Talent and tech: Critical elements for AI success

October 2024 Editor's Choice

AI has been generating a great deal of excitement around its potential to transform various sectors and industries across South Africa. Despite all the buzz, the business value of AI will only be fully realised when enterprises evolve beyond simply consuming the technology to wielding it for competitive advantage, because this is where the magic happens.

In other words, AI’s true enterprise power is leveraging it to project and act upon trends, such as proactively uncovering shifts in customer purchasing behaviours and pre-emptively pivoting to meet forecasted opportunities ahead of the competition.

Most local enterprises have significant AI ambitions — almost 70% of organisations in sub-Saharan Africa are deploying cloud and AI technology to improve productivity and drive growth. However, for these businesses to truly start developing a competitive AI advantage, they must first lay a solid foundation for AI transformation. A recent global survey, commissioned by HPE and conducted by Sapio Research, provides critical insights as to where current gaps lie, that can help local business leaders create a strong AI strategy.

The survey shows that the majority of organisations have thus far focused on tactical projects around organisational efficiency rather than strategic initiatives that promote growth.

The same research also reveals that enterprises are overly confident in their readiness to shift from tactical AI projects to strategic ones. Most notably, fewer than half of enterprise IT leaders reported having a full understanding of enablement needs across the complete AI lifecycle. As the foundation for any successful technology initiative, AI or otherwise, is fully understanding the needs, this is a significant deficiency.

Within that deficit, two of the most critical gaps occur in the areas of talent and technology. For the latter, this is specifically around the computing, storage, and networking infrastructure required for supporting AI. To help enterprises overcome these disparities, let us dig deeper into each and discuss how organisations can respond effectively.


Michael Langeveld

Gaining the right talent

When considering talent for AI projects, enterprises seem to hold some fundamentally invalid assumptions about the breadth of individuals required at the table. Companies in the Sapio survey named the obvious cast of characters, including the C-Suite, IT leads, AI/ML engineers, data scientists, and network managers, among others.

However, less than a quarter of them have legal teams involved in their business’s AI strategy conversations, and similarly, only about a third have looped in their HR department. This mirrors generally held views that legal, compliance, and ethics are the least critical elements for AI success, selected by only 13% and 11% of research participants.

Of course, this is a serious blind spot. From a legal and compliance perspective, new regulations are in the pipeline worldwide for which the requisite expertise is necessary to ensure an AI initiative complies. For HR, the ability to fill such critical positions relies upon strong representation on each AI initiative team to prevent talent acquisition and retention bottlenecks that lead to delays or even outright project abandonment.

Recognising IT infrastructure gaps

Like talent, enterprises are also overly confident in their AI-related technology capabilities, according to the report, as 93% of IT leaders say their network infrastructure is ready to support AI traffic. Further, 84% indicate their systems have enough flexibility in computing capacity to handle the unique demands of AI applications at scale.

Yet, the certainty breaks down upon deeper examination. For example, 40% or fewer of IT leaders fully grasp the unique networking and computing needs for either AI model tuning or inferencing. Clearly, this poses a danger to ensuring that AI applications have the required technology resources.

Take a pilot-to-production approach

To discover what talent and technology your enterprise needs to move from tactical to strategic AI adoption, start by pursuing the well-accepted path of using smaller pilots to uncover the gaps within your talent and technology pools.

Pilots also provide the opportunity for balancing cross-functional business and IT teams to determine what gaps need addressing. Regardless, it is imperative that every functional area — business and IT — be represented right from the start to ensure sufficient expertise for identifying and addressing blind spots and roadblocks during the pilot and when transitioning to a larger production initiative.

In this way, AI initiatives can be shaped appropriately to ensure insights gleaned from them are fit for purpose and solve critical problems, ultimately enabling the delivery of the envisioned business outcomes.

Infrastructure: go with private or hybrid clouds

From a technology infrastructure perspective, it is tempting to consider starting a pilot in the public cloud. However, utilising public cloud platforms introduces significant latency into computing processes, preventing AI models from operating at the necessary speeds for completing real-time computations at scale.

This makes a public cloud approach unattractive for an enterprise AI initiative as it will ultimately necessitate moving to a private cloud and retooling your AI applications, creating considerable inefficiencies that markedly delay production rollouts and substantially increase costs.

To gain both the capacity, performance, and scalability required, it is recommended that enterprises start with private cloud infrastructure at the outset. Private clouds are designed to handle the unique workloads of an inherently more complex strategic AI initiative.

