Rwanda Partnership Brief

Runs security for a residential estate or an office building. Guards work from an Android app: they log every visitor at the gate, scan QR checkpoints as they walk their patrol route, raise a panic alarm, and file incident reports with photos. A supervisor watches a live control room screen across every site. Residents and office tenants get their own portal for visitor passes, and a door buzzer that reaches them on WhatsApp or SMS so they can let a visitor in without being at the gate.

Two buyers, two pitches. Private security firms buy it as an operating cost and white-label it to the estates they guard — no board approval needed, and it gives them recurring revenue. Property managers and developers of gated estates and apartment blocks buy it outright. Office towers and serviced-office operators are the third group; the system changes its own wording from estate/resident/unit to premises/occupant/suite depending on the site.

Most competitors do community administration with visitor passes bolted on. This does guard operations properly: proof a guard actually walked the route, measured panic response times, a real investigation workflow, and a monthly report the guarding firm can hand its client. The apps work offline, and the whole thing can be hosted inside Rwanda rather than on a server in America.

A client photographs a real room. The platform repaints the walls, swaps the flooring or the furniture for products from a real catalogue, or restyles the entire space from a sentence of description — photorealistically, in seconds. It then turns that picture into a costed list of everything in it, so the client sees the room and the price together. It also produces 360° panoramas and short AI video walkthroughs.

Interior design studios and architecture practices, who use it to win the client at the first meeting instead of after a week of renderings. Furniture and finishes retailers, who use it to let a customer see the sofa in their own sitting room. And property developers selling off-plan apartments, who need buyers to picture a unit that does not exist yet.

Generic AI image tools make a pretty picture that cannot be bought. This one is tied to a product catalogue, so the output is a quote. It already runs in English and French, which matters in Rwanda, and the AI engine is swappable by configuration — when a better or cheaper model appears, we move to it without rewriting the product.

Lets any merchant run their own gift-card programme. They design the voucher themselves in a drag-and-drop editor — layouts, their logo, background images, QR placement — then sell it through a public shop page and redeem it at the counter by scanning. Each merchant connects their own payment account, so the money goes straight to them and never passes through us.

Supermarkets and retail chains, restaurants, salons and spas, hotels, and bookshops — anywhere a customer might want to buy something for somebody else. Corporates are the quieter opportunity: staff rewards and end-of-year gifts are currently done in cash or printed paper vouchers that are easy to forge.

Gift cards barely exist as a product in the region, and where they do they are printed card stock with no way to check whether one has already been spent. This is digital, QR-verified, and sold from the merchant’s own page with their own branding. Because each merchant uses their own payment gateway, there is no float sitting with a third party — which is usually the first objection a finance director raises.

Infrastructure for anyone who sells software. It issues licence keys, activates them, and checks over an API whether a given installation is genuine. Around that: a product catalogue with versioned downloads and documentation, a customer portal created automatically when someone buys, a discount and voucher engine, and a support desk.

Honestly, this is the narrowest of the six locally. The buyer is a software company that sells a product rather than consulting hours — accounting packages, point-of-sale systems, school management systems. There are fewer of those in Kigali than there are estates or shops. Worth knowing about, not worth leading with.

The alternative is a spreadsheet of licence keys and an honour system. For a local software vendor losing revenue to copied installations, the activation API pays for itself quickly.

Generates captions, images and short video for a brand, plans them into a campaign calendar, then publishes automatically to Facebook, Instagram, X, TikTok and LinkedIn. Team accounts with permissions, and a credit system so usage is metered.

Marketing and advertising agencies managing several client accounts at once, and medium-sized businesses with one overworked person running all their social media.

It puts generation and scheduling in one place, metered, with the AI provider swappable by configuration. I should be straight with you about this one: unlike the others, I did not write it from scratch. I took over an existing commercial codebase, rebranded it, repaired it and extended it. It works, and I know it well enough to support it — but I am checking the licence position before either of us sells it to anyone, and I will confirm before we include it.

Books six strangers who share an interest onto one table for sixty minutes. Interest-matched seating, a verification step before a seat is confirmed, and the conversation is deleted when the table closes. It is explicitly not a dating product and is designed against becoming one.

Nobody, in the sense the others are bought — and I would rather say that than pretend. This is a consumer product, not business software. If there is an angle in Rwanda it is a restaurant group or a hotel filling quiet weeknights, or an events company running a members’ club. Treat it as something I can build, not something to go selling.

It shows what I can do with real-time systems — live messaging, scheduling across time zones, a companion Android app. If a Rwandan client ever needs something that works like a live marketplace or a booking floor, this is the proof I can build it.