In this follow-up episode of the Pure Digital Passion Podcast, I sit down again with Simon Bransfield-Garth, Founder & CEO of Akili AI, this time joined by Shikoli Makatiani, CTO and technical lead at Akili AI. Our first conversation with Simon explored his remarkable journey from working on artificial neural networks at Cambridge in the mid-1980s, to Symbian, mobile security, off-grid solar in Africa, and eventually the founding of Akili AI. This second conversation moves from vision to deployment. We discuss what it really takes to put AI to work inside Kenyan financial services organizations, including banks, cooperatives, microfinance institutions, customer service teams, and regulated business workflows. This is a deeply practical conversation about shadow AI, agentic systems, AI governance, AI readiness, AI risk registers, pain owners, MVP-first deployment, staff training, and why Kenyan and African organizations need to focus less on AI hype and more on measurable business outcomes. Key themes covered: Why many organizations are already using AI even when leadership thinks they are notWhat “shadow AI” means and why it creates governance and risk exposureWhy AI is not just another digital transformation tool but a decision-layer technologyHow Akili AI identifies practical business pain points before deploying AIWhy “pain owners” are critical to successful AI projectsHow agentic AI differs from chatbots in financial services workflowsWhy AI governance, guardrails, human oversight, and staff training matterHow Akili Snapshot, Akili Assured, and Akili Passport help organizations manage AI readiness and responsibilityWhy “good enough is good enough” when selecting AI models for practical African use casesWhat CEOs, CTOs, CIOs, founders, and risk leaders should do next Timestamps: 00:00 Introduction and episode context01:29 Shadow AI and why unmanaged AI is already inside organizations03:04 Why the value is not in the AI model but in what you build around it04:43 Shikoli on practical AI deployment in Kenyan organizations06:48 Why AI operates at the decision layer of the business08:35 Moving from AI strategy to solving real business problems10:57 Why AI projects are different from ERP-style implementations12:32 Business cases, experimentation, and managing the risk of failure14:05 Why Akili AI prefers MVP-first deployment over big-bang transformation16:55 Reinventing business processes from the bottom up17:58 Building credibility through incremental AI wins19:07 Why Shikoli starts by asking: “Show me the pain owner”20:46 Customer service pain points, 72-hour response times, and 8,000 calls23:59 Moving from first AI projects to third and fourth deployments25:42 Managing scope creep in AI projects28:50 What agentic AI really means31:32 How AI agents can operate in microfinance and cooperative workflows34:33 Guardrails, human oversight, and safety in regulated financial services36:16 Reducing cognitive load for employees and customer service teams37:22 Akili Snapshot and what it reveals about AI readiness40:40 Akili Assured, AI governance, and Akili Passport training42:56 Real-world shadow AI risks inside banking environments45:52 Frontier models, sovereign AI, and why not every use case needs the biggest model48:08 Token maxing, cost control, and African AI pragmatism52:02 Advice for CEOs: start now, assess readiness, solve practical problems53:08 Advice for CTOs, CIOs, and founders on practical AI use cases55:19 Closing thoughts on responsible and practical AI deployment