In addition, private clouds are more cost-effective as you can begin with a solution sized for early piloting phases and then grow seamlessly to include the appropriate supercomputing solutions as you move into production. This approach enables you to avoid retooling costs and ensure the fast, efficient rollouts necessary for gaining business advantage in the marketplace.

For enterprises with AI pilots already underway in an existing public cloud, the most effective way to add private cloud is using a hybrid cloud approach. A hybrid cloud provides you with a centralised interface for managing both public and private clouds, significantly reducing administrative costs.

Open, flexible, and adaptive fuels AI innovation

Regardless of how your journey begins, plan to adopt open systems solutions for your infrastructure. This will provide your AI initiative with the ability to use the best tools for any given circumstance rather than being locked into a specific vendor’s offerings.

Even better, look for solutions that are also cloud-agnostic, as this permits IT staff to use a unified management interface for monitoring every AI workload, no matter where they are, or which cloud they reside in.

Most importantly, choose a solution with an adaptive architecture that can scale rapidly and appropriately as your AI maturity progresses. This will enable harnessing vast amounts of both public and private data to fuel analytics applications and AI model development, from training to tuning.

With common gaps illuminated and tips on how to close them, South African enterprises can take the next steps in their AI journeys. By constructing a solid talent and technology foundation, they can transition from operational to strategic AI to make the most of this exciting AI era.




Share this article:
Share via emailShare via LinkedInPrint this page



Further reading:

Surveillance on the perimeter
Axis Communications SA Hikvision South Africa Technews Publishing Editor's Choice Perimeter Security, Alarms & Intruder Detection
Cameras have long been a feature in perimeter security, with varying reports of success and failure, often dependent on the cameras’ planning, installation and configuration, as well as their integration with other perimeter solutions and centralised management platforms.

Read more...
Onyyx wireless alarm
Technews Publishing Editor's Choice Smart Home Automation
IDS has introduced Onyyx, a wireless alarm system engineered to provide complete system control via the Onyyx app or keyring, as well as seamless installation.

Read more...
Visual verification raises the security game
Technews Publishing Inhep Electronics Holdings Videofied SA Editor's Choice Perimeter Security, Alarms & Intruder Detection
Incorporating alarm signals with live surveillance footage, visual verification enables a human observer in a control room (onsite or offsite) to gain a clear understanding of the situation, thereby facilitating informed decision-making.

Read more...
The AX Hybrid PRO Series offers reliable wired and wireless protection
Hikvision South Africa Editor's Choice Perimeter Security, Alarms & Intruder Detection Products & Solutions
Hikvision has announced the launch of a new AX Hybrid PRO alarm system with innovative Hikvision ‘Speed-X’ transmission technology. This system offers reliable wired protection while delivering expanded flexibility with seamless wireless integration.

Read more...
A critical component of perimeter security
Nemtek Electric Fencing Products Gallagher Technews Publishing Stafix Editor's Choice Perimeter Security, Alarms & Intruder Detection Integrated Solutions
Electric fences are standard in South Africa, but today, they also need to be able to integrate with other technologies and become part of a broader perimeter security solution.

Read more...
SMARTpod talks to The Risk Management Forum
SMART Security Solutions Editor's Choice News & Events Security Services & Risk Management Videos Training & Education
SMART Security Solutions recently released its first SMARTpod podcast, discussing the upcoming Risk Management Forum Conference 2024, which will be held on 26 September 2024 at the Indaba Conference Centre in Fourways, Johannesburg.

Read more...
There is a SaaS for everything, but at what cost, especially to SMEs?
Editor's Choice Information Security Security Services & Risk Management
Relying on SaaS platforms presents significant cybersecurity risks as the number of providers in your landscape increases, expanding your attack surface. It is important to assess the strength of the SaaS providers in your chain.

Read more...
New State of Physical Access Control Report from HID
HID Global Editor's Choice Access Control & Identity Management News & Events
HID released the 2024 State of Physical Access Control Report, identifying five key trends shaping access control's future and painting a picture of an industry that has been undergoing considerable transformation.

Read more...
Addressing today’s mining challenges: cyber risks beyond IT
Editor's Choice Information Security Mining (Industry)
Despite the mining industry’s operational technology systems being vulnerable to cyberattacks, many decision-makers still see these threats as purely an IT issue, even though a breach could potentially disrupt mining operations.

Read more...
Workforce Consortium to reskill 95 million people
Editor's Choice News & Events AI & Data Analytics
ICT Workforce Consortium of global leaders has come together, committing to train and upskill 95 million people over the next 10 years, as 92% of jobs analysed are expected to undergo either high or moderate transformation due to advancements in AI.

Read more...