Robert Benvenuti has transformed how one of Africa’s largest financial institutions approaches data and automation. As the CIO of Data and Applied AI at Absa, he has successfully moved the organization past the pilot phase into a world where machine learning drives real-time results for millions of customers. This conversation explores how these technologies are alleviating the burden on customers in financial distress, the intricate balance between automation and human judgment, and the rigorous security frameworks required to keep a modern bank safe from emerging digital threats. We also look at the bank’s internal educational initiatives and why they view technology not as a separate expense, but as a fundamental multiplier for growth across the continent.
Processing debt applications often requires manually reviewing dozens of attachments, which can delay relief for customers in distress; how has your approach changed this dynamic?
In the past, our employees were bogged down by a sea of paperwork, often having to open emails and manually parse through 20 to 30 attachments for a single application. This labor-intensive process meant that customers already facing the heavy emotional weight of financial hardship had to wait four days for a response. By implementing AI-enabled optical character recognition, we have successfully slashed that turnaround time down to just one day. It’s not just a technical win; it’s about providing faster relief to people when they need it most. Watching the system extract relevant data with precision allows our teams to focus on the human side of financial recovery rather than just the exhaustion of manual data entry.
There is a lot of anxiety surrounding the idea of automation replacing jobs, so how are you positioning technology to augment your workforce instead?
We strongly believe that AI is here to enable the human, not replace them, which is a philosophy grounded in human-centered AI principles. We want our people to be the ones making the critical judgment calls and handling the nuanced, sensitive interactions with our clients. By stripping away the repetitive, soul-crushing tasks—like reading thousands of identity documents, passports, or electricity bills—we free up their mental energy for higher-level decision-making. It’s about taking the “robot” out of the human so they can do what they do best. This shift creates a more fulfilling work environment where employees feel their expertise is being used effectively rather than just performing rote calculations.
Upskilling a large organization is a massive undertaking; what does the internal journey look like for the thousands of employees learning to work alongside these tools?
To ensure no one is left behind, we have established a dedicated AI Academy that offers personalized learning journeys for our staff to complete at their own pace. So far, more than 10,000 employees have completed immersive training, which has transformed the way they view and interact with technology. We don’t just want them to use the tools; we want them to understand the underlying mechanics so they can innovate within their own departments. These internal initiatives have fostered a culture of curiosity rather than fear. It’s rewarding to see staff from various backgrounds gain the confidence to apply machine learning to their daily workflows to improve the customer experience.
With the rise of generative tools, how do you provide employees with the freedom to experiment while maintaining the strict security required by a bank?
Trust is the bedrock of banking, so we had to be very deliberate about how we opened the door to large language models. We created a secure AI gateway that gives our staff access to over 120 different models, but it operates within a strictly governed environment. This setup includes more than 40 specific guardrails that address everything from data protection and bias to cybersecurity and responsible usage. By providing these tools officially, we significantly reduce the risk of “shadow AI,” where employees might use unapproved and unsecure external applications. This balance allows for rapid experimentation without ever compromising the sensitive data our customers entrust to us.
Financial crime is becoming more sophisticated every day; in what ways are you using machine learning to stay ahead of bad actors?
We are using machine learning to cast a much wider and more accurate net over activities like money laundering, human trafficking, and terrorist financing. Our systems are now capable of identifying subtle patterns of suspicious activity that would be almost impossible for a human to spot in real-time across millions of transactions. We even deployed a specialized AI agent specifically to scan our email infrastructure for phishing attempts, stopping malicious links before they ever reach an inbox. Our efforts were recognized at the International Compliance Awards in London, which validates our strategy of using AI to find hidden software vulnerabilities. It’s a constant arms race, but these models allow our investigators to narrow down potential cases with much greater speed and precision.
Many firms treat innovation as a separate line item, but why has your organization chosen to avoid a standalone AI budget?
We don’t see AI as a separate bucket or just another tool to buy; it is an integral part of how we solve business problems. Every technology investment we make is evaluated based on the specific value it delivers to the customer and the business, rather than just its novelty. By integrating AI into our broader cloud and data strategy, we ensure that every project is grounded in reality and tied to a tangible outcome. This approach prevents us from chasing trends and keeps our focus on delivering personalized financial services across our African operations. It’s about building a foundation of trusted data that empowers every part of the bank, from customer onboarding to complex risk management.
What is your forecast for the role of AI in financial services across the African continent over the next few years?
I see AI as the great multiplier for financial services across Africa, as it will make it significantly more affordable to offer highly customized banking to millions of people. We will reach a point where services are automatically tailored to a customer’s specific language, local geographical context, and unique product needs without the massive overhead costs of the past. While we remain cautious about giving autonomous agents full control over bank accounts—security must always come first—the communication between agents and banks will eventually become a more common standard for efficiency. The integration of cloud infrastructure and trusted data will allow us to bridge the gap in financial inclusion faster than ever before. This technology isn’t just an upgrade; it is the engine that will drive a more accessible and personalized economic future for the entire region.