According to a recent Trans- Union survey, cash-strapped South Africans are shopping around for the best car and house insurance deals. The survey found that 59% of consumers had looked for more affordable car insurance in the past six months, while 41% planned to do so in the same time frame. The survey also revealed that more than half (55%) of those who compared options ended up switching insurers. The market is also getting more competitive, as new players enter.
TransUnion South Africa’s insurance vertical sales leader, Schalk Fischer, believes that the Holy Grail in the South African insurance space is finding ways to grow the collective pie. “It’s important to ask how we identify people who are not being serviced by the market and serve them so that the pie can get a little bigger.”
Meanwhile, many consumers are cutting discretionary spending, like insurance, so they can afford essential items.
“Retaining existing clients and/or tapping into the untapped insurable market means insurers must fight hard to stay relevant,” Fischer says. This is pushing the industry to become more innovative. Insurers have realised there are efficiencies to be gained by using analytics and big data platforms to create dynamic risk profiles and flag potentially fraudulent activity, Fischer says, or offer tailored insurance products and streamline claims submissions. Many are partnering with businesses such as TransUnion to gain access to broader and richer types of data so that they can better understand consumers and market their products in new ways.
South Africa's total insurance premium volume was estimated at R840.84bn in 2023. New players such as Naked and Pineapple are disrupting the industry, especially in personal insurance such as car, household, travel, and all-risk cover. But the market is still dominated by brands like Santam, Hollard, Old Mutual, and Liberty.
Clean, relevant, quality information applied in the right way is far more powerful than having endless volumes of unorganised data.
Sumarie Greybe, Naked
Fischer says there are about 12mn vehicles on the country’s roads. Of those that are registered, only between 30% and 40% are insured. “This has a huge impact on the insurance industry,” he says. “If you’re in an accident, there’s a one in seven chance that the other driver is uninsured. As more uninsured vehicles take to the roads, costs rise for insured drivers because insurers can’t recover expenses from third parties. And this drives up premiums.”
In theory, insurers could adjust premiums based on the likelihood of encountering uninsured drivers in specific regions, and drivers in areas with higher uninsured rates could be offered specialised cover options. Insurance businesses could also offer low-cost, entry level products, like pay-per-kilometre or accident-only cover, to encourage uninsured drivers to get insurance.
Sumarie Greybe, co-founder of Naked, says most people want simple and adaptable insurance products. Naked customers can change their cover instantly on its app, without paperwork or call centres. They can also pause accident cover on days they aren’t driving to save on their premium, or cancel their policy on the app without needing to speak to someone. Like most insurers, it has a constant flow of unstructured data coming into its business, says Greybe, such as call transcripts and selfie videos/photos and a wide range of supporting documents for claims. In traditional insurance companies, this data would be processed manually by large teams of back-office workers.
Naked uses AI models to parse unstructured information, turning it into structured data that supports automated business processes and enables quick and accurate decisions on straightforward claims. “For example, we can settle a windscreen claim almost as soon as a customer submits a photo of the damage to their windscreen,” she says.
When a would-be customer is looking for a quote, Naked’s chatbot, called Rose, will ask them a series of questions in natural language. Under the hood, algorithms will assess their risk profile and come up with a final quote in under 90 seconds. Another example of how Naked is using AI and data is in fraud prevention. An AI-driven claims system can track thousands of data points from a range of sources and systems to identify patterns that emerge in fraudulent claims. According to Greybe, this system outperforms traditional insurance in flagging fraudulent cases. “Reducing fraudulent payouts reduces costs, which results in lower premiums for honest, low-risk customers, who are in the vast majority. And it helps insurers make faster claims decisions, so that valid claims are settled as soon as possible.”
According to TransUnion’s Fischer, most of the fraudulent activity in vehicle accident claims is related to who was driving the vehicle at the time of an accident. “If someone other than the main driver was driving a car when an accident took place, insurers can now use things like telematics or GPS data to double-check if the person who claims to have been driving was anywhere close to where the accident happened at the time of the accident,” he says. Similarly, when someone moves into a new home, one of the first things they do is apply for home contents insurance. And one of the first things your insurer will want to know is whether your new home has a thatched roof. “You’d be surprised how many people lie about this kind of thing. The number one rule as an insurance customer is to never lie,” he says. “Today, insurers can use satellite imagery to check if you do, in act, have a thatch roof, and if you lie, you won’t get paid out.”
Dis-Chem launched its life insurance offering, Dis- Chem Life, in March this year. Greg Smith, Dis-Chem Life CEO, says it will use the data it already has about its customers to build them insurance products. Customers with chronic conditions such as diabetes, hypertension and HIV are traditionally seen as high-risk, irrespective of how well they manage their health.
He says insurance is built on spreading risk across members, so if someone is statistically more likely to need care, their premiums are adjusted upward. This makes it expensive for these people to take out life insurance because their premiums are loaded based on assumptions and not an individual’s behaviour.
Smith says a Dis-Chem store has a clinic, pharmacy and the retail shop, and with around 15mn customers, “the amount of data you’re generating is just endless. If we just look at the first two components of the ecosystem – the clinic and the pharmacy – we have information that allows us to tailor premiums to individuals because we can link data around their health checks and medication adherence.” If, for example, a diabetic client comes to the clinic for a checkup, Dis-Chem can monitor their blood glucose levels at regular intervals, and keep track of their adherence to their medication. If they’re taking their medication diligently, it’s clear that they’re trying to look after their health, he says. “Adherence to medication is directly proportional to being healthier.”
By using customer data, its ecosystem enables it to reward good behaviour by giving customers who try to be healthy benefits or discounts on their premiums. “In this way, we can treat a person with a chronic condition in the same way as we would treat a healthy person,” Smith says. He adds that customers who undergo regular health checks at Dis-Chem clinics and proactively manage their health will begin to see dividends from its rewards programme.
“Gone are the days when an insurer would assess your risk profile and that was just who you were for the next 40 years,” says Smith. Insurers can now use other metrics and data sources, and it’s possible to track changes over time and offer more tailored products and more dynamic pricing models. These alternative data sources can also be used to identify small, niche groups of new customers.
But the value of all of this data really lies in context. “If you have data, but you don't understand the context of that data, it becomes meaningless. You can have all the data, but you need to be able to convert that data into action if you want it to add value,” says TransUnion’s Fischer.
Naked’s Greybe has a similar view. Too much data can become a burden if it’s poorly structured, of low quality, or locked away in silos, she says. “Clean, relevant, quality information applied in the right way is far more powerful than having endless volumes of unorganised data.”
Fischer believes that using data properly will mean insurers will have a more granular view of their market segments, and that products can be tailored to address their needs. “In my view, it is less about trying to identify and tap into generic, big markets and more about drilling down into different, smaller segments and then using technology, data and tools like AI to sell to these people in very strategic ways.”
* Article first published on brainstorm.itweb.co.za
Share