Lesaka is scaling fast across consumer banking, merchant POS, cash logistics, prepaid and enterprise VAS — with Bank Zero pending. Mohamed's team needs one unified fraud engine across all of it, not more point tools. Fraudio is the centralised AI, multi-tenant platform built for exactly that consolidation, with an active Africa presence through Enza and Omnea.
Lesaka has grown fast since 2022 — through acquisition. Every business unit has come in with its own fraud posture, its own vendor stack, its own operational reality. Bank Zero is next, pending regulatory approval. The fraud, risk and financial crime team's own strategic call was clear: one scalable solution across every subsidiary, not a growing mesh of siloed, product-specific tools. That's exactly the problem Fraudio's centralised, multi-tenant model was built to solve.
Fraudio's multi-tenant architecture provisions a tenant per business unit. Each subsidiary gets its own segregated environment — rules, KPIs, alerts — while Mohamed's central team gets a group-level view across the entire book. One vendor, one contract, one operational language.
Mastercard & Reserve Bank switches
KFC, GAP and enterprise merchants
Drop safes & filling-station loans
Prepaid electricity meterage
22M customers/month via retailers
Centralised AI across every subsidiary. Tenant-isolated. Group-visible.
Fraudio's entity-detection rails cluster and identify normal behaviour versus anomalies for every entity type Lesaka's diverse book contains — not just cardholders. Merchants on the POS side, cash vaults in Cash Connect, ATMs across the consumer bank, retailer terminals in Enterprise VAS — all modelled the same way: peer-group analysis, sequence detection, and pattern-shift alerts.
Authorisation, decline, chargeback and refund ratios per merchant. Behavioural pattern shifts on KFC, GAP and other enterprise merchants. Bust-out and collusion detection.
Drop-safe deposit patterns per filling station. Peer-group analysis across the vault estate. Loan-and-repay anomaly detection for the business loan product.
Transactional monitoring on ATM flows — the layer your current local tool doesn't fully integrate. Sequence-based analysis for card-testing, cash-out and physical device compromise.
The 22M-customer VAS estate at retailer level. Pattern shifts, abnormal spikes, and peer-group deviation across the retailer network — modelled together, not in isolation.
Full CP and CNP card fraud, plus account monitoring for the consumer banking side. Push-payment authentication signals — the capability your ATM tool lacks.
Recharge patterns per meter and per outlet. Fraudulent voucher detection, chargeback-driven pattern shifts, and cross-outlet clustering.
Fraudio spun out of an ING acquirer in 2019, built by data scientists who found the existing fraud detection solutions insufficient. The differentiator is the centralised database — issuing, acquiring, wallets and accounts combined into one homogeneous schema for a genuinely holistic view.
Lesaka's diverse book benefits from network effect on transaction one — no cold-start period while your own data seasons. The same platform sees the pattern whether it appears in the consumer bank, the merchant POS book, Cash Connect or the VAS estate.
Beautifully simple front end: no-code rule editor for the risk team, LLM-driven investigation tools in current development, and classic ML for anomaly detection and peer-group analysis.
Fraudio has an active Africa presence through partnerships with Enza and Omnea. Happy to walk through what that means in practice — and where Lesaka fits — on the deep dive.
Fraudio's multi-tenant model gives each Lesaka subsidiary its own segregated environment — tenants for consumer bank, merchant/POS, Cash Connect, Recharger and Enterprise VAS. Each business unit runs its own rules, its own KPIs, its own analyst queue. Group risk gets the roll-up.
Each tenant is data-segregated. Rules, alerts and cases stay within the tenant. No accidental cross-tenant visibility.
Mohamed's central team gets the group view — VAMP/SMMP-style ratio drift across every subsidiary, comparative KPIs, cross-subsidiary trends.
Group-mandated rules (regulatory, scheme, Lesaka appetite) can be pushed to every tenant and cannot be silently disabled by a subsidiary.
We're not asking Lesaka to rip out every tool. Encompass is working for KYC/CDD — keep it. Where Fraudio replaces the fragmented, siloed local tooling is on transaction monitoring and merchant/entity fraud, unified across every subsidiary.
Encompass stays where it is. Fraudio focuses on transaction monitoring and entity-behaviour rails rather than duplicating KYC/CDD tooling that's already scaled and working.
Orca isn't integrated into the payment rails and lacks push-authentication capability. Fraudio unifies ATM, card, merchant, wallet and account monitoring under one AI-driven platform — no more per-product tooling.
The strategic story: one Fraudio contract replaces the mesh of subsidiary point tools inherited through acquisitions. Encompass keeps doing KYC/CDD. Lesaka gets one operational language across every business unit.
The consortium AI is already trained. Every Lesaka subsidiary gets the network effect on transaction one — no cold-start period while your own data seasons across each business unit.
One contract, one console, one integration pattern replicated per subsidiary. Onboarding Bank Zero when the approval lands is a tenant provisioning task, not a new vendor selection.
Active partnerships with Enza and Omnea in Africa — happy to walk through the practical implications for Lesaka on the deep dive.
No-code rule editor. LLM-driven investigation tools. Every AI alert ships with clear reasoning — designed to be usable by the risk analyst, not just the data scientist.
Viva Wallet is a European payment institution and acquirer that deployed Fraudio to score card fraud in real time and monitor merchants for fraud and money laundering. It scaled the merchant book past half a million without scaling the fraud team — the operational shape Lesaka's diverse, multi-subsidiary book needs.
“Fraudio enables us to detect fraudulent merchants and money laundering, ensuring the safety of our clients against fraud in payments. This has been extremely helpful to our growth by providing us the ability to focus our efforts in a much more accurate manner.”
Full API reference, integration guide, data schema, webhook specs.
Centralised AI, network effect and customer results in one page.
How the AI is trained, how alerts are explained, model governance.
Confirmed 23rd, 10am — or grab an earlier slot for a pre-demo call.
Demo confirmed for the 23rd at 10:00. Mohamed's already invited the wider business teams. If a pre-demo call would help align anyone internally before the session, grab a slot below.
Book a Pre-Demo Call