South African telecom operator Telkom says it will spend $6.1 million to build artificial intelligence skills, according to a report carried by Agence Ecofin. The move lands as companies across the country race to train workers in digital and AI tools amid tight hiring conditions and a push for productivity gains in telecoms and beyond.
Telecom operators are already leaning on AI for network automation, energy optimization, fraud detection, and customer service—areas that have become standard items on technical roadmaps. For Telkom, operating in a competitive market, the challenge is no longer just buying software. It’s finding and developing people who can deploy these systems, explain them, and secure them.
The pressure extends well outside telecom. South African banks, retailers, industrial firms, and public services are increasingly looking for hybrid skill sets—data science, cybersecurity, software engineering, and an understanding of real-world use cases. Building a local pipeline is difficult, especially when in-demand workers can be pulled toward international offers, often made easier by remote work.
Telkom commits $6.1 million to build AI skills
Telkom’s announcement centers on a $6.1 million commitment dedicated to AI skills development. In telecom, budgets like this typically translate into structured training tracks—certification courses, internal bootcamps, partnerships with educational institutions, or funded programs for recent graduates. At the scale of a large company, the money is less about covering everything than about accelerating training in roles the business considers urgent.
The most sought-after profiles often include data scientists, machine-learning engineers, and cloud architects, but the needs run wider. Operators also need specialists who can industrialize models—often referred to internally as MLOps—manage data quality, and document systems for production use. Support functions are part of the equation too: legal and compliance teams need to understand how models work and where they can fail, while managers need to run data projects without turning AI into a passing fad.
Use cases in telecom are broad. AI can help predict network failures, prioritize field work, reduce downtime, and optimize electricity consumption by equipment. Customer contact centers use conversational assistants and call-summarization tools, where quality control and error reduction are central concerns. Telkom’s focus on training—internally or through partners—aims to reduce reliance on outside contractors while keeping deep in-house understanding of systems that run critical functions.
The issue is also social. By investing in skills, a company can retrain part of its workforce rather than absorb abrupt automation. In telecom, AI changes day-to-day work—less repetitive processing, more analysis, supervision, and configuration. The budget could support pathways from existing jobs to new needs, including technicians, analysts, developers, and call-center staff, with the goal of helping the company manage how work evolves.
South Africa’s AI market is running into a talent shortage
Telkom’s decision comes as South African companies report sustained strain on digital skills—especially in data and AI. Hiring is difficult for several reasons: competition across industries, rising wages, international mobility enabled by remote work, and the long training runway required for experienced profiles. In AI roles, scarcity isn’t limited to theoretical experts; it also hits professionals who can deploy models in production under real constraints around security, cost, and performance.
The shortage feeds directly into budgets. When talent is scarce, companies face a choice: pay top dollar in the market, outsource to consulting firms, or invest in training to build an internal pipeline. Training is often favored for stability. An employee trained on internal data and systems can quickly outperform an external consultant, but retention becomes critical—career progression, concrete projects, and recognition—otherwise training can simply make workers more attractive to competitors.
The local landscape is also shaped by a gap between academic programs and industrial needs. Schools may provide foundations in math, statistics, and programming, but companies want experience with real data—noisy, fragmented, and governed by rules. Skills like model versioning, traceability, and continuous evaluation are increasingly central. Many environments are hybrid—on-premises and cloud—with data sovereignty and sector regulation adding complexity in telecom.
AI is also driving demand for cross-disciplinary skills. Cybersecurity is increasingly inseparable from AI: protecting datasets, resisting data-poisoning attacks, controlling access, and managing identities. Teams also need to audit bias, test robustness, and document decisions. As AI spreads quickly through the economy, the shortage shifts toward these hybrid profiles—people who can speak to business needs, IT, and risk. Telkom’s program is likely aimed at meeting that broader demand, not just training model builders.
Telecoms are deploying AI across networks, customer service, and fraud
Inside a telecom operator, AI deployments often cluster in three areas: networks, customer relationships, and risk control. On networks, the promise is a shift from reactive maintenance to predictive maintenance by analyzing performance histories and weak signals. Models can prioritize work on cell towers or fiber segments based on failure probability and user impact. That can improve uptime, but it requires high-quality data, fine-grained instrumentation, and teams able to interpret alerts correctly to avoid unnecessary operations.
On the customer-service side, automated assistance tools are multiplying—chatbots, voice agents, call summaries, ticket sorting, and contact-reason detection. The goal is faster resolution and fewer repetitive tasks for human agents. But deployment requires strict oversight, because a wrong answer can erode trust. Training therefore needs to cover journey design, controlling “hallucinations” in generative systems, and clear escalation procedures to a human. Field teams also use smartphone diagnostic apps, which raises integration and update-management demands.
Fraud and security are another major front. Operators analyze suspicious behavior—abnormal SIM use, automated calling, identity circumvention attempts, or inconsistent consumption patterns. AI can speed detection, but it must be tuned to limit false positives, which carry commercial costs and legal risk. Teams need to understand metrics—precision, recall, thresholds—and balance network protection with customer experience. Training people who can run these models becomes a direct profitability issue.
Against that backdrop, Telkom’s investment can be read as a bet on integration: building internal capability to control tools, reduce dependency, and respond faster. A model that performs well in a lab is useless if it can’t be maintained or explained. Skills become an invisible infrastructure—without them, software investments lose value quickly. Operators that organize data teams early can move faster to scale projects and avoid a pile-up of prototypes that never reach production.
Training partnerships and local jobs are central to operators’ strategies
Skills programs in South Africa often rely on partnerships—universities, technical schools, incubators, certification platforms, or alliances with cloud providers. For Telkom, the upside is twofold: access to up-to-date content and a broader candidate pipeline. AI changes quickly, and purely internal training can become outdated. Partnerships can deliver modules aligned with tools used in business—data pipelines, managed services, deployment practices—while providing certification markers that support internal mobility.
Local employment is a key variable. Training people on the ground can reduce reliance on imported skills and lower recruiting costs. But effectiveness depends on real pathways into work—internships, apprenticeships, supervised projects, and mentoring. Companies that invest without clear opportunities risk funding skills that end up leaving. Telecom operators have an advantage in offering concrete use cases—network data, customer tickets, field operations—allowing training on real problems while respecting confidentiality constraints.
Internal talent management becomes a natural extension of the budget: recognized technical career tracks, updated pay scales, time set aside for learning, and a culture of knowledge sharing all shape the return on investment. Operators also set up “data communities” and internal labs to share tools and prevent fragmented AI efforts. In an environment where every department may want its own AI solution, governance becomes a performance driver.
The approach can ripple through the broader ecosystem. When a major player funds skills, it can lift contractors, help local startups hire, and influence public-sector methods. There’s also a risk of talent concentrating inside a few large firms, which is where public policy and academic partnerships can help balance access by expanding training tracks. In Telkom’s case, the $6.1 million figure puts a number on a priority: skills as a condition for durable AI adoption in infrastructure and services, as South Africa looks to strengthen digital competitiveness in 2026.
