AI Agents Are Quietly Transforming South African Banking — Here’s How
Just a few years ago, large language models and generative AI were concepts confined to research labs and tech circles. Today, the conversation has already moved on. The next frontier — agentic AI — is here, and it is not waiting for anyone to catch up. These semi-autonomous or fully autonomous systems can reason, make decisions, and take action independently, and their adoption is accelerating at a pace few predicted.
A recent PwC survey found that 79% of senior executives globally already have AI agents operating within their organisations in some capacity. For a technology that is still relatively new, that level of penetration is remarkable. But what surprises many people is that South Africa is not lagging behind — far from it. The country’s financial services sector, in particular, is already deploying agentic AI at scale, with early results pointing to significant gains in efficiency and customer experience.
For everyday banking customers, the impact is tangible. Faster service, more personalised interactions, quicker resolution of routine queries — these are no longer aspirational goals. They are becoming the new baseline, driven largely by intelligent automation working behind the scenes.
How AI Agents Are Reshaping Banking in South Africa
Consider what happens when a small business tries to open a bank account. Traditionally, the process involved relationship managers manually gathering documents, verifying business registration with CIPC, conducting know your customer (KYC) checks, assessing credit risk, and configuring products across multiple disconnected systems. This could stretch across several hours over multiple days. The cost of acquiring each customer through this process was high enough that many banks found it economically difficult to serve smaller businesses at scale.
Now, that entire process is being compressed into as little as 20 minutes using biometric identity verification, automated CIPC lookups, real-time credit bureau integration, and instant KYC orchestration. The economic equation shifts completely. Banks can now profitably serve SMEs that were previously unviable as customers, expanding financial inclusion while reducing administrative burden for business owners who previously had to spend days navigating paperwork.
Absa made a significant move in this space in September last year, partnering with Salesforce to deploy Agentforce — the cloud-based CRM platform’s enterprise agentic AI solution — on African soil for the first time. The deployment involves three distinct AI agents being tested within the bank, each targeting a different operational challenge.
The first agent functions as a copilot for relationship managers. Previously, these professionals were spending up to 75% of their working day in preparation — pulling customer data from multiple systems, reviewing transaction histories, identifying potential opportunities, and drafting engagement notes. Only 25% of their time was actually spent speaking with clients. The AI agent flips this ratio dramatically, generating pre-meeting briefs, consolidating customer intelligence, and capturing post-meeting notes automatically. The result is a 75% to 90% reduction in administrative time, freeing relationship managers to do what they were hired to do — build client relationships.
The second agent handles customer enquiries as a multilingual support assistant capable of communicating in 11 South African languages, including isiZulu, Sesotho, and isiXhosa. Early data shows that approximately 40% of customer service queries are now resolved without any human intervention, with responses delivered instantly. Queries that previously kept customers waiting up to 30 minutes are now answered in real time, with success rates approaching 99%.
The third agent focuses on monitoring and resolving technical system issues around the clock. Since its introduction, it has handled more than 6,400 internal support queries with a 96% success rate — meaning the vast majority of technical problems are resolved automatically, without escalation to specialist support teams. This frees technical staff to concentrate on more complex challenges rather than routine troubleshooting.
Looking further ahead, the implications extend well beyond operational efficiency. Banks are expected to use AI to rethink credit assessment entirely, moving away from traditional balance sheet analysis toward evaluating transaction patterns and real-time cash flow behaviour. For small business owners, this could translate into faster access to financing and more proactive credit offers aligned to actual financial activity.
Routine banking interactions are also likely to migrate toward digital self-service channels, allowing customers to resolve everyday requests instantly while retaining access to human support for situations that genuinely require it. The shift will be gradual, and much of it will be invisible to the end user — but its cumulative effect will fundamentally change how South Africans experience banking over the coming years.
Agentic AI is still finding its footing, but the direction is unmistakable. As these systems grow more capable and financial institutions grow more confident in deploying them, customers can expect a banking environment that is faster, smarter, and more responsive to individual needs — one where the technology does the heavy lifting, so that the human experience becomes richer, not more complicated.