NowNow's South African AI Adoption Limitations: SMEs and AI Hallucinations
Summary
- Lars Gumede launched NowNow, an applied AI startup for SMEs, aiming to provide access to expert knowledge and tools.
- The platform has over a thousand users and several partners signed up to grow the platform.
- NowNow's approach focuses on grounding AI in locally relevant data to ensure accurate answers.
- Concerns about 'hallucinations' - AI confidently giving false answers - are growing, particularly for business owners relying on AI-generated information.
- Lawyers should be aware of the potential risks associated with relying on AI-generated information and inform clients accordingly.
What Happened
AI offers a leapfrogging opportunity for not just the country, but the whole continent.
Lars Gumede, a University of Cape Town alumnus, has launched NowNow, an applied AI startup that streamlines business tasks for small and medium enterprises (SMEs). The platform aims to provide SMEs with the power of large companies and teams by leveraging artificial intelligence. Currently, NowNow's first version is live, pairing local knowledge with real-world tool integrations to act as an expert assistant for local business owners.
Gumede's vision is to create a South African AI giant that can put the country and continent in the AI race. He believes that AI adoption across the continent is still low by global standards, primarily driven by chatbot tools and large language models like ChatGPT and Grok. NowNow represents the next stage in the AI chain, focusing on applying AI to support SMEs.
The platform has already gained traction, with over a thousand users and several partners signed up to grow the platform.
Legal Context
As AI adoption increases, concerns about 'hallucinations' - AI confidently giving false answers - are growing. This is particularly relevant for business owners who rely on AI-generated information to take action in the real world. To mitigate this risk, maintaining a large knowledge base of local information and data is crucial.
Gumede emphasizes the importance of exercising due care when using AI, especially in high-stakes decision-making. He notes that NowNow's approach focuses on grounding AI in locally relevant data to ensure accurate answers. The platform also connects AI to tools, allowing it to perform useful tasks in the real world.
The regulatory landscape for AI in South Africa is still evolving. A draft national AI policy was approved for public comment in March 2026, but it was officially withdrawn on June 12, 2026, due to references to non-existent sources. A revised version is expected by November 2026, with a new approach focusing on national guidelines for sectors to develop their own AI strategies, rather than prescriptive rules. However, experts like Gumede are pushing for more robust regulations to address concerns around AI adoption and its potential risks.
Why It Matters
NowNow's innovative approach has the potential to revolutionize how SMEs operate on the continent. By providing access to expert knowledge and tools, the platform can help bridge the gap between large companies and small businesses.
The success of NowNow also highlights the importance of addressing AI adoption limitations in South Africa. With over 50 million SMEs contributing significantly to the economy, leveraging AI can have a profound impact on job creation and economic growth.
Lawyers should be aware of the potential risks associated with relying on AI-generated information, particularly 'hallucinations.' Ensuring that clients are informed about these limitations is crucial when using tools like NowNow.
Practical Implications
Lawyers should watch for the potential risks of relying on AI-generated information, particularly 'hallucinations' where AI confidently gives false answers, and ensure that their clients are aware of these limitations when using tools like NowNow.
Source
How does this affect your business?
Get an AI analysis of this article grounded in your jurisdictions, practice areas, and any policy documents you've uploaded to Wansom.
Wansom is AI and can make mistakes.
