Nozipho Duma, AI programme business adoption and change management lead at FNB, shares insights on the bank's journey to equip its workforce with critical AI skills and create a culture of AI innovation.
According to McKinsey research, the AI industry for banking is projected to become a trillion-dollar sector, with significant value expected from both advanced and traditional AI applications across marketing, sales, risk and human resources. Subsequently, major organisations like FNB are scaling their investments in AI capabilities.
In 2023, FNB launched its AI programme to encourage employees to engage with and apply the tool more. Nozipho says they have seen a significant shift in the company’s approach to AI since then.
The bank reports that over 4,000 front-line employees now use Microsoft Sales Copilot as a virtual assistant embedded into Microsoft Dynamics 365, Outlook and Teams, providing customer context, helping with email drafting, improving customer experience and increasing sales productivity. Additionally, more than 5,200 employees access GPTatWork, the bank's custom AI agent that delivers group knowledge and insights to staff.
The bank has also developed specialised tools for its technical teams. Over 1,500 data scientists, analysts and developers now utilise purpose-built AI agents like CodeAssist and Codify to optimise development experiences and demystify complex code, helping these specialists work more efficiently.
“We’ve seen a clear mindset shift, with employees no longer waiting for AI solutions – they’re co-creating them,” says Nozipho. This transformation is supported by dedicated AI working groups and collaboration sessions that bring together teams from business, technology, governance, risk, architecture, platform and change management.
Nozipho notes that the bank has seen “a growing wave of AI-led innovation ideas submitted through our internal platforms, with employees shifting from passive adoption to active ownership – translating their ideas into tangible initiatives and projects that now form part of our AI programme”.
The bank’s customer-facing agent, for example, now facilitates over 500,000 secure chat interactions monthly across FNB and RMB brands, with positive feedback from both employees and customers.
Simplifying access
One employee testimonial highlights practical benefits: “FNB App Search has been really helping me to find buttons for other services and products offered by the bank that may not appear immediately on the Home Page.”
Similarly, customers have noticed improvements, with one noting: “The virtual banking assistant is cracking so many customer-related issues... and is available 24/7.”
In marketing and communications, AI models are being deployed to create personalised campaign content at scale, showing early signs of improved engagement and conversion rates across digital banking platforms. Operations teams are leveraging AI-powered Optical Character Recognition (OCR) to automate document handling and income verification – tasks that were previously time-consuming and error-prone.
Internal tools like GPTatWork are also proving value beyond technical departments, supporting functions from HR to compliance by simplifying access to organisational knowledge. As one employee observed, “it’s helpful for finding policy documents – like the social media policy, bereavement leave, or even the group’s AI policy – without needing to search through folders or emails”.
Given the sensitive nature of financial services, FNB prioritises responsible AI development and usage. The bank has embedded AI governance as a dedicated workstream within its programme, with comprehensive frameworks to address risks ranging from model risk and data privacy to ethics, bias, cybersecurity, and vendor management.
Balancing innovation, compliance
Nozipho explains that: “Our governance teams, which include representatives from operations, risk, legal, tech, and data science/ architecture, ensure AI is assessed through a multi-disciplinary lens.” These governance frameworks are actively integrated into project lifecycles, though the bank acknowledges that balancing innovation with compliance presents ongoing challenges.
The programme teaches employees to understand AI's fundamentals, assess its applications critically and identify opportunities where AI can add value to the bank’s operations and customer service.
Despite impressive progress, FNB’s AI journey hasn’t been without hurdles. AI fluency and availability of key talent remain constraints, with some teams overly reliant on a limited number of specialists. Skill gaps, rapidly evolving technologies and a growing pipeline of projects often stretch resources thin.
Technical integration presents another challenge. As Nozipho acknowledges;
“integrating AI into existing development environments and enterprise workflows isn’t always plug-and-play. Technical complexity and legacy constraints demand careful orchestration”.
Adoption varies across business units and tools. While there have been early success stories with AI productivity tools, some frontline and customer-facing employees still need additional support and confidence-building. The bank recognises that making AI solutions intuitive, tailored, and trusted is crucial for wider adoption.
Nozipho emphasises that “AI adoption is not just about access. It’s about making it meaningful, usable and trusted”. The bank is addressing these challenges through executive sponsorship, peer storytelling, and outcome-driven initiatives that cater to different maturity levels. Their “change, communication teams and AI champions play a vital role in fostering experimentation, creating safe spaces for learning and enabling feedback loops through forums and collaboration hubs”.
The future of AI at FNB
Looking beyond 2025, FNB plans to further scale AI impact across the organisation through continued focus on literacy, awareness, training, and adoption.
The bank is developing AI fluency pathways for product owners, executives and business leaders. The aim is to strengthen strategic thinking, ethical decision-making and vendor evaluation capabilities.
A unified AI platform experience is in development, where employees can access internal and vendor agents, training and tools in one space. Focus areas include customer-facing AI agents for service and sales, internal agents for operational efficiency and general productivity agents that democratise access to information.
Nozipho explains that their approach is to “empower all employees with the right mix of AI capabilities, ensuring we take learnings from this to continuously improve agent performance and increase the scope of coverage. It’s all part of building an AI-native, future-fit workforce – but never losing the personal touch”.
As the financial services industry continues its transformation, FNB’s comprehensive approach to AI literacy serves as a valuable case study for organisations seeking to empower their workforce with the skills needed for an AI-driven future.
