Generative artificial intelligence, or GenAI, is the talk of the town, with entire industries rethinking their business priorities and goals regarding the emerging technology.
By Tshepo Motshegoa, CIO of Seacom
According to one Salesforce survey of IT leaders worldwide, the majority of those in South Africa (87%) expect GenAI to have a prominent role in their organisations. This is on top of 83% saying the role of AI in their organisations is already well defined.
Looking more broadly, Bloomberg Intelligence reports that the global demand for GenAI over the next decade could add over $280 billion in new software revenue in the form of specialised assistants, copilots, and infrastructure products.
It’s not always immediately clear what the value of GenAI is to organisations, especially if they are only familiar with it in the form of its most popular use cases such as chatbots. GenAI is also a technology that requires a significant reallocation of IT resources and comes with a whole new set of pressure points companies need to address.
From small businesses to entire industries
The applications for GenAI range across multiple various sectors. In agriculture, data sourced from satellites and IoT devices can help farmers analyse factors such as weather conditions, soil and air quality, and overall crop yield to optimise their harvest and associated processes. In healthcare, algorithms can personalise the patient experience and assist doctors by automating administrative tasks such as updating patient files.
At a business level, GenAI can be integrated into various workflows. For example, a programmer or software developer can use it to write, update, and maintain code as well as quickly put together all the documentation related to that code such as user manuals.
A product designer can use algorithms to automate, evaluate, and optimise adjustments to a new product, for example by reducing the amount of materials used in a product build, which would then help reduce overall costs.
A project manager can tap into historical data to determine new project timelines and requirements, while a marketing manager can automate and hyper-personalise campaigns across all channels and social media platforms.
Already, there are local use cases such as law firms using natural language processing tools to provide legal advice to small businesses, and retail brands using chatbots to automate administrative tasks. But when it comes to implementations, or organisations that want to build their solutions from the ground up, what should you be wary of?
Legacy, complexity, and other challenges
For many organisations, the first obstacle to embracing GenAI would be surrounding legacy systems, and how integrating it into older technology environments can result in capacity and compatibility quandaries. An example of this would be whether current systems can handle the computing needs of a language model, leading to questions about whether enterprises should replace those systems outright, which then becomes a question of cost.
Additionally, GenAI can end up contributing to an organisation’s technical debt, with the resulting optimisation achieved doing little to minimise existing workloads. Organisations also need to keep an eye out for algorithmic bias that could compromise the quality of the content it generates.
A final challenge, one that is very relevant to the local market, is the availability of talent that’s required to action and manage GenAI implementations.
The technology has spurred demand for AI specialists and professionals in South Africa over the last year. Employers are on the hunt for AI candidates who are proficient in machine learning, predictive modelling, and programming languages such as Python and SQL.
This has the potential to exacerbate the country’s existing skills gap and spotlights the need for not just an overhaul in the education sector in terms of ICT skills, but the need for organisations to seize the initiative and nurture those skills in-house.
Resources and tools
Whether enterprises are looking to build their own GenAI models and solutions, or just leverage existing ones as part of their implementation strategy, there are several considerations to be aware of. The decision to develop custom GenAI solutions should be based on various factors, including the availability of prerequisite skills, knowledge and resources, and the use case in question.
The first and foremost consideration is data. Data acquisition, storage, and processing sit at the heart of any GenAI ambition, and so organisations need technical infrastructure that can cope with those needs and scale when necessary.
South Africa’s growing cloud landscape and the popularity of infrastructure-as-a-service (IaaS) models means companies now have the choice between on-premise, cloud, or a combination of the two to meet their infrastructure needs.
AI innovation and the regulatory landscape are in constant flux, so even if it is just a case of adopting popular APIs for the short term, South African enterprises need to move ahead in any way they can. Competition in any business sector is fierce and, by investing in emerging technologies and the infrastructure that underpins them, enterprises can gain an edge and build a launchpad on which they can build the AI-powered businesses of the future.