How AI is breaking banking’s old employment model

Banking jobs in Kenya and, by extension, most African markets have carried a certain social prestige thanks to stable salaries, pension plans, and confidence in a sector that seems too important to shrink. 

But the proliferation of artificial intelligence (AI) is threatening to rewrite this promise. Looking at a bank like Standard Chartered Kenya (StanChart), the numbers tell a big story before its executives do.

In 2013, StaChart had over 2,200 employees. At the time, the bank operated a large branch network, sizable operational teams, several middle-management roles, and thousands of employees handling most processes manually—from onboarding customers and processing paperwork to compliance reviews and reconciliations.

By the end of 2025, its workforce fell below 1,000 employees for the first time in history.

These shifts at Stanchart signal re-pricing of labour inside Africa’s banking sector. The work that used to justify thousands of entry- and mid-level roles is now being done by systems that are cheaper and involve far fewer people.

In May, the lender’s parent company signalled that the decade-long cuts are not temporary, but part of its new strategic focus. During an investor event in Hong Kong on May 19, the British bank said it plans to cut more than 15% of its support-function staff by 2030. 

These are the people working in areas like human resources, compliance, procurement, operations, and administration. The bank openly said AI will help replace many of those tasks as its acceleration will “deliver faster execution and clear financial outcomes.”

It is moving toward what it calls a “simple, connected and fast” operating model, in which every task is assigned to automation, AI-assisted workflows, or humans. 

By 2027, it expects 90% of key technology controls to be continuously monitored by AI, while 80% of controls will be fully codified into executable rules. Operational processes are also being automated, with AI document processing targeted at 95% accuracy (up from 85%) and virtual assistants expected to resolve up to 60% of internal queries without human intervention.

The bank has deployed more than 300 AI use cases, including 43 high-impact generative AI applications, and trained about 85,000 staff on Microsoft Copilot. It is reporting early efficiency gains, including a 40% reduction in false positives in digital asset surveillance, an 88% cut in monitoring manpower through centralised systems (saving roughly $10 million annually), and a 30% reduction in manual effort tied to regulatory change implementation. 

AI is coming for the people

The first wave of digital banking killed some branches, but AI is now coming for the few ones that are remaining and even the headquarters. 

In essence, the first era was customer-facing. Banks spent the last 15 years persuading customers to stop visiting branches and use online or mobile banking, ATMs, and agency banking. This removed the need for physical interactions, moving a majority of transactions outside bank halls.

The first phase of the transition affected only frontline workers, such as tellers. As branch footprints shrank, cash handling declined.

But the next stage of automation, as signalled by Stanchart, is more consequential because it targets the institutional backbone inside banks themselves.

Banking functions like human resources, compliance, call centres, and customer onboarding employ thousands of people across African markets precisely because banking remains one of the continent’s most administratively complex industries. The sector must navigate fragmented identity systems, cross-border regulations, paper-heavy documentation requirements, anti-money laundering obligations, and diverse payment infrastructures across multiple markets.

Historically, a large workforce addressed most of these inefficiencies, but AI now threatens to do so more cheaply. That is the significance of StanChart’s announcement. The bank is arguing that many support functions no longer need to be labour-intensive.

For instance, a large language model (LLM) can review documents continuously without overtime costs and flag suspicious transactions faster than human analysts. Automated compliance systems can process vast amounts of regulatory information instantly, while customer-service chatbots can handle thousands of queries simultaneously.

What once required floors of junior employees requires software infrastructure supervised by a smaller number of specialists.

The middle-class jobs

The danger of faster AI adoption in banks for African economies is not simply unemployment. It is the erosion of middle-tier professional work.

Banking has historically been one of Africa’s most important engines of the urban middle class. It created structured graduate recruitment pipelines, management-training programmes, pension-backed careers, and relatively stable white-collar employment.

Notable African political and business elites passed through banks early in their careers. What AI threatens to remove are precisely the kinds of jobs that created those pathways.

These jobs are repetitive enough to automate but skilled enough to have historically supported middle-income urban life. That creates a bigger social risk.

If banks continue to earn strong profits while employing significantly fewer people, the sector may cease functioning as a major employer. Banking could resemble the technology sector itself, becoming highly productive and highly profitable while employing small numbers of specialised workers.

And transformation may already be underway. Across Kenya’s banking sector, hiring is concentrated around cybersecurity, data engineering, AI, and specialised relationship management rather than traditional operations. Some banks like KCB Group and Equity Group continue expanding overall staff numbers, but the composition of hiring is changing.