Edge intelligence.
Commerce in the hand.
Twenty-two verticals, one device.
ElaraRise is a platform, not a product. A single architecture serves Vision AI at the site, WhatsApp-first commerce for customers, and a reserved expansion lane for every vertical our operators will ask for next.
Vision AI
Cameras become intelligence. Intelligence becomes action.
Every Vision AI deployment consists of an NVIDIA Jetson Orin Nano edge device installed at the customer site, connected to existing IP camera infrastructure. Models trained on Azure AI Foundry run locally on the device, producing alerts, grading events, counts, and incident records in under fifty milliseconds — without sending raw video to the cloud.
ElaraRise Vision AI analyzes objects, inventory, and site operations. We do not use facial recognition, biometric identification, or automated profiling of individuals.
When an event of interest is detected, the device publishes a structured payload to Azure IoT Hub. Our cloud tier enriches the event, notifies the correct operator via the Owner App, and writes to an immutable audit trail. Sensitive frames are blurred for operational review; unblurred evidence is sealed in a separate container, released only under authorised investigation workflows.
- Edge device NVIDIA Jetson Orin Nano, 8GB, installed per site
- Inference latency Approx. 50ms on-device, using ONNX Runtime
- Camera compatibility Standard Hikvision and Dahua IP cameras via RTSP / GigE
- Cloud backbone Azure IoT Hub, Functions, SQL, Service Bus, Blob Storage
- Alert delivery Owner Mobile App (iOS & Android), WhatsApp Business API
- Privacy design Dual-container evidence isolation, encrypted at rest, retention controls
- Provisioning Zero-touch, plug-and-play via Azure IoT Hub DPS and X.509 certificates
WhatsApp Commerce
Ordering where the customer already is.
In the markets ElaraRise serves, WhatsApp is not an additional channel — it is the primary channel. Our commerce layer embeds the entire ordering and receipt flow inside the conversation the customer already has with the operator. No new application to download. No sign-up friction. No change in customer behaviour.
All customer orders, receipts, and customer support are handled via the official WhatsApp Business Platform. ElaraRise does not process payments or store card data on this website.
Orders are captured through natural message exchange and structured catalogue responses, confirmed via M-Pesa or TIGO PESA through our Pesapal integration, and returned as a fiscal-ready receipt. The operator’s Owner App reconciles every transaction against site activity captured by the Vision AI tier — closing the loop between what was ordered, what was prepared, and what was collected.
- Channel WhatsApp Business API · Meta Cloud API
- Subscription payments Pesapal — single integration covering M-Pesa and TIGO PESA
- Tanzania collections DSM IOT AND ICT COMPANY LIMITED. settles weekly to India
- Receipt architecture TRA VFD fiscal receipt columns reserved for Tanzania regulatory activation
- Reconciliation Bi-directional with Vision AI event log — orders cross-check against site activity
Vertical Intelligence
A foundation wide enough for every industry we will serve.
ElaraRise is architected to host twenty-two vertical applications on the same platform — each one inheriting a shared foundation of identity, catalogue, transaction, and credit-ledger primitives, and adding only the intelligence that is specific to its operating domain.
Vertical-specific AI outputs — freshness grades for seafood, volumetric measurements for agriculture, spectral analysis for mining, attendance and zone integrity for workforce operations — are captured in a single, schema-flexible output layer. Adding a new vertical is a software configuration; never a platform rebuild.
Verticals are activated progressively. Each activation follows a structured commercial sequence and never compromises the shared foundation.
Four layers. Clear boundaries. Open at every seam.
We structure the platform as four explicit layers. Each has a single responsibility, and each is swappable without touching the others.
- L1 Edge — Jetson, ONNX, camera pipeline Active
- L2 Cloud Infrastructure — Azure IoT, Functions, SQL Active
- L3 Application Framework — ABP Commercial, .NET 8 Active
- L4 AI & Analytics — model-abstracted, interface-first Active
Operator data is operator data. We protect it accordingly.
Edge-first privacy.
The vast majority of video data never leaves the site. Only structured events, and — when required for evidence — specific frames are transmitted to the cloud. Raw continuous video streams are not part of our architecture.
Encryption everywhere.
All data in transit is encrypted with TLS. All data at rest is encrypted using managed keys stored in Azure Key Vault. Every credential in the platform is secret-managed — no secrets appear in source code or configuration files.
Audit-grade AI.
Every AI-generated output is logged to an insert-only audit table — model, version, prompt template, tokens consumed, duration. Every incident report has a permanent, immutable provenance record. We retain this for seven years.
Tenant isolation by design.
The platform is built on a multi-tenant foundation with row-level isolation enforced at the application layer. One operator cannot see another operator’s data under any circumstance.
Regulatory readiness.
Architecture is designed for Tanzania Revenue Authority (TRA) VFD fiscal receipt integration and for India’s Digital Personal Data Protection Act 2023. Both integrations are architecturally provisioned and activated per-tenant as operational readiness demands.