Zabble Qualifies for the IOSCO-FCA TechSprint 2026, One of Just Two or Three South African Teams
Out of 67 applicants worldwide, 16 teams made the IOSCO-FCA TechSprint 2026, and only two or three of them are South African. Zabble is one of them, and we are taking an auditable-AI answer to Madrid in October.
By ZabbleSouth African consultancy building bespoke operational systems
ByZabble
Reviewed20 August 2026grounded in 2 primary sources
Zabble has qualified for the IOSCO-FCA TechSprint 2026, advancing from 67 applicants to the final 16 teams worldwide. We are one of just two or three South African teams at this level. Our build tackles how retail investors can learn to invest with AI safely, and the finals take place in Madrid this October.
Zabble has qualified for the IOSCO-FCA TechSprint 2026. Out of 67 applications from around the world, 16 teams were selected, and we are one of them. We are also one of just two or three teams from South Africa at this level.1 The original announcement is on our LinkedIn post.
What is the IOSCO-FCA TechSprint?
The International Organization of Securities Commissions (IOSCO) is the global body of securities regulators. In 2026 it launched its first TechSprint, "Investor Education in the Age of AI", delivered with the UK Financial Conduct Authority's AI Lab. The programme brings regulators, technologists and educators together to build practical, scalable tools that protect and educate retail investors as AI reshapes how people learn about money. It runs as a cohort-based, largely virtual programme and culminates in an in-person Demo Day on 8 October 2026 in Madrid, hosted at the margins of the IOSCO C8 Plenary.2
Why this matters for South Africa
Applications came from across IOSCO's global membership, and 67 teams put their hands up. Sixteen made it through. Only a handful of those, just two or three, are South African, and Zabble is one of them.1 We are working alongside the Financial Sector Conduct Authority (FSCA) and Khula Capital, with guidance from Raveen Ramlakan CFP, and we intend to represent the country properly in Madrid.
The problem we are taking to the TechSprint
There is now a generation of people for whom AI is the first, and often the only, place they turn for guidance about money. That is not a future risk, it is already how the next wave of retail investors learns. The tools they use can hallucinate, rarely explain themselves, and offer little transparency about where an answer came from. Our challenge is to make AI a safe and genuinely useful learning tool for those investors.
What we are building
Our submission has three parts:
- Auditable AI infrastructure. The technical layer that makes an AI answer reliable and traceable: where it came from, what it is based on, and whether it can be trusted. It is the same citation-first thinking behind our knowledge assistant and our audit trails work.
- A behavioural journey. Information alone does not make a competent investor. The second part moves a user from reading answers to demonstrating competence, step by step.
- A proposed IOSCO standard. A framework we describe as a nutrition label for AI in the financial sector: a simple, visible declaration of what an AI tool does, what it draws on, and where its limits are.
"AI is already sitting in the pocket of the next generation of investors. Our job is to make sure it earns their trust, and that the guardrails exist when it does not."
- Zabble team, TechSprint announcement
What happens next
The cohort phase runs through the coming weeks, and the finals take place in Madrid on 8 October 2026, where the selected teams present their prototypes to a global audience of regulators.2 We will share the journey as it unfolds on LinkedIn. The thinking we are taking to the TechSprint, evidence-first AI with a real audit trail, is the same thinking we build into the analytics systems our clients run every day.
Frequently asked questions
- What is the IOSCO-FCA TechSprint 2026?
- It is the first TechSprint run by IOSCO, the global body of securities regulators, delivered with the UK FCA AI Lab. The 2026 edition, Investor Education in the Age of AI, asks teams to build tools that help retail investors learn and stay safe as AI changes how people get financial information.
- How many South African teams qualified for the TechSprint?
- Only a handful. Of the 16 teams selected from 67 applications worldwide, just two or three are South African. Zabble is one of them, which makes the selection a milestone not just for our team but for the South African fintech community.
- What is Zabble building for the TechSprint?
- A three-part solution: technical infrastructure that makes AI answers auditable and reliable, a behavioural journey that moves a retail investor from information to demonstrated competence, and a proposed IOSCO standard we describe as a nutrition label for AI in the financial sector.
- When and where are the TechSprint finals?
- The programme culminates in an in-person Demo Day on 8 October 2026 in Madrid, hosted at the margins of the IOSCO C8 Plenary, where the selected teams present their prototypes to a global audience of regulators.
- Who is Zabble working with on this project?
- We are working alongside the Financial Sector Conduct Authority (FSCA) and Khula Capital, with guidance from Raveen Ramlakan CFP. The FSCA is the South African market conduct regulator, which keeps the work grounded in the realities of our own market.
Sources
- LinkedIn - Zabble: IOSCO-FCA TechSprint 2026 selection announcement (2026).Zabble advanced from 67 applications to the final 16 teams in the IOSCO-FCA TechSprint 2026, one of just two or three South African teams at this level.
- IOSCO - IOSCO TechSprint: Investor Education in the Age of AI (2026).The inaugural IOSCO TechSprint, powered by the UK FCA AI Lab, focuses on investor education in the age of AI and culminates in a Demo Day on 8 October 2026 in Madrid.
Keep reading
The company playbook stops being a folder no one reads, it becomes the system that answers questions for the whole team.
The submissions regulators, auditors, donors, and boards expect, assembled from the data you already generate, no quarter-end scramble.
Background monitoring across a stream of activity no human could realistically watch, only the things that matter surface, with their evidence already attached.