On September 1, South Africa’s retirement savings landscape is set to undergo one of its biggest shake-ups since the dawn of democracy 30 years ago. On that day, the two-pot retirement system will come into effect. Retirement product providers can expect a massive uptick in queries – as many as 200 000 or more calls per month.

Could intelligent automation smooth the path forward and avoid thousands of frustrated people queuing for answers? Ryan Falkenberg, CEO of CLEVVA, with deep insight into the challenge, believes virtual agents could smooth the way. He writes

Essentially, the two-pot system aims to preserve retirement savings while allowing fund members to draw a part of their savings in an emergency. It sounds simple but it’s far more complicated than it looks on the surface, largely as the system still needs to work within existing rules and regulations.

This means some people will encounter constraints while others won’t be able to take advantage of it at all. An Allan Gray explainer states “if you are a member of a provident fund and you were a member of that fund and 55 or older on 1 March 2021, you will be excluded from the two-pot system”.

Therefore, retirement product providers are likely to be inundated with queries about the system. Customers will want to know how much they can withdraw, and how to do so, among other things. The Financial Sector Conduct Authority says there are 867 active retirement funds and it’s a safe bet that many aren’t equipped to deal with queries.

Further complicating matters is that generic answers won’t typically suffice. The answer any customer gets will depend on their specific circumstances – and giving the right answer requires a high level of training and competence. A simple chatbot or set of frequently asked questions on the website won’t cut it.

Providers have three options:

1. Wait and see what happens. The expected volumes may not materialise, and the distributed information may be sufficient. The risk is that the query volumes may materialise and the resulting frustration due to long call waiting times may prove very damaging indeed.
2. Alternatively, providers can choose to contract a large number of temporary agents ahead of September. The number required is difficult to predict and they may end up with too many or too few. Either way, these temporary agents will need training to handle financial queries. The costs would be significant and pension fund providers have already said withdrawals will attract fees in a bid to recoup the administrative costs.
3. Another option is to deploy intelligent virtual agents to service queries 24/7. A virtual agent is a digital version of a human, capable of human-like conversations via voice or text channels, that can effectively answer customer queries like a financial expert in the context of each person’s situation. It can also work with business systems to ensure resulting actions are processed and customers receive the desired outcomes without waiting for a human agent.

Virtual agents are not chatbots or voicebots. They really can have conversations like people. As a result, huge volumes of engagements can be automated effectively in line with business and regulatory rules.

Given the risk of getting the call volume predictions wrong, it makes the most sense to invest in a capability that can scale immediately with changing volumes without negatively impacting customers. This takes the pressure off existing human agents and offers customers immediate support, no matter the channel.

With so many people anxious about what the changes mean, providers should be looking to give customers the support and answers they need without the inconvenience of long waiting times. Intelligent automation offers a way to do this at a lower cost to all.”