Banks in South Africa have significantly changed how they interact with customers, including account opening, payment processing, and receiving assistance. Fast and intuitive digital service is no longer novel but has become a necessity for operation. However, ease of contact does not always equate to ease of problem resolution.
A customer might upload documents for updates according to the Financial Intelligence Centre Act (FICA) via an app, then contact the call center, and finally visit a branch if their request remains unresolved. Even if all channels function as intended, the customer is forced to repeat information because no one sees at what stage their request is. For someone buying extra connection time or taking a vacation, this friction incurs direct costs.
The real test comes when the digital process fails or the request falls outside standard procedure. These are the moments that shape the relationship between the bank and the customer: an employee with authority and situational understanding can turn this friction into trust, whereas one who demands the customer restart the process does the opposite.
The Cost Hidden Between Systems
For technology executives, the main task is connecting various channels to the operations that fulfill requests. According to a PwC analysis of major banks for September 2026, the total cost-to-income ratio was 49.6% in the first half of 2026, down from 50.6% the previous year. The report also noted increased competition for customer relationships.
It is possible to simultaneously improve service quality and reduce costs by eliminating unnecessary work. Expenses arise when employees re-enter information, requests move between queues without an assigned owner, or teams use email because the process cannot handle an exception. As a result, someone must compile the customer's history from multiple systems before making a decision on further action. Customers interact with one bank and expect it to act as a single entity; internal task handoffs should not force them to fill in the gaps themselves.
Completed Interaction Does Not Equal Resolved Request
Wait times for calls, retention in chatbots, form filling, and application availability demonstrate channel efficiency. However, they do not always answer the most important question: did the customer receive what they needed? Each interaction may meet its own KPIs, while the request itself continues to move within the organization. Initial resolution is useful where the request can be reasonably completed on the first contact. Complex issues also require assessing the time to final resolution and the frequency with which customers inquire about the status. Improving these metrics implies that the request itself must become the primary unit of work.
What Is Actually Required to Connect Work
This begins with continuous case logging: a common context that accompanies the request across different channels, teams, and systems. This log must reflect what the customer needs, what happened, who is responsible for the next step, and what decisions or approvals are still pending. Then, an orchestration layer can coordinate the work without replacing every underlying platform. It can direct activities, invoke existing systems, manage exceptions, and apply controls, making delays visible before the customer has to make subsequent inquiries. Core platforms, payment systems, and specialized applications can continue to perform their functions, while orchestration connects their contribution to the service journey.
The process is as important as the architecture. Automating unnecessary checks or poorly designed task handoffs only accelerates the movement of inefficiency. First and foremost, banks need to determine which controls are vital and which steps truly add value.
Resilience Goes Beyond the Application
Fragmentation also hinders the understanding of operational resilience. South Africa's Joint Standard 2 on cybersecurity and cyber resilience came into effect on June 1, 2025. Its requirements concern protecting critical information assets, ensuring business continuity, and strengthening institutions' ability to prevent, detect, and recover from cyber threats.
A broader implication for service delivery is that application recovery is only part of the picture. Banks must understand the systems, teams, task handoffs, and third parties supporting the service from start to finish. The application might be available, but the underlying process for the request could be stalled. For the customer, the service is still not working.
AI Cannot Create Continuity By Itself
Artificial intelligence can summarize histories, classify requests, extract information, and recommend actions. But a faster chatbot achieves little if its request disappears into an opaque queue. Banks also need to track what led to an automated recommendation, what happened next, and where human oversight intervened. The strongest AI use cases are within processes that are understandable, controllable, and measurable. Otherwise, banks risk automating gaps instead of closing them.
Start With One Process
Attempting to connect all operational activities is too broad a starting point. One should select one high-volume service with a high level of friction and trace its path from initial contact to full resolution. Where is the context lost? Which task handoffs add value? Who owns the request if an exception occurs? Can the bank see the delay before the customer asks about it? The answers will show where process redesign, continuous case logging, and service orchestration can improve the customer experience. They will also reveal areas requiring deeper platform modernization and places where targeted changes can bring benefits faster and with fewer disruptions.
South African banks have already changed how customers access banking services. The next competitive advantage lies in making one promise a reality: as soon as a customer requests something, the bank must accompany that request, its context, and responsibility until the moment of resolution.



