Behavioural Banking with Hylton Kallner

In this Huawei FSI Series conversation, Discovery Bank CEO Hylton Kallner tells Colin Iles why a health insurer built a bank, and why the whole thing rests on removing the cross subsidy between customers who manage money well and those who don't. Kallner, a 25-year Discovery veteran and actuary by training, walks through the founding purpose, the branchless cost structure, an admitted early mistake with Discovery Miles, and how AI and grocery basket data are already shaping personalised financial advice.

The behavioural wedge: why Discovery built a bank instead of buying one

Kallner says Discovery Bank was built on the same purpose that drives the rest of the group: making people healthier, extended into financial behaviour. He describes the model as a "behavioural wedge," the banking equivalent of flattening a mortality curve: use incentives, nudges and data to push customers toward lower default rates and better savings habits, in a way that benefits both the customer and the bank. He is explicit that the business case only exists because the product works first. "We're product led, and it's only once we have a product breakthrough that we believe the business has got legs," he says, describing how the bank's business plan followed the product, not the other way round.

Killing cross subsidies: how Vitality Money prices good behaviour

Kallner argues traditional banks cross-subsidise bad debts by charging good and bad customers similar rates, because pricing has historically been segmented by income rather than behaviour. Discovery's Vitality Money system instead scores customers on actual financial behaviour, emergency savings, insurance coverage across the whole market, and settlement history, regardless of income level. He states plainly that Discovery Bank has recorded zero defaults at its highest Vitality Money statuses since launch, and that customers who manage money well are offered the best interest rates and rewards in the market, while weaker performers get market-related rates rather than being cross-subsidised by better customers.

The Discovery Miles mistake: why cashbacks failed and Miles came back

Kallner names killing the original Discovery Miles currency at the launch of Discovery Bank as one of the bank's clearest early mistakes. The team believed cashbacks were the better proposition, but customers hated it: cashbacks "just got lost in their bank account" and didn't feel like a reward, whereas Miles could be stored and spent on tangible purchases. Discovery reversed course and rebuilt Miles as what Kallner now calls "the definitive e-currency, or stablecoin, in the group," used to monetise good behaviour across exercise, driving and money management in real time.

No branches, ever: the cost structure behind the pricing

Kallner says branch infrastructure is a structural barrier that protects incumbent banks, and that Discovery Bank could not have launched at all if branches were required. The bank has one nominal head-office branch, closed since the start of COVID, and delivers every service through the mobile app, a website, and a "Live Assist" feature that lets call centre agents see a customer's screen in real time and guide them through the app. Close to 50% of new accounts are opened outside traditional bank hours, which Kallner attributes to not needing physical branch presence for any customer journey.

Grocery baskets and AI: what Discovery already knows about its customers

Kallner describes a data set he calls globally unique: over 20 years of clinical health outcomes data, geolocation and spend data, and SKU-level grocery basket data captured through partnerships with Woolworths and Pick n Pay, where every item is categorised and cashbacks of up to 75% are offered for healthy purchases. Discovery uses this to map basket composition changes to long-term health outcomes and to price credit risk at an individual level. Asked directly whether this data is sold on to advertisers, Kallner is unambiguous: "We don't sell the data at all to third parties."

Innovation discipline: a 25-year weekly meeting and a fixed launch date

Kallner attributes Discovery's innovation culture to formal incentive programmes, a direct staff email channel he personally manages, and a weekly R&D meeting that has run for 25 years, roughly three hours long, 50 weeks a year. He says nine out of ten products that enter that pipeline never reach launch day, and that the fixed annual launch date, drawn originally from the medical scheme industry's January renewal cycle rather than modelled deliberately on Apple, forces prioritisation and keeps teams iterating up to the last minute before presenting on stage.

Does Discovery sell customer data to advertisers?

No. Kallner states directly that Discovery does not sell customer data to third parties, despite holding what he describes as a globally unique data set spanning over 20 years of health outcomes, spend, geolocation and SKU-level grocery basket data. He says data sharing is consent-based and strictly controlled, used only to generate personalised incentives for the customer who provided it.

What's next for Discovery Bank?

Kallner points to a product launch for intermediaries scheduled for the following week, focused on deeper AI-driven personal financial analytics. Home loans and vehicle finance are named as gaps the bank still wants to fill. About 30% of new client acquisition already comes from outside the existing Discovery ecosystem, and the entry-level account costs R10 a month, which Kallner says shows the bank isn't income-constrained in who it can serve.