Sound data management and governance are at the heart of successful AI initiatives, according to experts from InfoVerge, who were speaking in a series of three webinars on AI roadmaps, scaling and governance.
The webinars, presented to help organisations address challenges in optimising their AI initiatives, provided a roadmap from early adoption to scaling AI with confidence.
In the first webinar of the series, "Building the Foundation for AI", Skhumbuzo Mjoji, CTO at InfoVerge Solutions, said: “AI is having a huge impact on society and business. But any AI initiative’s success depends on a robust data foundation.”
He outlined key steps for building this foundation, including building an AI-first data strategy, establishing a single source of truth and continuously managing, governing and securing data.
Webinar two of the series, "From Data to Intelligence: Activating Smarter Operations", addressed challenges such as struggling to turn data into clear, actionable insights, scaling AI initiatives, proving the value and return of AI investments, managing resistance to change in decision-making processes and expanding isolated pilots into broader adoption.
Ntiyiso Mayile, Chief Data Officer at InfoVerge Solutions, said: “It all goes back to governance and having the fundamental data foundations in place.”
Mjoji and Mayile emphasised that AI success also depends on operational maturity, while a lean and structured MLOps approach helps reduce project failure. They recommended starting small, validating projects early and scaling based on measurable value.
Mjoji said: “It doesn’t have to be as complicated as it might seem. InfoVerge helps organisations by guiding them through the entire AI-to-operations life cycle, with a Starter Ops Blueprint to accelerate time to value.”
In the final event of the series, "Scaling AI Success: Governance, Ethics and Future-Proofing", Mayile noted that AI investment in the US alone had topped $100 billion – an indication of the scale and speed of adoption.
“However, as fast as the technology grows, you have regulatory bodies trying to catch up with governance policies. South Africa does, in fact, have a national AI policy and we will see more regulations drafted as AI is more widely adopted,” he said.
Mjoji added: “Scaling AI requires structure and governance – without that, you cannot take a solution enterprise wide.”
A poll of the audience found that only 5.56% had a fully established governance framework in place for AI, 44.44% said they had one in progress, 27.78% planned to establish one, and 22.22% didn't have a policy and said it was not a priority yet.
On the question of how confident they were in their organisation’s ability to navigate AI-related regulatory and compliance requirements, 33.3% of webinar attendees were not confident, 50% were somewhat confident, 8.33% were very confident, and a further 8.33% said they were unsure of what applied.
A poll on challenges with AI-related data quality found 16.67% said inconsistent or incomplete data were a top challenge, 38.89% cited a lack of standards or ownership, 27.78% said integrating data across silos was their key challenge, and 16.67% were not sure.
Highlighting potential risks, Mayile said: “Many organisations are using a Copilot, and in doing so, they accept all the terms and conditions and opt in to allow it to access all your data – and some of it could be sensitive, such as contracts and personal information. If you simply accept the terms and conditions, you are at risk.”
He advised: “There are some things you need to think of when implementing an AI risk management framework. NIST’s AI Risk Management Framework gives you a good starting point. It was specifically designed to help organisations manage risk throughout the AI life cycle, but it is also flexible enough to allow you to bring in other policies such as South Africa's framework. Microsoft Purview helps you implement your NIST framework, define your IT policies around who can see what and how, and monitor and scan the environment."
He highlighted Microsoft Purview data security posture management (DSPM) for AI, and demonstrated the Microsoft NIST AI Risk Management Framework dashboard. Purview offers comprehensive risk assessments with actionable points and specific solutions to improve risk management, he said.
Mjoji noted: “If your organisation harnesses AI without the necessary protections, you are at high risk of being in breach of regulations. Scaling AI requires structure and governance. With the NIST framework, we are talking not just about documents you refer to, we are talking about comprehensive system implementations to reduce the risk to the organisations. InfoVerge can help organisations implement and tailor those frameworks, and help implement ethical AI practices within your environment. We offer all the tools, processes and expertise to scale AI within your organisation.”
You can watch InfoVerge’s AI webinar series on demand here and see how data is fuelling AI-driven smart business.
Share