Back to past projects

Whizplan

B2B travel safety platform with AI risk reports, dual-billing, and multi-tier RLS. Sole developer.

Visit the live product

The problem

Corporate travellers visiting unfamiliar locations need real-time intelligence on what is happening between their origin and destination. The regulatory environment around travel-risk intelligence demands that no user be identifiable to anyone, including the platform itself. Off-the-shelf travel platforms log everything for monetisation, which immediately disqualifies them from this market.

Whizplan splash — "Plan trips. Brief your team. Be ready for anything." The pitch a corporate traveller actually wants: one structured plan, not seven half-finished tabs
Home screen — for travellers, for HR & managers, privacy-first. No analytics, no tracking, no third parties. The pitch corporate duty-of-care buyers cannot get anywhere else

Why custom over SaaS

There is no SaaS that simultaneously delivers real-time AI-summarised travel intelligence and complete user anonymity. A privacy-first version of the same product has to be built from scratch with anonymisation baked into the architecture, not added on later.

HR Dashboard — Global Watch. Real-time visibility of who is where, what is missing, what to do if something goes wrong — the org-side view competitor SaaS only delivers by tracking everyone individually

What we built

A cross-platform mobile app (Flutter for Android, iOS, and web via CanvasKit). Every API call routes through a Supabase edge layer that strips identifying metadata before forwarding to upstream services. Google Maps API provides route data. Gemini AI with Google Search grounding generates a brief on conditions and notable events between any two points the traveller is moving through. Push notifications alert the user to changes en route. A Stripe and RevenueCat dual-billing layer handles paid tiers without exposing user identity to the payment processors at the app level.

Trip builder with Privacy Mode toggle — the traveller drives the inputs, the edge layer strips identity before anything reaches the model
Destination search powered by Google Maps — type an airport or address and the app autocompletes, all routed through the anonymising edge layer
Drop a pin on the map for an exact location — the same Google Maps + OpenStreetMap stack, with no identifying metadata leaving the device unmasked

The hardest part

Building a system that delivers personalised AI-generated content while maintaining complete user anonymity. The model needs context to write a useful brief, but the user must remain unidentifiable end to end. Every data path from the user's phone to the model and back was designed around routing through an anonymising edge layer. Add a real-time AI service, three platform targets, and dual-billing reconciliation, and the architecture stays demanding regardless of how clean the surface looks.

Notifications inbox — Gemini AI with Google Search grounding watches the corridor between your origin and destination and surfaces critical events as they happen
Full AI-generated risk article — situation report, biggest risk, action plan. Personalised to the trip without ever exposing who the traveller is

What it shows

Whizplan is the strongest demonstration we can offer of how to build a privacy-first AI consumer app. Cross-platform from a single codebase. AI integration with search grounding. Multi-tier subscription with dual payment providers. Multi-tenant data isolation. All of it shipped and maintained by one developer.

Subscriptions — single-user Pro tier billed through Google Play and Apple App Store keeps the consumer experience friction-free
Business tier billed through Stripe with per-seat pricing and a 30-day free trial — the same app serves both consumer and corporate buyers from one codebase
Account menu — Join with invite code (team onboarding), Manage Subscription (deep-links into the app-store billing portal). Privacy-first means no email-list capture
Timezone awareness — IANA-zone trip legs with a live SAST clock preview so a traveller knows what time their arrival airport actually thinks it is
Exported trip PDF — personal info, emergency contacts, full travel-leg detail. Hand it to your assistant, your spouse, or your HR officer. The same brief the AI watches against, in a format anyone can read offline

Want one like this?

Tell us about your operation and we will follow up to schedule a discovery call.

Book a Discovery Call