There is an issue with the Youtube video embedding, so we attached the link as well for easy reach.

Demo Video Link

https://www.youtube.com/watch?v=Ah32LRLnh5s

About the Project

The Problem

Across Nigeria and India, hundreds of millions of people participate in the economy without having reliable access to conventional banking and business-management tools. They run shops, trade goods, provide services, and manage significant amounts of money, but many operate outside the systems designed for formal businesses.

For a small retailer, running a business can mean purchasing inventory from several wholesalers, selling dozens or hundreds of products every day, extending informal credit to customers, keeping track of cash manually, and deciding what to purchase next, often without a POS system, accounting software, standardized inventory records, or a dedicated business bank account. The result is a fundamental information problem. A merchant may know that their shop is busy, but not know their actual profit. They may have money in the business, but not know how much capital is tied up in inventory. They may sell hundreds of products, but have no reliable record of what is running low. They may need financing, but lack the financial records required by conventional financial institutions.

The problem is not necessarily that these businesses do not want technology.

The problem is that much of today's technology expects the user to adapt to the software.

We asked:

What if the software adapted to the way the merchant already works?

Our Insight: The Phone Is the New Interface

In many underserved communities, the phone is already the most accessible piece of technology.

People may not have a laptop, POS terminal, accounting system, or sophisticated business software, but they have a phone. They already use it to communicate, make transactions, coordinate suppliers, and run parts of their businesses.

That led us to a simple idea: Instead of asking people to learn another complicated application, let them run their business by talking to an AI agent through the device they already understand. We are building an agentic AI business and financial assistant that uses the phone as the interface.

A merchant could receive a call or interact through their phone in a language they are comfortable with. The AI can ask simple questions such as:

What did you sell today?
What inventory did you purchase?
How much did you spend?
What products are running low?
How much money is currently in the business?
What do you need to purchase tomorrow?

The agent can turn those conversations into structured business information and progressively build a financial picture of the business.

Over time, this can support inventory management, sales tracking, accounting, cash-flow management, procurement, financial planning, bank-account creation and management, credit access, and other financial services.

The long-term vision is not simply an AI chatbot.

It is an agent that can actually help operate the business.

Why Agentic AI?

Traditional software waits for the user to navigate menus, enter data, understand dashboards, and perform tasks. We believe agentic AI creates a fundamentally different interface. The merchant can simply state what happened, and the agent can determine what needs to be recorded, what needs to be calculated, and what action should happen next.

For example:

"I bought 20 bags of rice for ₦800,000 today and sold five."

Instead of requiring the merchant to manually update an inventory database and accounting system, an agent can interpret the conversation, update the appropriate records, calculate the implications for inventory and cash flow, and potentially remind the merchant when it is time to reorder.

This changes the relationship between people and software from:

Learn the software → enter information → perform the task

to:

Tell the agent what happened → the agent understands → the agent performs the work.

What We Built

We built a voice-first agentic AI system designed around this interaction model. The system is designed to:

Communicate through the phone rather than requiring sophisticated hardware or software.
Understand natural conversation, reducing the need for forms and complicated interfaces.
Work across local languages and communication styles, including languages and dialects commonly used by merchants.
Capture business activity such as sales, purchases, inventory, expenses, and cash movement.
Turn unstructured conversations into structured business records.
Provide business insights and recommendations based on the information collected.
Create a foundation for financial services, including banking, financial management, credit, and other products.

The architecture combines conversational AI, agentic workflows, structured data processing, and financial/business logic. The goal is to make the complexity invisible to the merchant. They should not need to understand databases, accounting software, APIs, or AI. They should simply be able to talk.

What Inspired Us

We were inspired by seeing how much economic activity happens outside conventional digital systems. A business can be real, profitable, and economically important while still having very little digital financial history. That creates a paradox: The people who may benefit most from better financial tools are often the people least able to use the tools currently available. We believe AI creates an opportunity to reverse that. Rather than forcing underserved populations to become sophisticated software users before they can participate in the digital economy, we can build technology around the interfaces they already use. The phone provides the bridge.

What We Learned

One of our biggest lessons is that accessibility is not only about making technology affordable. It is also about making technology understandable. Adding more features does not necessarily make a product more useful. For many small-business owners, the most powerful interface may be a conversation. We also learned that trust is fundamental. When technology interacts with someone's money, business records, or financial information, users need to understand what the system is doing and why they should trust it. This means our product has to be designed around transparency, simplicity, reliability, and user control, rather than treating AI as a black box.

Challenges We Faced

1. Building for informal businesses

Informal businesses do not always follow standardized accounting or inventory practices. Different merchants describe the same business activity in completely different ways. We therefore had to think beyond traditional forms and schemas and design the system to interpret natural language and convert it into structured information.

2. Language and voice

A voice-first product has to work in the real world, not just in a controlled English-language demonstration. Accents, local expressions, code-switching, background noise, inconsistent connectivity, and local languages all create challenges for speech recognition and AI reasoning. This is particularly important in markets such as Nigeria and India, where language diversity is enormous.

3. Trust and financial information

Financial data is highly sensitive. We have to think carefully about authentication, privacy, security, consent, and how much autonomy an AI agent should have when dealing with financial actions. We believe agents should not simply be given unrestricted access to people's money. High-impact financial actions need appropriate safeguards, user authorization, and clear boundaries.

4. Designing for simplicity

One of the hardest product challenges is deciding what not to show. A conventional financial application may expose users to dashboards, charts, categories, accounting terminology, and dozens of settings. Our goal is to hide that complexity behind a simple interaction. The complexity should live inside the system, not with the user.

Where This Can Go

We see this as a much larger opportunity than inventory management. Once an AI agent understands how a business operates, it can become the interface through which that business accesses an entire financial ecosystem. The progression could look like:

Business records → Accounting → Procurement → Banking → Credit → Payments → Financial services

Instead of requiring an underserved entrepreneur to discover, understand, and separately operate each financial product, an intelligent agent could help them navigate the ecosystem through a single conversational interface.

Our vision is to make the phone more than a communication device. We want to make it an intelligent business and financial interface for people who have historically been underserved by conventional digital systems. We are starting with Nigeria and India because of the enormous scale of their informal economies, but the underlying problem exists across emerging markets around the world. Ultimately, we believe AI can make formal financial and business infrastructure accessible without requiring everyone to first learn how to use formal software.