South Africa’s home services market has had a trust problem for years. Finding a plumber, electrician, or builder through any available channel typically means navigating referrals, WhatsApp groups, Facebook recommendations, and the occasionally useful Yellow Pages successor. The platforms that have tried to bring order to this haven’t solved the underlying issue: knowing who the pro is before they arrive, and understanding what you actually need before you book. Kandua’s new AI interface, Jess, attempts to fix both at once.
Launched today by Kandua, a wholly owned Santam subsidiary, Jess is a conversational AI interface built into kandua.com. Through its “Ask Jess” function, homeowners describe a problem in plain language and the system helps them diagnose what’s actually wrong before connecting them to a vetted service provider through a “Book a Job” flow. The journey stays within a single interface, which matters more than it might seem. Most existing platforms either help you find someone or help you understand the problem; Jess tries to do both, and the drop-off between those two steps is where a lot of service requests fall apart.
Locally, the closest competitors are the home services offerings embedded inside banking apps. FNB’s nav» Home Services gives app users access to a marketplace of vetted SMEs, and similar plays have come from Nedbank and Vodacom, positioning professional services access as a value-add to their core products. None of these have introduced a conversational diagnostic layer. The distinction matters because the problem Jess is solving isn’t availability of contractors; Kandua already connects over 40,000 vetted home service companies to roughly R50 million worth of work each month. The problem is earlier in the process: most homeowners can’t accurately describe what they need, which leads to mismatched bookings, wasted time for both parties, and service provider frustration.
That’s where the platform model shift is also significant. Kandua has moved away from a pay-per-lead structure in favour of one-to-one job matching, where service providers are only connected to work they’re equipped to complete. We’ve written before about how inDrive’s transparent pricing model disrupted the commission orthodoxy in ride-hailing; Kandua’s shift away from lead-volume incentives carries a similar logic. When the business model aligns with job completion rather than lead generation, the quality floor lifts.
The vetting process underpinning Jess’s provider network is also notable. Providers must pass identity verification, criminal background checks, certification validation, and business registration checks before being listed. CEO Vinolan Pillay describes this as among the most stringent vetting processes in the global home services marketplace sector. Given how frequently South African homeowners have been defrauded by unverified contractors, that credibility mechanism matters more than the AI interface itself. Jess’s conversational front end is only useful if the pros it connects users to are reliable.
Santam acquired Kandua in 2024 as part of a strategy to expand digital innovation and deepen support for small businesses. The strategic value runs deeper than that framing suggests. Santam is South Africa’s largest short-term insurer, with a market share above 22%. An insurer that owns the platform through which homeowners diagnose and repair home problems has visibility into maintenance activity before claims are filed. The better the quality of repairs facilitated through Kandua, the better the risk profile of the properties it insures. Jess’s cost benchmarking and complication-flagging features also serve Santam’s claims environment, potentially shifting more incidents toward early, competent remediation rather than late, expensive claims.
This isn’t a knock against the product. It’s the kind of structural alignment between platform value and parent company incentives that makes this more durable than a standalone startup with the same feature set. Kandua has institutional backing, an established provider network, and an AI interface that addresses a real behaviour problem in home services discovery.
What Jess doesn’t yet do is anything that requires physical context. The diagnostic capability is based on user descriptions, which means its accuracy depends entirely on how well someone can articulate what they’re seeing. A leaking geyser that’s also tripping the circuit breaker is diagnosable from a description; intermittent electrical faults or slow structural problems are harder to communicate accurately. As AI diagnostic models improve, the gap between what a user can describe and what a platform can understand will narrow. For now, Jess is most useful for common, describable problems.
The interface is live at kandua.com and available without registration for the diagnostic function, which removes friction at exactly the moment a homeowner is stressed about a leak or an outage.
The category has needed a proper diagnostic layer for a long time. Whether Jess delivers on that in practice will depend on how well its AI handles the messy, ambiguous descriptions that real homeowners give it under pressure.