SA is entering a new era in data management, characterised by a rise in agentic AI and exposure to a new frontier of reasoning large language models (LLMs).
This is according to Nkululeko Thangelane, group executive head: big data, AI and ML at Vodacom, who delivered the opening keynote address at the ITWeb BI Summit 2025 on 12 March at The Forum, Bryanston.
As opposed to LLMs that use statistical guesswork to provide outcomes, reasoning models are designed to handle complex issues by adding steps within a logical ‘chain of thought’ process to help derive more accurate outcomes.
Thangelane explained that Vodacom began its AI/ML journey in 2016 and has integrated GenAI technology into its KPIs, with 116 use cases having gone live in 2024. The company has upskilled 1 000 data scientists and other members of its ecosystem across Africa.
Vodacom has applied GenAI to support the implementation of its ‘ask once’ philosophy, based on the intention to address client requirements based on a single conversation, eliminating the need for customers to keep calling and having to explain the issue multiple times.
Thangelane added that the technology is used to address several challenges – particularly with call management within the call centre environment. These include a lack of contextual awareness, data blindness for management and an ineffective knowledge base.
The company continues to back smart service use cases, including the application of GenAI frontline enablement, a dashboard to provide insights to bolster management decision-making, as well as knowledge base and sharing.
“One of our key considerations is whether calls have been resolved. LLMs are used to tag the sentiment of calls, to recategorise these and ‘bucket’ them. The issue of non-disposition is a challenge,” said Thangelane.
Disposition is the way the outcome of calls is categorised.
He said agentic AI can be used to better understand non-disposition.
“There are massive benefits to AI-driven data management, but it must be looked at holistically. There are benefits to anyone who works with data. GenAI is transforming data management by automating processes, improving data quality, unlocking structural data and enabling real-time insights for smarter decision-making,” Thangelane added.
Data storage and architecture were highlighted as examples.
In terms of a GenAI use case, Databricks integrates Mosaic MLS GenAI to enable AI-driven data structuring and organisation.
According to Thangelane, the South African market is heading towards agentic AI, and businesses that are prepared can capitalise.
Share