The application of AI in the country’s healthcare industry means faster insights, more personalised services and solutions that reflect the realities of the industry’s system, says InterSystems South Africa, the local branch of US-based data technology provider InterSystems. However, cost, complexity and value-add in real-time are challenges that need to be overcome to realise benefits.
At the company’s flagship READY 2025 Summit, hosted in Sandton this week, executives offered an overview of its AI strategy, underpinned by what the company describes as its “secret sauce” – a fast, flexible common data plane.
The company's strategy is to make it simpler and faster for specific types of businesses to use powerful, real-time AI.
InterSystems plans to do this by creating pre-built, industry-specific toolkits rather than making each customer build everything from scratch.
A key part of this strategy is to address the real-time challenges of AI in electronic health records, including complexity, regulation, clinical safety and cost.
Dr Rami Riman, director of clinical and business improvement at InterSystems, reflected on the impact of these challenges in hampering successful application of GenAI in healthcare. He said 30% of GenAI projects are abandoned, and 50% of projects reach production in eight months.
This is what the company’s AI strategy is looking to address.
MGAIC affiliation
InterSystems underlined the significance of its affiliation as a founding member of the Massachusetts Institute of Technology (MIT) Generative AI Impact Consortium (MGAIC), a membership organisation that brings MIT researchers together with industry leaders to enhance and advance AI in key areas, including healthcare, business and education.
Its objective is to investigate how GenAI can be used to develop solutions for real-world challenges.
AI-ready data
The company emphasised the importance of AI-ready data, not only in terms of completeness and cleanliness, but the ability to feed AI systems with data that can be immediately checked and processed – and quickly retrieved for use.
InterSystems demonstrated how its InterSystems IRIS data platform is evolving to meet this reality, with recent 2025 updates adding native vector search capabilities. This means organisations can now perform semantic queries and retrieval-augmented generation directly within the InterSystems IRIS data platform, unlocking faster insights from both structured and unstructured data without moving information to external systems.
In healthcare specifically, explainable AI use cases highlighted how vector search and the InterSystems IRIS data platform for health can surface similar patient histories, provide transparent recommendations and support clinicians in delivering more personalised care.
“Data is only powerful when it is usable. With vector search and generative AI now built directly into the IRIS data platform, our customers can move from simply storing information to actively applying it in ways that transform care,” said Gokhan Uluderya, director of product management – data platforms at InterSystems.
Mike Fuller, regional director of marketing at InterSystems, added: “We are moving into an era where applications must think and adapt just as quickly as the people who use them. By combining interoperability, AI and advanced data tools within the IRIS data platform, we are helping our community build a new generation of applications that are faster, more intelligent and designed for real-world impact. In Africa, this agility is vital as it means healthcare, finance and public services can innovate without being held back by legacy constraints.”
Henry Adams, country manager for InterSystems South Africa, said: “The conversations we have had in both Cape Town and Johannesburg showed how Africa can lead with innovation when we put data to work in the right way. From using generative AI with vector search to build the next generation of applications with advanced UX, to focusing on creating solutions that make a measurable difference for users and organisations.”
Share