https://images.ctfassets.net/ic8vz4cuikua/68hPCp16LCYCdO6dLE7KW9/ae9b77729e6b008f336250db3005412b/Intro.png?w&h&fm&fl

AI Mobile Payments Platform for FuturePay

Yellow partnered with FuturePay, a Nigerian fintech company, to build a mobile payments ecosystem for consumers, merchants, and sub-merchants.

Client:

FuturePay

Time:

January 2026 – ongoing

Type:

AI development, mobile development

Industry:

Fintech

futurepay

24/7 AI-powered assistance

We transformed customer support from conventional workflows into 24/7 in-app assistance, answering FAQs, explaining FuturePay functionality, and helping users understand transaction statuses at any time.

58% faster KYC investigation

While using Claude Sonnet we reduced suspicious KYC review time from 12 minutes to 5 minutes per submission by analyzing KYC information, highlighting potential inconsistencies, and preparing cases for human review.

100% user confirmation for AI-assisted payments

We added AI assistance to traditional payment workflows by preparing payment interactions while keeping 100% of money-moving actions subject to explicit user confirmation through an authorized mechanism.

45% less manual operations work

Admin AI reduced manual operations work by 45% by replacing fully manual case review and summarization with AI-assisted classification, case summaries, and recommended next steps before human approval.

About the project

FuturePay, a Nigerian fintech company, approached Yellow to build a mobile payments platform for consumers, merchants, and sub-merchants across Nigeria. The goal was to create a secure and scalable ecosystem that supports digital payments, merchant operations, and financial services while complying with local regulatory requirements.

Beyond the core payment infrastructure, FuturePay wanted to introduce AI-powered assistants that could improve operational workflows and help compliance teams identify potential issues more efficiently.

The client had

  • Mobile-first payments platform

  • Instant consumer and merchant payments

  • Multiple payment and funding methods

  • Merchant and sub-merchant management

  • Strong security and compliance requirements

  • AI-powered support and operations

We were responsible for

  • Mobile app development

  • Backend payment infrastructure

  • Admin portal, wallets and virtual accounts

  • Merchant and sub-merchant management

  • QR, NFC, and USSD payments

  • AI-powered support and operations

Team

  • Project manager

  • 2 mobile engineers

  • Backend engineer

  • AI engineer

  • QA engineer

AI-based features

AI-powered customer assistant

The in-app AI assistant, powered by Claude Sonnet, answers FAQs, explains FuturePay features, checks transaction statuses, provides contextual support, guides users through payment flows, and prepares payment actions for explicit user confirmation. It cannot independently execute money-moving actions.

futurepay

Admin AI agent

Claude Sonnet supports FuturePay’s internal operations by identifying failed transactions, flagging suspicious KYC submissions and potential fraud signals, summarizing operational issues, drafting support responses, and generating reports. Sensitive administrative actions still require human approval.

futurepay

Main features

Instant money transfers

Users can send money instantly to phone numbers and bank accounts, giving them multiple ways to transfer funds from the same application.

FuturePayID virtual accounts

Each customer receives a virtual account that can be used to receive and manage funds within the FuturePay ecosystem.

QR code payments

Customers can scan QR codes and complete payments directly from their FuturePay wallet.

NFC tap payments

The application supports NFC-based payment scenarios, creating a faster in-person payment experience for supported devices and merchants.

Bill payments

Scheduled payments

Multiple wallet top-up methods

Merchant/sub-merchant management

Biometric security

How Claude Sonnet is used

FuturePay uses Anthropic Claude Sonnet as an intelligence layer across customer-facing and internal workflows.

Context processing

The application provides Claude with the context required for the specific workflow, like the user's support question, relevant transaction information, KYC information, or operational event data.

AI reasoning and assistance

Claude Sonnet interprets context and provides the right response or recommendation. It supports customers with product and transaction guidance and helps admins analyze risks, summarize issues, and draft responses.

Controlled actions

Claude does not directly authorize financial transactions. When an AI-assisted interaction involves moving money, the system creates the relevant action for the user to review rather than allowing the model to execute it independently.

Tech stack

Mobile

React Native, React, TypeScript, Kotlin, Swift

AI

Anthropic Claude Sonnet

KYC & security

ML Kit, Vision Camera, Biometrics

Payments

NFC tap-to-pay, QR payments

Development challenges and solutions

Building for a regulated financial environment

Challenge: FuturePay operates in a highly regulated industry. The platform needs to account for CBN requirements, KYC/AML processes, NDPR data protection, PCI DSS requirements, and integrations with Nigeria's financial infrastructure, including NIBSS.

Solution: We designed the platform around controlled transaction flows, identity verification, secure authentication, access controls, payment confirmation, and auditable operations. Compliance considerations were incorporated into the system architecture from the beginning.

futurepay

Supporting multiple payment methods and user scenarios

Challenge: FuturePay needed to combine bank transfers, wallet payments, virtual accounts, QR transactions, NFC payments, cards, bills, scheduled payments, merchants, and sub-merchants inside one ecosystem.

Solution: We created a modular mobile and backend architecture that separates payment channels from the core account and transaction logic. This makes it easier to introduce additional payment methods and merchant capabilities as FuturePay expands.

futurepay

Payments under unreliable connectivity

Challenge: Mobile connectivity is not always reliable, and some potential users may not have access to smartphones or consistently available mobile data.

Solution: We designed the payment ecosystem to support USSD alongside smartphone-based payment flows, expanding access beyond users who can depend on an Android or iOS application and continuous internet connectivity.

futurepay

Introducing AI without giving it uncontrolled access to money

Challenge: A financial AI assistant becomes significantly more useful when it can understand transactions and help users perform financial tasks. However, allowing an LLM to authorize payments or sensitive administrative actions autonomously would introduce unacceptable risk.

Solution: We separated AI reasoning from financial authorization. Claude Sonnet can interpret requests, retrieve information, flag risks, recommend actions, and prepare workflows. But payment execution requires explicit customer confirmation through PIN or biometrics, while sensitive administrative actions require human approval.

futurepay

Results

Yellow developed a secure mobile payments platform for FuturePay, combining instant transfers, virtual accounts, QR/NFC/USSD payments, merchant management, and AI-powered support and operations. By integrating Claude Sonnet, the platform reduced KYC investigation time, cut manual operations work, and introduced AI-assisted workflows while keeping all sensitive financial actions under user or human approval.

Results
AI engineering pod$16k/monthFixed
monthly cost.