Trace Id is missing
October 15, 2024

Capitec Bank employees save more than 1 hour per week with Microsoft 365 Copilot and Azure Open AI

Capitec Bank, South Africa’s largest retail bank centered on customers, strives to keep things simple, affordable, and accessible. The bank's Head of Data Transformation wanted to make things even simpler for employees.

Starting with Copilot for Power BI, exploring Copilot Studio, and graduating with Azure OpenAI, Capitec streamlined a number of processes across various departments in the bank.

With Copilot and Azure OpenAI Service, Capitec’s employees save one hour per week, enhancing efficiency and driving innovation across departments

Capitec Bank

Established in 2001, Capitec Bank serves over 22 million clients with a focus on simplicity, affordability, and accessibility. Operating 850+ branches and employing more than 16,000 people, it continuously explores technology to serve clients better and simplify its operations. "At Capitec, we continuously push the boundaries, question conventions, and adapt to the growing demands of our clients. ‘Better never rests’ embodies our unwavering commitment to excellence, constant improvement and forward momentum," explains Michael O'Carroll, Head of Data Transformation at Capitec Bank.

Gradual yet decisive change

As an early adopter of technologies, O’Carroll has always been fascinated with AI. He started experimenting with custom chatbots 10 years ago and today leads a company-wide effort to improve efficiency and augment employee capabilities with GenAI. “Generative Al has the potential to catalyze an era of transformation. It stands to reshape industries, enhance productivity, and forge new paradigms of innovation and human-machine collaboration,” O’Carroll shares. Leading a small team of engineers, instructional designers, and change managers in a transformation office, he steers discussions on the latest generative AI technologies. “I work with various departments—marketing, HR, finance, audit, risk, and product teams—who come to me with ideas for generative AI use cases. In each area, I’ve set up a champion or lead driver who identifies, explores and prioritizes use cases,” he adds.

However, Capitec treads carefully when introducing new generative AI tools, aiming to manage risks, drive responsible use, and always keep a human in the loop. So, the bank employs the ADKAR (Awareness, Desire, Knowledge, Ability, and Reinforcement) model of change management, ensuring that employees have the necessary skills, knowledge, and motivation to adopt these AI tools. “Generative Al has potential to enhance client experience and increase operational efficiency. By aligning these goals with individual roles, the transformation feels relevant and beneficial to employees,” shares O’Carroll. With initiatives like creating a Centre of Excellence and a Community of Practice, the company gradually builds employee capacity and knowledge in Generative AI

Exploring AI solutions

With this calculated approach, Capitec Bank started introducing generative AI solutions. By integrating Copilot with Smart Narrative in Power BI, O'Carroll saw an opportunity to automatically generate insights on the daily reports used across the bank’s 850 branches. Staff previously received Power BI reports with pages of data, which had to be manually reviewed before meetings. “Using Microsoft 365 Copilot’s Smart Narrative feature, the reporting process was streamlined. Now, people just click a button, and Copilot generates a report summary for consideration," highlights O'Carroll. “It was a good first project because it scaled easily. We have 850 branches with about 10 to 20 people in each, so a huge number of people quickly realized the benefits of Copilot.”

Using Microsoft 365 Copilot’s Smart Narrative feature, the [Power BI] reporting process was streamlined. Now, people just click a button, and Copilot generates a report summary for consideration.

Michael O'Carroll, Head of Data Transformation, Capitec Bank

Capitec Bank’s adoption of Copilot continued to expand with Microsoft 365 Copilot. “Microsoft 365 Copilot enabled our people to test, play, and run their own prompts in a fairly low-risk environment, streamlining internal processes. We ran hundreds of demos, building enthusiasm among our staff for Microsoft 365 Copilot,” shares O’Carroll.

Capitec also saw how extending Copilot with a finance agent could further unlock efficiencies. One striking example involved a finance team member who was responsible for a reconciliation process. "It previously took her six hours each week to run her reconciliations. It involved too many lookup formulas, and she was constantly stressed," says O’Carroll. Capitec turned to Microsoft 365 Copilot for Finance, a Copilot agent purpose-built to support financial processes. “With Microsoft 365 Copilot for Finance, the reconciliation process was reduced to just one minute per session, saving our colleague six hours a week. And the fact that she now trusts the process makes a big difference."

With Microsoft 365 Copilot for Finance, the reconciliation process was reduced to just one minute per session, saving our colleague six hours a week. And the fact that she now trusts the process makes a big difference.

Michael O'Carroll, Head of Data Transformation, Capitec Bank

Building solutions from scratch

Seeking to develop even more advanced AI solutions, Capitec Bank turned to Microsoft Copilot Studio to build custom copilots. “Copilot Studio gave us a well-packaged, safe, and secure platform within Microsoft 365 to learn, grow, and explore,” O’Carroll describes. “Copilot Studio remains cost-effective and low maintenance, making it a good steppingstone in our journey.”

