As Data Officer for Retail Banking Operations of Standard Bank South Africa, Itumeleng Monale knows that strong data governance and management are the foundational capability needed to ensure that teams across Business Intelligence and Data Analytics are better able to perform their functions of data exploitation and monetization.  She has a keen interest in the proliferation of AI, not only for the purposes of data exploitation but for enhanced customer experiences.
What was the business opportunity you sought to address by using AI?
Our organization was racing down a path to invest in data science initiatives, including AI and, paying data scientists who weren't able to achieve these aspirations because the fundamentals of data management were not in place.  There are numerous use cases for AI in a banking context. The data was however of poor quality, difficult to access and our legacy architecture meant that data was just in overall disarray across the organization. For a bank, this meant that only certain islands of data sources were kept "holy" for business-critical outcomes such as credit scoring, but everything else was sub-optimal. All our models initially were being applied to data that was at quality levels of approximately 20%.  We then embarked on a journey of four years of implementing a robust data management practice that would enable data exploitation to reap business benefits through various use cases and applications - from RPA, to supporting the building of accelerated models for data analysis and using AI.
What has your company achieved with AI?
Data monetization is becoming an increasing paradigm in our organization - to the extent that we are able to quantify the financial uplift from data analytics work above and beyond BAU. Relationship bankers utilizing some of the models the analytics teams have applied are experiencing 40% revenue uplift relative to their peers who have not adopted the personalized offer tools that have been made available to them.

Using intelligence to automate our metadata management processes has shrunk our analysis cycle for new systems migrations and implementations down to 20% of the time it used to take. This is accelerating many business initiatives, including our migration of work packages to the cloud.

Another part of our bank is utilizing AI to test and validate API's as we increase our modular delivery of digital features. We definitely see AI as the tool through which we will accelerate business delivery across various disciplines; free up human capacity to do value-added and relationship building work and elevate tasks to root cause-based problem solving and higher-order design engagements.
What advice would you share with others who are considering using AI?
The initial process can feel slow but making the right efforts early on will accelerate it and create benefits down the line. It is very important to never automate anything you don’t fundamentally understand. The algorithms may be very complex for human comprehension, but the underlying principles must be clear.
Why should women be more involved in the creation and use of AI?
AI is currently the technical manifestation of thought processes that were, initially, human. With this in mind, the more diverse the people who shape it, the more representative the machine outcomes will become. Woman make up just about half of the global population, to get optimal outcomes, their voices must be taken into account. Secondly, on average, women are empirically proven to impact more human beings around them per individual relative to men. So just based on good common sense and what is best for humanity, you really can’t exclude women.