# TradePulse AI

- **Event:** [Meta Global AI Developer Hackathon](https://cerebralvalley.ai/e/meta-ai-developer-hackathon)
- **When:** Oct 26, 12:00 PM – Nov 4, 9:00 PM (EST)
- **Where:** Online
- **Team:** [Desmond Adedosu](https://cerebralvalley.ai/u/Desmoade)
- **GitHub:** https://github.com/adekunlewriteswell-code/tradepulse-meta-hackathon-2026
- **Demo video:** https://www.youtube.com/watch?v=bwaBhbh3K2I
- **Gallery:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery
- **Page:** https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3461

TradePulse AI: Listen. Understand. Activate.

In FMCG distribution across emerging markets, millions of small retailers form the backbone of product availability. Yet many outlets become inactive, and conventional sales-force automation systems can identify which retailers stopped purchasing without explaining why.

TradePulse AI is an evidence-driven retail activation prototype that explores this missing intelligence.

The solution begins with a simple numbered-menu interaction, demonstrated through a browser-based USSD simulator. Retailers can identify barriers such as limited working capital, product availability, insufficient shelf space or unreliable deliveries.

Their responses are preserved as traceable observations. Meta AI helps interpret those signals and generate evidence-linked recommendations, explicitly identifying uncertainty and risks.

A route manager then reviews the recommendations, decides whether to plan an intervention and can create a bounded incentive mission. The system separates AI advice from human authorization.

TradePulse also includes simulated purchasing, retailer commitments, reward eligibility checks and a financial intelligence layer that calculates revenue, costs, incentives and contribution margins using exact monetary arithmetic.

The prototype demonstrates the complete journey from understanding an inactive retailer to evaluating a financially controlled activation experiment.

Built with Python, FastAPI, SQLite and Meta Model API integration, it includes validation, audit controls and safeguards against duplicate transactions.

The current demonstration uses synthetic retailers and simulated transactions. It does not claim proven sales growth, live telecom USSD connectivity or real reward payments.

TradePulse AI explores a practical future for African route-to-market intelligence: understand the barrier before prescribing the incentive.

## More from Meta Global AI Developer Hackathon

- [Sentinel AI](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3458)
- [Taskforce 141](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3459)
- [Mihiretu (Didymos) Likassa](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3460)
- [Zero Echelon - Uplink](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3462)
- [Songjie Huang](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3463)
- [Kurta Squad](https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3464)

---

Markdown version of https://cerebralvalley.ai/e/meta-ai-developer-hackathon/hackathon/gallery/3461. Site index for agents: https://cerebralvalley.ai/llms.txt · full text: https://cerebralvalley.ai/llms-full.txt