Your customers make thirty commercial decisions a month. You monetise one. Flood is the AI commerce layer that turns the app they already open into the place those decisions happen — and turns the merchants on the other side into a second, higher-margin business you own.
Start the 4-week testThe premise, in two sentences. You already pay for the audience, and you monetise one thing they do with it. Everything else they do — deciding where to shop, what to buy, where to walk, who to pay — happens somewhere else, on somebody else's surface, and is worth more than the thing you charge them for.
Live in South Africa, India and Ghana, with agreements progressing in Mauritius, Panama, Puerto Rico and Türkiye. Definitions behind every figure available on request.
The argument on this page is the same for everyone. Your wedge, your first revenue line and your first metric are not. Tap your business.
95% of retail in emerging markets is still physical, and classical e-commerce does not work there — fulfilment costs exceed basket sizes. So the money is not in moving the goods. It is in moving the shopper. Low-income South African metro residents were reported spending roughly 29% of income on transport, and fares rose again in 2026. Move the sliders.
Illustrative worked example, not a study.
Today you own a customer relationship and monetise one product against it. The retailer owns the purchase, the card scheme owns the payment, and nobody owns the decision.
The base you already pay to acquire and retain. Flood adds nothing here — that is the entire point of the model.
The offer appears at the recharge confirmation, the balance check, the credit landing. Never behind a tab. AI serves it at each user's moment, not on a campaign schedule.
Hyperlocal discovery, search and saved lists — captured and structured, then turned into street-level demand forecasting. Prediction, not observation.
The attribution layer, adapted to your business type. This is the difference between a media business and a coupon app. Miss it and nobody funds a second campaign.
Wallets, gateways, cashback returned in-app, tiered loyalty. The loop closes where you want the next purchase to start.
Merchant turnover and consumer demand data, structured for credit underwriting and retail media. This is the part nobody copies quickly.
Build the measurement before the merchandising. If you cannot prove the walk-in, you have built a banner ad on your own property — no brand shifts trade budget to you, and you cannot underwrite a merchant on anything better than a bank statement. If you can prove it, you own the first offline attribution layer in your market.
A marketplace with one strong side is a directory. Flood runs both with the same intelligence layer, which is why the loop closes instead of leaking.
Near me, this week, in stock, affordable to me — ranked by what this person actually buys and can reach on foot, served at the recharge or salary-credit moment. Cashback lands back in your wallet, because points that expire elsewhere build somebody else's habit.
The merchant sees real walk-ins. The AI copilot tells them what to stock, price and message so those walk-ins convert. They adopt your rails to capture it. You now underwrite on observed behaviour, not a declared application — and a better-funded merchant converts more of your traffic. Back to step one.
Why this is hard to copy. A lender who can send its borrowers customers — and then tell them how to serve those customers — is not competing on rate. It is competing on outcome. That requires a consumer audience, a merchant book and an intelligence layer at once. You have the first. We bring the second and third.
Every competitor building for informal retail eventually ships a merchant dashboard, and almost none get used. Our merchants are time-poor and data-poor. They do not need information. They need decisions, pushed to them on WhatsApp, in their own language, before they think to ask.
Illustrative of the interaction model. Language, channel and prompts adapt by market.
Why this matters commercially. A merchant who uses the copilot daily generates the transaction, inventory and repayment data that makes them underwritable. A merchant who ignores a dashboard generates nothing. The AI is the mechanism that converts a listing into a data relationship — and that relationship becomes a credit book, a retail media product, and the moat.
"What should I stock this week?" · "Why did sales drop?" · "Can I afford to restock?" A CFO, ops manager and marketing assistant in one — on WhatsApp, USSD and in-app.
Our edge is not data volume — national retailers have more. It is granularity. Block-level prediction, auto-suggested orders, pre-positioning ahead of pension and pay days.
Instead of "do you have payslips?", the system states "this shop is stable and creditworthy" — on observed sales, repayment, foot traffic we generate, seasonality and repeat rate.
Segmentation, offer design and copy written in the merchant's voice. Photograph the shelf and stock updates — no typing, which is what makes the data consistent enough to forecast on. Voice and multilingual support expands the base to merchants every competitor wrote off.
Copilot, baseline forecasting, promotional engine, credit scoring v1. Produces the training data everything downstream needs.
Lending infrastructure, marketplace optimisation, computer-vision inventory. The platform becomes a fintech rather than a marketplace.
Embedded finance, dynamic marketplace intelligence, cross-border. Attempting Phase 2 without Phase 1 is the most common failure in this category — lending models trained on nothing.
Kaspi.kz started with payments and a wallet, then in 2014 opened a marketplace to any shopkeeper with a phone. The bank stopped being a bank and became the mall. Benchmark the ratios, not the scale.
Kaspi.kz 3Q 2024 investor presentation (9M'24) — roughly two years old and pre-Hepsiburada. Marketplace and Payments together produced 68% of net income, from zero for Marketplace in 2013.
