The Training Gap No One Budgets For

Every year, African enterprises spend between 8 and 14 percent of their IT budgets on new software licenses. They allocate another 5 to 10 percent on infrastructure upgrades. They set aside funds for consultants, for integration, for data migration. And then, almost universally, they allocate less than 2 percent for training the people who will actually use these systems. This is not an oversight. It is a structural failure that explains why most enterprise technology investments across the continent underperform by 40 to 60 percent against their projected returns.

The pattern is so consistent it has become invisible. A tier-one bank in Lagos deploys a new core banking system at a cost of $12 million. The vendor provides a two-week training programme for 30 staff members, who are then expected to cascade knowledge to 3,000 colleagues across 200 branches. Six months later, branch staff are still processing transactions manually and entering data into the new system as an afterthought. The bank has a $12 million system operating at roughly 35 percent capacity. No one calls this a training failure. They call it a change management challenge, or worse, they blame the software.

McKinsey's research on technology adoption consistently demonstrates that organisations investing less than 15 percent of total project cost in training and capability building achieve, on average, only 30 percent of targeted benefits. Those investing 20 percent or more capture 80 percent or higher. The gap between these two groups is not marginal — it is the difference between a successful transformation and an expensive disappointment. Yet across sub-Saharan Africa, the average training investment in enterprise technology projects sits between 3 and 7 percent of total project cost, well below the threshold where meaningful adoption begins.

The Anatomy of Under-Investment

To understand why training budgets are systematically inadequate, you need to understand how enterprise technology projects are approved in most African institutions. The business case goes to the board with three categories of cost: software licensing, implementation services, and infrastructure. Training, if it appears at all, is folded into the implementation line as a minor deliverable — typically five to ten days of vendor-led sessions focused on system navigation rather than process transformation.

This structure creates three compounding problems. First, the training is designed by the vendor, not the organisation. Vendors train users on features. They do not train users on how to rethink workflows, question existing processes, or leverage the system to solve business problems the vendor has never encountered. A vendor training a Kenyan insurance company on claims processing software will demonstrate how to enter a claim. They will not address the fact that the company's claims workflow involves seven manual handoffs, four of which exist solely because the previous system could not automate them. The new system can. But no one is trained to redesign the process.

Second, training is treated as an event rather than a process. The two-week training window closes, the consultants leave, and the organisation is expected to be self-sufficient. This ignores everything we know about adult learning and behaviour change. Research from the Association for Talent Development shows that 70 percent of learning in professional contexts occurs through on-the-job experience, 20 percent through coaching and mentoring, and only 10 percent through formal training. The standard vendor training programme invests exclusively in the 10 percent channel and abandons the other 90 percent entirely.

Third, training budgets are the first casualty of cost overruns. When an implementation runs over budget — and in Africa, 67 percent of large enterprise technology projects exceed their original budget by 20 percent or more, according to Standish Group data adapted for emerging markets — the training phase is compressed or eliminated to recover costs. The logic seems sound: the system is built, the licences are paid, we just need people to start using it. This logic has destroyed more technology investments than any technical failure.

The Compound Cost of Untrained Teams

The financial impact of inadequate training is significantly larger than the cost of the training itself. Consider a mid-sized commercial bank in East Africa that deployed an enterprise resource planning system in 2022. The total project cost was $8.5 million. Training allocation: $280,000, or roughly 3.3 percent of total spend. Within the first year, the bank documented the following costs directly attributable to inadequate system proficiency among staff:

Manual workarounds consumed an estimated 14,000 person-hours per month across the organisation, as staff reverted to spreadsheets and manual processes when they could not navigate the new system efficiently. At an average fully-loaded staff cost of $18 per hour, this represented $252,000 per month in productivity losses — nearly the entire training budget consumed every single month in waste.

Data quality degradation was equally severe. When staff do not understand how to use a system properly, they enter data incorrectly, inconsistently, or not at all. The bank's data quality audit six months post-deployment found that 23 percent of customer records had at least one critical field error, up from 8 percent under the old system. These errors cascaded into regulatory reporting, credit decisions, and customer service failures. The estimated cost of poor data quality: $1.8 million in the first year alone, including regulatory penalties, mis-priced loans, and customer attrition.

Staff turnover added another dimension. Employees who feel incompetent in their daily tools become disengaged. The bank saw voluntary turnover in operations roles increase by 11 percentage points in the 18 months following deployment — from a baseline of 14 percent to 25 percent. Each departure cost approximately $15,000 in recruitment and onboarding. With 120 additional departures attributable to the new system, this represented $1.8 million in avoidable turnover costs.

The total first-year cost of inadequate training: approximately $6.6 million. The cost of comprehensive training that would have prevented most of these losses: approximately $1.7 million, or 20 percent of total project cost. The bank saved $1.4 million by cutting training and lost $6.6 million as a consequence. This is not unusual. This is typical.

What Adequate Training Actually Looks Like

The organisations that consistently capture full value from technology investments approach training fundamentally differently. They do not treat it as a phase of the technology project. They treat it as a parallel workstream with its own budget, its own leadership, and its own success metrics.

The first principle is role-based design. Rather than training everyone on every feature, effective programmes map system capabilities to specific job roles and design training paths that connect the technology to the work each person actually does. A branch manager needs different system competencies than a credit analyst, who needs different competencies than a compliance officer. This seems obvious, but fewer than 20 percent of enterprise deployments in African institutions implement role-based training programmes.

