When Teddy Ogallo began building conversational banking tools in 2019, there were sporadic conversations, at least in the media, surrounding artificial intelligence (AI) chatbots or no-code platforms.
At the time, there were no large language models (LLMs) in the public domain, and few banks believed customers would transact through chat apps. But the need was obvious.
As banking teams struggled to reach customers during lockdowns, Nairobi-based WayaWaya introduced the region’s first conversational AI solution. The platform enabled users to make financial requests and complete transactions across WhatsApp, the web, and SMS.
Today, as global players like OpenAI and Google rush to release no-code “agent builders,” WayaWaya’s head start is interesting to look at. The company has quietly developed a full-stack solution that combines conversational AI, payments, and a drag-and-drop workflow builder, serving banks, telcos, and small businesses across Africa, Eastern Europe, and the Middle East.
The core architecture
WayaWaya runs on two main backend layers. The first is a payments platform, which serves as the backbone of the system, having been built over several years and directly integrated with banks and payment processors.
It connects to Mastercard’s core banking sandbox, a testing environment used by Mastercard’s principal member banks, as well as to Interswitch’s merchant and payment APIs across Africa.
In practice, this means a bank in Nigeria or a fintech in Kenya can launch a new chat-based banking service simply by updating their API credentials, testing the workflow in Mastercard’s sandbox, and going live, all without building or modifying their own backend systems.
The second one is a conversational AI platform, which Ogallo claims was built before LLMs like ChatGPT were available. WayaWaya’s conversational engine was designed to translate natural-language inputs into financial requests and route them to the payments layer.
When early enterprise clients wanted to use the same interface for customer support, product sales, and utility payments, Ogallo’s team realised that every implementation required weeks of custom development.
That bottleneck led to the creation of a third layer, called the No-Code frontend platform, a web-based builder that enables teams to design their own conversational journeys.
How the no-code layer looks like
The drag-and-drop interface is powered by a Next.js front end, backed by a Python API layer. Lightweight versions exist on Android and iOS (built in Java and Swift). Through it, enterprises can define conversational flows, such as loan applications, bill payments, or marketing campaigns, using modular components that connect directly to WayaWaya’s backend APIs.
Each business can upload its own data and frequently asked questions (FAQs), which the AI engine uses to train contextual models. The system then tracks intent and confidence in real time. If the model’s confidence falls below a set threshold, the query is escalated to a fine-tuned LLM hosted on WayaWaya’s own servers. That model performs classification and prediction tasks to ensure precise, context-aware responses.
This approach eliminates the need for third-party AI providers, allowing enterprises to maintain more control over data security and performance.
Orchestration and data flow
WayaWaya’s orchestration logic ties its conversational, payments, and no-code layers into one continuous system. Each user interaction begins as a session, identified by a unique key or messenger ID. The system maintains a unified customer profile across all channels, including WhatsApp, web, and SMS, and tracks conversation state in real-time.
If a user switches devices or channels during a conversation, the context is preserved. A customer can initiate a bill payment on WhatsApp and complete it using a web widget without losing progress. Every request, from checking a bank balance to buying airtime, is classified and routed through the same backend pipeline.
This architecture allows the same logic to power a marketing chatbot for an e-commerce brand, a loan application for a microfinance firm, or a bill payment for a utility provider.
How security is executed
WayaWaya embeds PIN, password, face, and identity verification within its chat workflows. The company’s first deployment with HFC Bank used external pop-ups for SIM-swap validation while keeping mobile-banking PIN entry outside the chat interface.
More recently, the team introduced secure in-chat forms that support encrypted document and PIN entry directly within WhatsApp. These forms leave no trace in the chat history and are decrypted only on the enterprise client’s servers using private keys. For less secure messaging channels, the system automatically redirects verification to a secure URL.
The result is a hybrid model that meets banks’ compliance standards while keeping the experience as seamless as ordinary chat.
Integration with payments
The Mastercard and Interswitch integrations show how WayaWaya’s stack is built for scalability rather than one-off custom projects.
For Mastercard principal member banks, the sandbox integration means faster testing and deployment across multiple regions.
For Interswitch merchants, the API integration provides instant access to one of Africa’s broadest payment networks, which links cards, wallets, and bank accounts in real-time.
Because the system is API-based, the same conversational journey that powers a payment in Nigeria can work in Kenya or Ghana with minimal configuration. This multi-market interoperability is what makes WayaWaya valuable to banks that operate across borders but lack the engineering bandwidth to build new digital interfaces for every country.
Proprietary tech
While competitors like Chpter, also Nairobi-based, focus on single-channel sales automation, WayaWaya’s edge lies in its proprietary backend, much of it built before AI toolkits became widely accessible.
Ogallo says that the company gained an early understanding of how conversational intent maps to financial actions by developing its own models, an orchestration layer, and a builder interface. This experience is difficult to replicate quickly, even with access to modern APIs or LLMs.
Ogallo says the growing global awareness around no-code and AI builders is helping, not hurting, WayaWaya’s business. “The timing works in our favour,” he says. “It validates what we’ve been building since 2019 and lets us focus on helping more enterprises and developers create and scale their own conversational AI journeys.”
Why this matters
As generative AI moves into mainstream enterprise software, most new products aim to simplify complexity. WayaWaya’s story shows what that process looks like when built from the ground up for emerging markets, where messaging apps are the dominant interface, not a secondary channel.
Its blend of banking-grade integrations, secure chat verification, and no-code configurability reflects how African engineers are designing solutions that work within existing infrastructure rather than around it.
“With our latest no-code builder, we’re now focused on using that momentum to help more enterprises and developers easily create and scale their own conversational AI journeys across different markets,” Ogallo said.
from TechCabal https://ift.tt/hQoUVf2
via IFTTT
I am the CEO of 9JAGISTLAND.My name is Adedeji Timothy and my co-owner is Oyekola Damilare.This blog is born out of our passion to educating to people in all aspect of internet life.We are enthusiastic about writing posts and articles online.
Write your views on this post and share it. ConversionConversion EmoticonEmoticon