What's Next with Aki Anastasiou

The episode contains a handful of useful data points and a three-part failure taxonomy for AI projects, but most of the runtime is spent on high-level framing and reassurance rather than dense, actionable ideas. The ratio of novel claims to padding is moderate at best.

“statistically in this last year, based on surveys, uh, data that we have of the AI use cases, typically at the moment we're only seeing about 11% of AI deployments being successful”

“there's a human element when we ignore the user experience that employees have. There's a lot of times that employees just won't use the AI”

Every major argument here - Africa's leapfrog opportunity, AI pilots failing due to poor data and tech-first thinking, CIOs needing to be more strategic - is well-worn in B2B technology circles. There is no contrarian position, no first-principles challenge to received wisdom, and the closing advice ('stop believing in limitations') is motivational rather than analytical.

“I think in the past we've always been talking about the digital divide and how, you know, Africa needed to spend lots of, lots of uh, economic money behind digital infrastructure”

“what problem are we trying to solve for”

Naidoo holds a legitimate senior leadership role (CIO Program Leader for Africa at Deloitte) and demonstrates genuine familiarity with the African market context, making her relevant and credible. However, she is on the consulting and advisory side rather than an operator who has personally built and scaled AI systems, which limits the depth of practitioner-specific insight.

“The Deloitte Trends 2026 report is 17 years in the making”

“in the Deloitte world, we call it this need for having four faces”

The 11% AI deployment success rate is the only hard, sourced statistic in the episode. A few named technologies (Starlink, M-Pesa) and sectors (fintech, financial services) are mentioned, but there are no named client examples, dollar figures, timelines, or case study detail. The cybersecurity section is entirely abstract.

“typically at the moment we're only seeing about 11% of AI deployments being successful”

“we've seen M Pesa being successful in the past, but we starting to see the fintech industry really start to take off”

The host asks structured, topically relevant questions and lands a solid closing 'stop/start' prompt, but consistently affirms rather than challenges ('you're 100% right,' 'absolutely fascinating'), never follows up on the 11% statistic or presses for a concrete African example of success, and allows all claims to pass unchallenged.

“Look, it's absolutely fascinating and you 100% right”

“But when you talk to CIOs and C level suites executives, those pilots, those AI pilots are all over the place. But the real returns are quite rare”