basis. The lesson Simon takes from it is not about technology at all. It is about what happens when a company optimizes for a future anchored in its experience of the past. It is, he notes, a lesson that is entirely relevant to what he is watching happen in AI governance today.

Then came Africa. In 2012, Simon founded Azuri Technologies — a pay-as-you-go solar company that has now brought clean energy to over a quarter of a million households across nine African markets. What struck me is how he describes it: not as a philanthropic mission, but as a market creation exercise. The key insight was that the cost of the solar subscription was less than the cost of the kerosene lamps it was replacing. From the customer’s perspective, it was not a sacrifice. It was an upgrade. The lesson, which he quotes from Paul Polak, is this: you cannot donate your way out of poverty. You have to create markets.

The Akili AI Thesis

Akili AI was founded in July 2024, and its focus is on something that most AI companies treat as a footnote: governance. Simon’s argument is that the AI itself — the underlying models — are increasingly a commodity. Any one of ten AI labs will produce a result that is good enough for most business use cases. The differentiated value is not in the model. It is in knowing what to do with it, how to implement it safely, and how to ensure that the organization can be held accountable for what the AI does on its behalf.

The entry point for this conversation in Kenya is important as roughly 42 per cent of Kenyan internet users use ChatGPT in a given month — one of the highest adoption rates per capita anywhere in the world. But when Simon walks into Kenyan boardrooms and asks leadership teams what AI they are running, the answer is almost always the same: none. He then starts asking questions. How many of your software products have added AI as an update in the last eighteen months? How many of your employees are using ChatGPT at home to produce work they then submit internally? Does your GDPR policy mention AI? The mood in the room changes quickly. The gap between what leadership believes is happening from an AI adoption perspective and what is actually happening on the ground is, in his experience, almost universal.

What Akili Snapshot Actually Does

This week, Akili AI launched Akili Snapshot — a free, entirely anonymous AI readiness assessment available at assured.akili-ai.com. You answer approximately twenty to thirty questions covering five dimensions: shadow AI exposure, usage policy, staff training, accountability structures, and risk management. The tool produces a maturity score benchmarked against industry data and gives you a radar plot showing where you are relative to where an organization of your size and sector should be. If you put your email in, you receive a full AI-generated governance report with five specific remediation recommendations.

The insight behind the design is interesting. Most companies that want to understand their AI readiness go to a consultant and pay a significant fee for a twenty-page document that tells them what they should be doing. Akili AI has put the equivalent of that first engagement online, for free, with no sales call required. The logic is that if you can show organizations where they are — quickly, without friction, without cost — you create the honest conversation that makes the subsequent work possible.

Akili Assured, the paid governance product that sits behind Snapshot, provides the infrastructure for that work: pre-filled governance document templates, supplier contract AI clauses, an AI risk register, an oversight calendar, a monthly AI insight briefing, and what Simon calls the Assured Agent — an AI-powered tool that manages and schedules the compliance activities the governance framework requires. Akili Passport, launching later this year, adds the training layer: short, role-specific video modules with quizzes and completion certification, designed to ensure that every department head — not just the IT team — understands what the AI in their department is doing and what they are personally accountable for.

Kenya’s and Africa’s AI Opportunity — and the Inherent Risks

Simon’s view of the next decade in Kenya and the rest of Africa has two likely scenarios. The positive scenario is one in which Kenya and African businesses embrace AI, build indigenous applications on top of global AI models, and export those solutions across the continent and beyond. The negative scenario is one in which adoption happens slowly, and the value created by AI is extracted by the handful of companies — mostly in Silicon Valley and China — who build the underlying platforms. We have seen that pattern before in Kenya and the rest of Africa. It’s not a great outcome!

His advice to business leadership teams is highly insightful: do not try to boil the ocean. Do not commission a five-year AI strategy. Find the specific business problem that needs solving, solve it, and find the next one. He gives the example of what Akili AI did for a microfinance client: they took a loan processing workflow from two weeks down to five minutes. That is the kind of return that pays for itself within the same financial year. That is the starting point.

The Most Important Insights 

The biggest takeaway I got from this conversation is the quality of the long view. Simon has been watching AI go through waves — enthusiasm, disillusionment, resurgence — since before most of the people writing about it today were born. His practicality is not cynicism. It’s the confidence of someone who has seen enough cycles to know that the question is never whether the technology works. It is whether the organizations deploying it understand what they are actually doing.

Kenya and Africa are, right now, some of the most interesting places in the world to be having this conversation. The consumer adoption of AI is confirmed. The enterprise governance is not. The gap between those two things is where Akili AI is positioned. Whether organizations move to close it — and how quickly — is the question that will define the next decade of African technology.

If you want to see exactly where your organization stands on AI, take the free Akili Snapshot assessment at assured.akili-ai.com. It takes as little as five minutes. It is entirely anonymous. It may tell you things you really need to know and take action on from an AI Governance perspective as soon as possible since knowing is what matters the most. 

Watch The Podcast:

Listen To The Podcast: