CHRO Community Conversation hears how AI and analytics enhance the employee experience

Discovery’s Tswelo Kodisang presented a case study on how AI has transformed the company’s people function.

Discovery chief people officer and 2020 CHRO of the Year Tswelo Kodisang impressed HR leaders with a case study on how AI and analytics have been instrumental in transforming the company’s people function, during this week’s CHRO SA Community Conversation, which is sponsored by Workday.

Together with colleagues Kammy Sing, the group head of People Operations, and Meena Makan, the head of DP analytics, Tswelo provided background to the business case for people analytics, and attendees also had the opportunity to view a live demonstration of the people dashboards that are now in place.

“The transformation journey started a while back, with analytics and digitisation as key enablers, as part of our strategy. We have to build a strong people function to support the ambition of the business. To have the best person in every role, we have to offer them an unrivalled employee experience,” he said.

Tswelo further explained, “There were three tenets that we looked at. These were that the candidate experience has to be an inspired one, the employee experience has to be fit for purpose to inspire, motivate and engage, and that we had to create the right type of environment and support for the leader experience as well.”

Within these principles, Discovery was able to look at the employee lifecycle holistically, from talent acquisition and management to culture, values and learning and chart a path forward.

Shared Services 3.0

The plan was to have all people operations on Shared Services 3.0, to leverage robotics and AI so that shared services can service all our operations from anywhere in the world.

“The move is from gut-feel people decision-making to much more data-driven insights, to use predictive analytics and science to measure business impact. We were adamant that we would not lose humanity in all of this,” he highlighted.

Kammy explained that the company has over the past 25 years accumulated a staggering amount of data and information, ranging from wellness to mortality, on its customers. However, from a people function perspective, there was a lack of rich data and insights.

“HR had disjointed systems that did not speak to each other. For example, payroll and learning data were held separately and there were no integration points so the data was viewed in isolation,” she explained.

The journey towards an integrated HR solution and single employee view began in 2017, when the business case was built for people decision-making to be data-driven rather than based on perceptions. This meant that a single source of truth was sacrosanct and data integrity was key.

“The first driver was the customer (our people) and in line with this the desired end state was to create exceptional candidate, employee and employer experiences with self-service features that can be accessed anywhere, anytime and on any device,” she said.

This meant appropriate technology, infrastructure and leveraging multiple data points for that end state, implementing workforce analytics for critical employee information, power BI technology to visualise and present data in an acceptable format, and self-service analytics revolving around the employee and manager experience. It also included stakeholder analysis mapping and taking operational expectations into account.

“We set up a dedicated people analytics team to ensure ownership and accountability. Previously data and analytics was used from across the organisation and HR wasn’t a core function of that,” she added.


This was followed by building a robust data warehouse to aggregate and rearrange data and move legacy data into the warehouse over time. A baseline survey was also conducted with business.

“It was so disappointing, as the results of the baseline showed a rating of 1 out of 5 for maturity and effectiveness of the function. When we went back to business two years after starting the project, the maturity rating was 5, so it showed that we were moving in the right direction,” Kammy said.

She added, “It used to take us three months to issue a quarterly people report. We had to get data from different sources, put that on a PowerPoint. Now we have digital interactive dashboards from multiple sources in real time. It is a valuable asset for business partners.”

This was a long process that took about three years and included upskilling business partners, implementing digital training tools and running masterclasses to better understand the technology and its applications in the course supporting business decision-making.

“We started to see value materialising. Two key considerations when setting up an analytics function is to build for sustainability and scale,” she said.

Meena echoed this view, adding that analytics and tech are used as enablers to optimise business value creation and impact.

“The approach to AI was to start small but design for scale. We recruited data analysts from within the organisation, data scientists were recruited externally and the team also includes data engineers. This complements the data science skills with the understanding of HR.

Predictive modelling

“Now we are using AI and advanced algorithms for predictive analytics for focused guidance for HR. We are also able to use sentiment analysis like employment equity and by 2023 we will transition to be able to deliver augmented analytics. Decisions can be made with some assurance of the outcome, so 20 percent of key insights that are more strategic and has highest business impact can be generated,” she explained.

Some of the capabilities built so far are an interactive people dashboard, recruitment dashboard with pipeline metrics including efficiency measures, self-service portals and reports, supported by virtual assistants and smart bots, finance-based dashboards for budgeting purposes, employee recognition dashboards, fully automated real time monthly KPI dashboards, risk reporting on absenteeism as well as predictive modelling for flight risk.

“The flight risk predictive modelling looks at the likelihood of an employee leaving. The business is then empowered to proactively engage employees who may not be satisfied. It is run every six months and currently reflects 95 percent accuracy. The dashboards can provide tactical insights and offers a holistic view of all elements of the employee journey. It is insights driven, smarter and objective,” she noted.

Tswelo further advised, “It is also important to get the right capabilities in, not to convert HR generalists to data scientists. Transformation is not a step one to five process. It’s more like three steps forward, five back, one to the side. There is still much to be done in leveraging tech, automation and increasing capacity for more strategic work among line and HRBPs.”