Leadsbox was created to solve a real, daily frustration: tens of thousands of dollars in sales lost inside WhatsApp and Instagram threads because there was no system to track, qualify, or close them at scale.
The average Nigerian or African small business receives hundreds of buying messages across WhatsApp, Instagram DMs, and Telegram every week. These conversations are chaotic — mixed with support requests, complaints, old threads, and ghost buyers.
Sales teams spend hours manually reading through DMs to find which customer is ready to pay. Invoices are sent via screenshots. Follow-ups happen from memory. Deals get lost because someone didn't reply in time.
This isn't a productivity problem. It's a structural problem — and it requires a structural solution built for the way African businesses actually operate.
A four-stage AI pipeline that runs from the moment a message arrives to the moment a payment is confirmed.
WhatsApp, Instagram DMs, and Telegram messages flow into a unified inbox in real time via official APIs.
Each message is vectorised using LLM embeddings and indexed via HNSW. The AI classifies intent: inquiry, price negotiation, order intent, or complaint.
Hot leads are automatically surfaced with a confidence score. Sales teams see the highest-intent conversations first, not the most recent ones.
Invoices are AI-generated from conversation context. Teams can send, track payment status via Paystack, and follow up — all from a single screen.
Deliver enterprise-grade AI sales automation to every African SME that sells via social messaging — regardless of their technical capacity or team size.
Simplicity over complexity. Transparency in data practices. No lock-in. Radical focus on the SME owner who just wants to sell more and stress less.
Become the default revenue infrastructure layer for the next 10 million businesses across Africa and emerging markets who sell via conversational channels.
LeadsBox is an AI-first, cloud-native system designed for high-concurrency real-time inference at scale.
Incoming messages are converted to high-dimensional embeddings using OpenAI or Google models and indexed via HNSW (Hierarchical Navigable Small World) graphs for sub-10ms nearest-neighbour retrieval. The result: intent classification at a scale no keyword rule system can match.
A Redis pub/sub backplane powers the WebSocket layer, enabling horizontal scaling across thousands of concurrent active agent sessions. All inference is stateless, idempotent, and tenant-isolated, ensuring zero data leakage between accounts regardless of load.
Chizoba Victory is a software engineer and the Sole Director of Tribus Global Ltd (RC 9204949) — the Nigerian-registered company behind LeadsBox. With a deep background in systems engineering and distributed infrastructure, she founded LeadsBox with a singular mission: to empower African SMEs with an enterprise-grade "Revenue Engine" that uses cutting-edge AI — Large Language Models and Vector Databases — to automate the chaotic sales cycles of social commerce.
Having worked with and observed hundreds of Nigerian and African businesses, Chizoba identified that the biggest gap wasn't ambition — it was infrastructure. Small teams were beating themselves against the limitations of WhatsApp, manual invoicing, and zero lead-qualification tooling. LeadsBox is the answer she built.
The goal: deliver the same AI sales infrastructure used by Fortune 500 companies to small businesses that previously lacked the technical overhead to scale — turning WhatsApp DMs and Instagram threads into a structured, data-driven revenue pipeline that any SME can operate without an engineering team.
Built specifically for the constraints and opportunities of operating a business in Africa's digital economy.
Native Paystack integration, multi-currency support, and features designed for the realities of emerging markets — Pidgin English, informal trade, and mobile-first buying behaviour.
GDPR-compliant, Nigeria NDPR-aligned, PII redaction before LLM prompts, AES-256 encryption at rest, TLS 1.3 in transit, and isolated Redis namespaces per tenant.
Our semantic processing layer is trained to understand local code-switching, informal pricing language, and high-intent buying signals unique to African SME commerce.
We use LeadsBox ourselves every day to run sales. Every feature exists because we felt the pain of managing WhatsApp and Instagram sales manually at scale.
LeadsBox works where your customers already are — inside WhatsApp threads and Instagram DMs — so there's zero behaviour change required from your team.
Redis-backed WebSocket backplane handles thousands of concurrent real-time inference requests. Whether you process 100 or 100,000 conversations, the architecture scales.
Sub-Saharan Africa is home to over 44 million SMEs. The vast majority use WhatsApp as their primary sales and customer engagement channel. Yet every major CRM — Salesforce, HubSpot, Pipedrive — was built for Western enterprise workflows, not for a Jos fabric seller or a Lagos cosmetics brand managing 500 DMs a day.
LeadsBox fills a gap that no existing product has addressed: a vertically integrated AI sales operating system designed from the ground up for the social commerce economy that dominates African markets.
LeadsBox is a living product. Here's what's in active development.
AI agents that autonomously send follow-up messages, negotiate price ranges, and nudge cold leads back into the pipeline — all within WhatsApp, without human intervention.
Automatic transcription and intent classification of WhatsApp voice notes — one of the most common communication formats in Nigerian commerce.
Assign conversations to sub-teams (sales, support, billing) with role-based access, internal notes, and SLA tracking. Scale from solo to 50-person sales floors.
Revenue attribution per channel, per agent, and per campaign. Know exactly which Instagram post drove the most paying customers this month.
Join businesses already using LeadsBox to convert DMs into invoices, and invoices into revenue.
Built and operated by Tribus Global Ltd (RC 9204949) — proudly Nigerian, globally scalable.