Built Nu Shopping — in-app marketplace, cashback, 200+ partner stores — to rebalance revenue toward fees. 255 million visits in 2023 against 110m+ customers.
Live Better rewards across 30+ partners, plus an MVNO, on a mass-market banking relationship. 26% of group headline earnings from fintech in the year to February 2026.
Third-party merchant discovery layered onto an existing payments habit rather than a separate destination app. 70M+ user base.
A retailer buying the fintech layer: Flash + Shop2Shop into a R21.3bn business. ~176,000 traders, R200bn+ a year. Announced, not yet closed.
Read the Pepkor line carefully. The threat to a telco or bank here is not another telco or bank. It is the retailer — who already owns the store, the shopper and the supplier budget, and is now buying the payments and merchant rails to complete the set. In most markets that race has already started, quietly.
A standalone super app that had to win a daily habit from scratch against WhatsApp. Peaked above 35 million monthly actives, switched off after seven years, because it was a platform in search of a P&L. → Never build a destination. Embed, and give the layer a named P&L owner.
Six years, 76,000 active users. The Kenyan playbook imported into a market with bank penetration, card rails and a different informal economy. → Find the local friction first.
A tab requires a habit the app does not have. People do not navigate to a discount section; they respond to an offer when money is already on their mind. → Placement beats product.
Rates are set deal by deal, so none are stated here. Anyone who tells you all six start on day one has not run this.
Placement sold to the brand whose product sits on the merchant's shelf, priced on verified footfall rather than impressions. Paid from FMCG trade budgets that already exist and are currently unmeasurable.
Phase 1 — fastest first revenueListing tiers, promoted placement, campaign tools, subscription. Charge before you can prove the walk-in and you lose the merchant permanently.
Phase 2Take rate on marketplace GMV, payment margin, wallet float. Marketplace take rates run an order of magnitude above payment take rates.
Phase 2–3, compoundsMerchant working capital and deposits, consumer BNPL, cross-sell into an existing credit book. Kaspi reached 41,000 merchants on a new deposit product within a month.
Phase 2–3Premium copilot tiers for larger merchants; anonymised demand signal licensed to brands and suppliers once governance is complete.
Phase 2–3Zero-rated data sold to advertisers (telco); interchange and deposit retention (bank); supplier trade spend (retailer). Usually the most defensible line, because a competitor structurally cannot copy it.
Phase 2No national launch, no rebuild, no budget in phase one. We ask for a placement and a defined geography. Every phase has a gate — if a gate is missed, the correct action is to diagnose, not to scale.
One cluster. 40–60 merchants in weekly-cadence categories only. Randomised exposed and holdout cohorts. Offers at the chosen high-attention moment, nowhere else. Copilot live.
300–500 merchants across two or three clusters. First supplier-funded campaigns. First attributable case study.
Widen categories. Merchant services tiers. Your partner-specific revenue line switches on. Credit scoring running in shadow mode.
Financial services attach to the merchant base. Consumer BNPL where licensed. Demand-signal data product. Design the next market.
Most launches in this category are measured on the wrong things, which is why they survive two years and then get switched off.
Supply that is not merchandised is a database. → Merchants with a live, funded offer in the last 30 days.
Borrowed from a business model this deliberately is not. → Verified in-store redemptions, and cost per verified footfall.
Measures curiosity, not behaviour. → DAU/MAU on the layer, and the delta against your app's baseline.
The most seductive vanity metric in the market. A copilot nobody opens has no value at any level of model quality. → Copilot weekly-active merchants, and the share of recommendations acted on.
Verified in-store transactions per active user per month, and the share of merchants who act on an AI recommendation weekly. The first proves the layer changed real-world behaviour. The second proves the AI is an operating system, not a feature.
The app is yours, the brand is yours, the data governance is yours. Flood does not build a consumer brand that competes with partners.
Where the basket cannot carry a delivery cost, the store is the fulfilment centre. Warehouses are how the classical model dies here.
SDK into the app you already have, at a placement you already control. Front end untouched.
Category and term-limited exclusivity in a defined channel, yes — you deserve protection while the model proves out. Open-ended exclusivity is a free option on a whole market, and it removes the pressure that makes the layer improve.
Thirty minutes is enough to scope a cluster and agree how we will measure it. Bring whoever carries the revenue number — we will bring the numbers.
superapp@flood.finance Flood on LinkedInHow long does money stay with your customer before it leaves your rails? For a bank: hours between the credit landing and the cash withdrawal. For a telco: how much of your base has usable data at any time. The answer determines the design, the sequencing, and sometimes whether this is the right first market at all — and you can answer it from your own data in an afternoon.
Third-party figures: Kaspi.kz 3Q 2024 investor presentation; Nubank company release (2023); Capitec annual results to February 2026; GCash/GLife materials; Moneyweb on Pepkor–Flash–Shop2Shop (announced July 2026, not closed); TechCabal on the Ayoba shutdown (March 2026). Transport and basket figures are an illustrative worked example based on reported South African commuter cost data, not a study. Flood figures are first-party; definitions on request.
