iTOO Special Risks, in partnership with global reinsurer Munich Re, has announced the launch of aiSure, a pioneering artificial intelligence (AI) performance insurance solution in the South African market.
The product, developed by Munich Re, is the first of its kind on the African continent and is designed to protect organisations against liability and financial losses arising from errors made by agentic AI systems and large language models and it provides an insurance guarantee for AI developers.
As AI adoption accelerates across industries, from financial services and payments to agriculture, logistics and enterprise automation, businesses are increasingly relying on machine learning models to make or support critical decisions.
Yet AI systems, by their nature, are statistical and nondeterministic, and can “hallucinate”, misclassify, or generate inaccurate outputs, even when functioning as designed. These errors can lead to significant operational and financial consequences.
“AI is not like traditional software, and it doesn’t always produce the same output from the same input, and even the best models carry a measurable error rate,” says Stefano Cavalieri, product lead at iTOO Special Risks.
“aiSure gives organisations a safety net by insuring the performance of the AI models they develop or deploy. If the AI underperforms and causes a loss, the policy responds.”
Unlike cyber insurance, which responds to malicious attacks, data breaches, ransomware, or system compromise, aiSure covers losses arising from the non-malicious failure of an AI system to perform as intended. It also differs from professional indemnity cover, which applies when a human professional makes an error.
“aiSure is specifically designed for autonomous AI decision-making that impacts critical decisions and business processes,” explains Cavalieri. “If a human is still reviewing or approving the AI’s output, like a lawyer checking a draft generated by an AI tool, then that falls under professional indemnity.
“But when an AI model is left to make critical decisions on its own, that’s where this product becomes essential.”
Examples of agentic AI-generated risks include fraud detection models incorrectly approving fraudulent transactions, credit scoring engines misclassifying risk, or AI-driven agricultural systems recommending incorrect inputs. These are not cyber events, but performance failures of the AI itself.
Although Munich Re launched the aiSure brand in Europe four years ago, South Africa and Africa had not previously been a focus market. That changed when iTOO chief executive Justin Naylor identified the opportunity to bring the product to Africa.
“South Africa is at a point where AI adoption is accelerating rapidly. We’re seeing enterprises and developers integrating AI into their systems at a rapid pace, but the regulatory environment is still catching up. That creates a real need for risk mitigation tools that allow companies to innovate confidently,” says Naylor.
South Africa is home to a fair number of software developers, many of whom are now building or integrating AI models. For these developers, particularly startups and scale-ups, aiSure offers a competitive advantage.
“If you’re a small AI vendor trying to convince a bank or enterprise to adopt your model, it’s difficult to prove performance upfront. With aiSure, developers can go to market with a guarantee backed by iTOO and Munich Re. It helps them land deals and build trust,” says Cavalieri.
Naylor adds that the timing aligns with a broader shift in the local market. “AI is the buzzword in every boardroom right now. Companies want to experiment, automate and innovate, but they also want assurance. This product enables that innovation by reducing uncertainty.”