Once the team grew comfortable with the capabilities of Copilot Studio, their ambitions grew bigger. “We realized that if we’re going to scale, we would need a bigger platform. Azure OpenAI Service made sense,” recalls O’Carroll. It took about a month of learning to get the first solution running. As the bank’s GenAI Center of Excellence developed, it became far more efficient. “Previously, it took a month to get a chatbot running on Azure OpenAI. Now, we can do it in a day because we’ve built the capacity to support development, monitoring, and testing,” O’Carroll shares. This increased agility enabled Capitec to experiment rapidly, implementing Generative AI-driven solutions across the business.

Cross-departmental improvements

O’Carroll worked with various departments like marketing, human resources (HR), finance, and risk management to implement Generative AI solutions tailored to their needs. One such bespoke solution is for customer service consultants within the branch. “A client might ask about a specific product, and the branch staff might not have memorized that item,” elaborates O’Carroll. “So, we’ve implemented a chatbot using Azure OpenAI, which will be available on every branch employee’s screen. It has 1,000 product-related questions along with answers stored in a database, so people can just engage the chatbot and respond quicker to clients, serving them better, and selling more effectively.” The bank is now experimenting with adding the speech-to-text capability to the agent, empowering teams to talk to the chatbot. “Direct speech engagement to our data assets will open up a myriad of new transformation initiatives as we can now talk to our data,” O’Carroll adds.

Another example relates to the chatbot that was developed on Azure OpenAI and grounded with the last 5 years of bank’s integrated annual reports. “This has enabled teams within Capitec Bank to review how they can enhance sustainability reporting against global standards like the Global Reporting Initiative (GRI). This ultimately enables us to enhance the quality of information that we disclose to the market,” O’Carroll adds.

In other examples, the communication team started generating unique images tailored to its clients and the South African market with models like DalleE3, instead of using generic stock images. “We’ve created amazing imagery for internal communications,” shares O’Carroll. The legal team started analyzing contracts with Generative AI, extracting information from contracts and comparing it against the bank’s criteria. The human resources team began perfecting job descriptions and reviewing resumes faster with Generative AI. The audit team also uses Copilot to explore optimized risk control reviews and testing processes.

The bank also used Azure Open AI to develop a Copy-write chatbot which is being used to review and rewrite internal content using the bank’s specific style, tone and writing requirements. “This agent helps to standardize how we communicate with our teams and improve the speed at which new content is generated, reviewed and approved,” O’Carroll says.

With Generative AI solutions being implemented across various departments, teams soon started noticing efficiency gains. “We did a survey that revealed around 80% of our staff with access to Microsoft 365 Copilot save an hour per week. Some of the respondents are saving up to five hours a week,” O’Carroll points.

80% of our staff with access to Microsoft 365 Copilot save an hour per week. Some of the respondents are saving up to five hours a week.

Michael O'Carroll, Head of Data Transformation, Capitec Bank

For O’Carroll, the key benefit is empowering employees to achieve more. “It’s not only about efficiency but also about increasing creativity and innovation,” he shares. “A lot of tasks that were previously impossible are now achievable with our Generative AI. This process has demonstrated that these tools can transform our operations and push boundaries in ways we hadn’t considered before.”

A lot of tasks that were previously impossible are now achievable with our Generative AI.

Michael O'Carroll, Head of Data Transformation, Capitec Bank

Looking ahead to future success

As Capitec Bank looks to the future, it recognizes that Generative AI will require a strong foundation. “Our strategy now is to continue building a solid governance framework for the responsible use of Generative AI. We've rewritten our responsible Generative AI usage standard, developed trainings and began revising job descriptions,” O’Carroll emphasizes. This approach will ensure that everyone in the bank, from entry-level employees to the chief operating officer, is equipped to handle Generative AI responsibly. “As our use-cases mature, we will be able to run even faster and enable more innovation across the bank,” O’Carroll says.

Beyond augmented workforce, Capitec’s Generative AI initiatives have sparked new opportunities in areas the bank hadn’t previously considered, such as search engine and large language model optimization. By improving how clients search for information about products and services, the bank continues to refine its client engagement strategies. “Copilot sparked a new way of thinking,” O’Carroll shares. “This shift in perspective has led us to rethink processes, embrace innovation, and embed Generative AI throughout the bank’s operations, laying the groundwork for future success.”

Take the next step

Fuel innovation with Microsoft

Talk to an expert about custom solutions

Let us help you create customized solutions and achieve your unique business goals.

Drive results with proven solutions

Achieve more with the products and solutions that helped our customers reach their goals.

Follow Microsoft