Johannesburg, 13 Mar 2024
It's staggering that generative artificial intelligence (AI) has only been mainstream for 14 months. In late December 2023, OpenAI released ChatGPT, the front end of its GPT AI models. Typically, such a splash would run its course, cooling down substantially as people work through their curiosity and throw a few use cases at the new technology.
But this isn't what's been happening with AI.
Instead, adoption has been massive. So is the growth of AI through competing start-ups and AI appearing in vendor applications. In a survey posted late last year by the Conference Board, 56% of US office employees said they use generative AI, and 10% use it daily. An Organisation for Economic Co-operation and Development (OECD) survey of 5 000 employees across the world revealed that 80% felt AI improved their performance at work.
"You often hear about 'democratising' technology, which I think is completely appropriate for AI," says Muhammed Omar, ServiceNow's Africa Manager. "It's closing that gap between business and technology. When providing AI, you're empowering business people to be more autonomous and have more control over what they want to create through tech systems. They can articulate their ideas in a way to create their own solutions with limited overhead from the back office."
Closing the gaps for business
AI closes many gaps. Many people already use it for summarising, such as compiling case files or meeting minutes. AI makes no-code development easier, allowing more non-technical people to use technology for business improvements. It's improving companies' front, middle and back, helping employees and technologies behave more agilely.
"One of the biggest frustrations for companies is the time that it takes to meet internal business demands in an agile way. AI is helping to close that gap. It puts businesses on the trajectory to close gaps between the front, middle and back offices," says Omar.
Most companies will rely on several AIs, some serving most users and some specialising in specific areas. Specialised and domain-specific AIs are emerging: for example, Microsoft is currently testing Copilot For Security, an analyst-facing AI that supports reporting and remedying security issues. Similar vertical and domain-focused AIs are also gaining traction in fields such as finance and medicine.
One of the AI revolution's most exciting characteristics is how easily non-technologists create feasible use cases and approach these with pragmatic attitudes.
"Normally, business people keep technology at arm's length or have ideas that are not feasible for the technology they have in mind. But that's not what I'm encountering out there with AI. In most of corporate South Africa, enterprises have got a view of three or four use cases. And more striking is how grounded these use cases are. Companies want to build on AI and grow with it, which tells you how much AI's potential aligns with their ambitions."
De-risking AI in your organisation
Yet AI systems come with risks. The two primary concerns are that companies will lose control over their data and compliance, or that AI will replace employees. These worries are not unfounded and are more likely to occur if a business doesn't have a central platform to manage its AI services and integrations.
"Will I lose control? Will I lose my job? I hear these concerns often," says Omar. "The best way to use AI is as a collaborator, not something that takes over. Generative AI can make recommendations, such as around process bottlenecks. But you want somebody who looks at the recommendation and executes it. This new generation of AIs are companions, not employee replacements. And if you establish safety rails and establish appropriate policies informed by how people use AI, you won't lose control."
Business services platforms are the best candidates to create such environments. "There are several reasons why a services platform is the gateway for AI projects. These are not AI-first platforms, which means the platform doesn't exist just to make the AI happen. Instead, AI becomes an integrated part of a larger business ecosystem powered by the platform. The second factor is that the best services platforms provide a single data model and system of record so that everyone and everything, including the AIs, use the same data. Companies can create guardrails, sandboxes and policies on the platform, thus giving their employees more autonomy with AI without creating unnecessary risks."
The best platforms facilitate integration with outside AIs while helping maintain data control and visibility. They also develop and introduce AI features that support business staff and processes, further reducing the risks associated with using AI services that may have opaque access, data usage or costs.
"African organisations can use AI to improve their global competitiveness, become more data-driven and look after their skilled employees. The challenge with AI is not whether your business should use it, but how to do so without destabilising operations or creating unmanaged risks. Service platforms are starting to show that they put companies in the best position to deal with these concerns, become more agile and get the benefits of AI on their terms and use cases."
Share