The second principle is sustained engagement. The initial training event is just the beginning. Effective programmes include weekly coaching sessions for the first three months, monthly refresher modules for the first year, and ongoing advanced training as proficiency increases. Safaricom's deployment of its enterprise analytics platform in 2021 included a 12-month training programme with embedded coaches in every business unit. The result: 89 percent system adoption at the end of year one, compared to an industry average of 45 percent. The training programme cost 22 percent of total project budget and delivered an estimated 340 percent return on that incremental investment through superior adoption rates.

The third principle is measurement. You cannot improve what you do not measure. Effective training programmes track system proficiency scores for every user, correlate proficiency with business outcomes, and adjust training intensity based on actual performance data. This requires instrumentation — logging how users interact with the system, identifying where they struggle, and deploying targeted interventions. Less than 10 percent of African enterprise deployments implement any form of post-training proficiency measurement.

The fourth principle is internal capability building. Relying exclusively on vendor trainers or external consultants creates a permanent dependency. Effective organisations identify and develop internal trainers — power users who understand both the system and the organisational context — and invest in building their capability as coaches and knowledge leaders. These internal champions become the primary mechanism for sustained adoption long after the implementation team has departed.

The AI Amplification Effect

The training gap that undermines traditional software deployments becomes exponentially more damaging with AI adoption. Traditional enterprise software has a defined interface: menus, forms, buttons, workflows. Users can learn to navigate these interfaces through repetition even with minimal training. AI systems operate differently. They require users to understand not just what buttons to press but how to frame problems, interpret probabilistic outputs, recognize when the AI is wrong, and integrate AI-generated insights into human decision-making processes.

Consider a credit risk department implementing an AI-powered scoring model. Without adequate training, analysts will either blindly accept the model's recommendations — creating concentration risk and regulatory exposure — or dismiss them entirely and revert to manual assessment, rendering the entire investment pointless. Effective adoption requires analysts to understand the model's methodology at a conceptual level, recognize its strengths and limitations, interpret confidence scores correctly, and maintain the judgment to override the model when contextual factors warrant it.

This is a fundamentally different kind of training than learning to navigate a software interface. It requires developing what researchers call AI literacy — the ability to work effectively alongside intelligent systems. A 2024 study by the World Economic Forum found that organisations with structured AI literacy programmes achieved 2.6 times higher returns from AI investments than those that deployed AI tools without corresponding human capability development.

For African institutions, this challenge is compounded by the relative novelty of AI in enterprise contexts across the continent. While global financial institutions have been experimenting with machine learning for over a decade, many African banks are encountering AI-assisted decision-making for the first time. The learning curve is steeper, the conceptual gap is wider, and the consequences of inadequate training are correspondingly more severe.

Building the Business Case for Training Investment

The data overwhelmingly supports higher training investment. The challenge is not evidence but institutional incentives. CIOs and project sponsors are evaluated on budget adherence and delivery timelines, not on adoption rates 18 months post-deployment. Training is an easy target for cost reduction because its benefits are diffuse, delayed, and difficult to attribute directly. The costs of inadequate training, by contrast, are absorbed across the organisation — blamed on user resistance, change fatigue, or system limitations rather than the actual root cause.

Changing this dynamic requires reframing training as a value driver rather than a cost centre. Three approaches work consistently. First, tie training budgets to adoption metrics and make adoption metrics part of the project's success criteria. If a core banking system is expected to process 95 percent of transactions through its automated workflows, measure that metric and hold the project accountable. Second, ring-fence training budgets so they cannot be raided when other project costs overrun. This requires board-level commitment and explicit governance.

Third, measure the cost of under-training retrospectively and use those numbers to justify future investment. Most organisations have never quantified the cost of their training gaps. They know, intuitively, that adoption is poor. They have not translated that intuition into financial analysis. When they do — when they calculate the person-hours lost to manual workarounds, the revenue lost to data quality issues, the turnover costs attributable to system frustration — the business case for adequate training becomes overwhelming.

The Path Forward

African enterprises stand at an inflection point. The next five years will see the largest wave of technology investment in the continent's corporate history, driven by digital transformation mandates, regulatory modernisation, and competitive pressure from fintech and digital-native entrants. The institutions that capture value from this investment wave will not be those with the largest technology budgets or the most sophisticated systems. They will be the institutions that invest proportionally in the human capability to use these systems effectively.

The formula is not complicated. Allocate 15 to 20 percent of total project cost to training and capability building. Design training around roles, not features. Sustain engagement for at least 12 months post-deployment. Measure proficiency and correlate it with business outcomes. Build internal capability rather than relying on external dependency.

None of this is innovative. All of it is well-documented in global research and proven in practice. The only barrier is the institutional willingness to budget for it. And that willingness starts with a single recognition: the most expensive training programme is the one you did not fund. The gap between a $12 million system operating at 35 percent capacity and the same system operating at 85 percent capacity is not a technology gap. It is a training gap. And it is the gap that no one budgets for.

The institutions that close this gap first will define the next decade of enterprise performance across the continent. The rest will continue to accumulate expensive, underutilised technology assets and wonder why their digital transformation never quite delivers what the business case promised.