Hyper-Automation and Year-end meet
Hyper-automation (HA) and data represents a fundamental shift in how we solve business challenges, and it also helps to guide our decision-making. At Medscheme, the team used non-human digital workers to great effect to meet extremely tight deadlines for annual benefit design requirements.
The medical scheme industry in South Africa has never been more challenged than right now, given the unknowns of a National Health Insurance (NHI) system and lack of clarity in the future role of medical schemes.
Administrators and managed care companies remain vested in the sustainability of their medical scheme clients, particularly in supporting them with effective benefit design combined with the best tariffs possible, and ensuring their members continuously experience exceptional service delivery. Despite finalising the benefit designs later in the year, medical scheme clients expect systems and set-up to be ready by the start of the benefit year.
Medical schemes are under pressure to effectively balance financial and membership risk.
World-class automation
Within Medscheme, the use of a world-class decision-rules engine has enabled an extremely high level of claims automation together with the ability to easily configure complex claims rules. Currently, only 1% of associated provider claims require any form of human intervention.
The timeous, accurate payment of claims is a key deliverable as scheme members expect their medical schemes to expertly settle their medical-related expenses. Meeting these exacting service expectations requires that we have the appropriate smart technology solutions with the necessary flexibility and agility to effectively meet the annual benefit design requirements of our client medical schemes.
Since we received the 2019 FICO (Fair Isaac Corporation) international award for innovation, our claims capability has continued to grow. The decision engine now houses rules for multiple considerations, with approximately 10,5 million claims processed each month. To configure a new rule can now be completed in a matter of hours and with no system downtime. For example, a recent new scheme take-on took just three days to configure in its entirety, compared to taking weeks previously.
Scheme limits remain held within the core transaction processing system (Nexus) and are applied once the claim is ready to pay. Loading new limits and rates typically took many days of human intervention at each year-end. Because of the time-heavy nature of this work, the Digital Business Solutions team together with Afrotech seized the challenge to automate this task to reduce the time and improve the accuracy compared to when done manually.
Navigating complex operations
Over the years, benefit designs have become increasingly complex, and the Regulator approval of scheme rules takes longer, resulting in a much-reduced timeframe to meet the deadline. For Medscheme, this means the challenge during this time is to make sure that the benefit design needs of all 11 clients with almost 50 different scheme options are ready by 1 January of each new benefit year.
The full process takes scheme administrators two to three months to complete all system development, rule setups and testing to be ready to process and pay claims correctly from 1 January. Failing this, claims are placed on hold and only released once the intended changes have been completed. This means that members, healthcare providers and facilities could experience delays or errors in payments – which is an unacceptable result.
What did we do differently last year-end?
To meet these challenges, Medscheme has grown its hyper automation capability over the past five years to take on new automation challenges supported by its partner, Cogent.
Automated process solutions were designed for several key year-end functions, with the aim to use non-human digital workers (robotic automation or bots) to reduce the time to deliver, reduce errors, and remove the mundane, repetitive tasks, allowing human workers time to focus on the cognitive work of validating and approving rule testing outcomes.
The process design ensured the necessary controls were in place to mitigate errors and any risk associated with rule and system set-up errors. The process automation had checkpoints to validate its own accuracy and efficacy.
What was achieved?
Despite a significant delay prior to the process being able to start, the responsible teams with the support of digital workers were able to meet the 1 January deadline. The digital workers collectively performed the functions within 111 days, whereas it would have taken our human staff 350 days to achieve the same outcome. To further demonstrate the effectiveness of the automation, and if one excludes the running of the test scenario packs and the handoffs to humans during the testing process (after limits and rates were loaded), the digital workers completed their work within six days compared to previously about 236 days – a massive improvement in efficiency.
These innovations have been driven by a need to be flexible, reduce production time, reduce errors, and offer clients a competitive advantage in their benefit design whilst not compromising on the member experience.
A fundamental shift
Hyper-automation (HA) and associated processes represent a fundamental shift in how we now approach, react to and solve problems. More importantly, given the availability of data, it should help translate and guide our decision making, based on data. While robotic process automation (RPA) handles routine tasks, trained industry professionals are better utilised for tasks requiring creativity, emotional intelligence, empathy and strategic thinking.
Aside from additional process improvements learnt on our current functionality, the next steps will be to explore how we use generative AI together with the digital workers to further accelerate the full process for the next exciting year-end period.
"The only limit to our realization of tomorrow is our doubts of today."
