Version 1.0
Status Phase 1 Pilot Launched
Product Manager Yvonne Gitata

1. Executive Summary

1.1 What We Built

An Embedded Inventory Financing Module that integrates Pezesha’s lending infrastructure directly into the Twiga Foods supply chain. The product allows Micro, Small, and Medium Enterprises (MSMEs/Kiosk Owners) to access instant, data-driven working capital to purchase inventory and restock their businesses on time.

What makes this product distinct from standard digital lending is its hybrid high-touch/high-tech model. It enables Twiga Call Centre Agents and Sales Reps to act as “human interfaces” for low-income vendors who may not own smartphones - allowing them to apply for, receive, and repay loans entirely through their existing ordering behaviour via USSD and SMS.

1.2 The Data Bridge That Makes It Work

The central technical and commercial innovation is a data bridge between Twiga and Pezesha. Pezesha previously had no visibility into the purchasing behaviour of Twiga’s vendor network. Twiga had rich order history data but no lending infrastructure. By connecting these two assets:

2. Discovery & User Research

2.1 Research Approach

Discovery involved conversations across three groups: kiosk owners (the borrowers), Twiga sales reps and telesales agents (the human interfaces), and internal Twiga and Pezesha stakeholders. Each group revealed a different dimension of the same problem.

2.2 What We Heard

Kiosk Owner Insight: The Mid-Week Cash Wall

<aside> 💡

"By Wednesday or Thursday, the money from Monday’s stock is gone. I can’t order again until Friday when customers pay me. So my shelves are empty for two days." — This pattern was confirmed directly with kiosk owners. The problem wasn’t lack of demand — it was a timing mismatch between cash inflows and restock cycles. These were viable businesses being throttled by a cash flow gap.

</aside>

Stakeholder Insight: The Invisible Data Asset

<aside> 💡

Twiga had years of vendor purchase data showing exactly which kiosks were reliable, high-frequency orderers. But this data was sitting unused from a credit-scoring perspective. Pezesha, meanwhile, was trying to lend to exactly this demographic with no behavioural data to underwrite against. The integration was an unlock for Pezesha’s entire risk model for this segment.

</aside>

Agent Insight: The Manual Attribution Problem

<aside> 💡

Telesales agents and sales reps were tracking credit-linked orders manually to attribute which loan corresponded to which stock delivery. This was error-prone and created reconciliation gaps between Twiga and Pezesha. There was no automated bridge between a confirmed order and a disbursed loan.

</aside>