“We didn’t just need more data. We needed smarter data.” – Muzi Dlala, Sasria
In Episode 5 of TIA Talks, host Jason Mizen is joined by Muzi Dlala, Executive Manager at Sasria and guest judge on TIA 2025, along with Evashen Mooninathan, the show’s third evictee—whose departure shocked both contestants and viewers alike.
What followed was a deep, technically rich and emotionally honest discussion that unpacked the challenge at the centre of the episode: how to prevent another national catastrophe like the July 2021 unrest, using intelligence, community-driven data, and systems thinking.
This wasn’t just a task—it was a real-world crisis simulation. And the model behind it? Already built. Already tested. Already live.
The brief: can unrest be predicted?
As Muzi Dlala explained, Sasria’s involvement in the July 2021 unrest was more than just about claims—it was a wake-up call.
“We did a lot of research to understand why this happened,” Muzi explained. “Even the state admitted to not having enough intelligence. So, intelligence became the big buzzword.”
That’s when sentiment analysis became part of the conversation: a digital tool used to gauge the emotions and intentions of people by analyzing language across various data sources—social media, app reports, even USSD.
But as Muzi noted, social media alone was never going to be enough. “In South Africa, a lot of people don’t have smartphones or internet access. We had to think beyond that.”
Four levels of defence: a systems approach
In response, the Sasria team developed a framework based on four layers of defense in society, drawn from catastrophe response models:
- Individuals – People have a natural instinct to protect their property.
- Communities – Often underestimated, but hugely powerful in crisis moments.
- Police – Formal enforcement, requiring real-time data to deploy effectively.
- Army – The final line of response when escalation reaches a national level.
“We found that community defense, like the one at Maponya Mall, had far more impact than we expected,” said Muzi. “The taxi industry took to the streets to protect their areas. That told us something—we had to account for them in our risk model.”
From data to intelligence: triangulation is key
The model’s real innovation lies in its triangulation of multiple data sources to build an early warning system with up to 90% accuracy. This includes:
- Social Media Monitoring
- USSD Codes for low-tech reporting
- App-based push/pull data
- Supply chain monitoring via Consumer Goods Council partnerships
- Historical protest mapping and migration patterns
- Claims history overlays
“We used historical data to understand where protests start, how they move, and what routes they take—like from Alexandra to Sandton,” Muzi shared. “When you overlay that with claims history, you see exactly where the pressure points are.”
From model to framework
The framework, built over months by experts and data scientists, now forms the backbone of Sasria’s real-world early warning system.
“If we have the intelligence,” Muzi said, “we can notify a policyholder in Sandton today that tomorrow at 9 a.m., a protest is likely to pass through their area. We can say: protect your property.”
“We can’t stop the protest. People have a right to protest. But we can do something about the destruction of property.”
Contestants vs. The clock
Evashen Mooninathan, who participated in the challenge and was eliminated, reflected on the experience.
“We considered so many components, but we just couldn’t bring it all together in time,” he admitted. “We focused too much on the context and didn’t get to really explain the solution.”
Host Jason Mizen acknowledged the complexity of the task:
“You had two hours. That’s hardly enough time to build an entire system. But even with that, some of the thinking you all brought forward was incredibly close to the real model.”
“It took a few PhDs and a few months to build the original,” Muzi added. “So for contestants who aren’t risk modelers or systems architects to come up with what they did? That’s impressive.”
Jason closed out the episode by reflecting on the importance of bringing the public into these conversations:
“It’s one thing to talk about policy and risk in boardrooms. But bringing it to a national stage, with real people, real stakes—this is how we change perceptions and build trust.”