NewsMacroAI Agents Raise New Risks for Banks and Fintechs in Nigeria

AI Agents Raise New Risks for Banks and Fintechs in Nigeria

Author: TechNext24·

Key Takeaways

  • •Federal Reserve Governor Christopher Waller said in September 2026 that authentication, fraud and liability are the central challenges when AI agents shop and pay on a person's behalf.
  • •According to the Central Bank of Nigeria's 2025 fintech report, real-time payment channels handled almost 11 billion transactions in 2024, double the five billion recorded in 2022.
  • •Fraud losses across Nigerian banking and payment channels dropped to ₦25.85 billion in 2025 from ₦52.26 billion in 2024, according to the CBN's Nigeria Payments System Vision 2028.
  • •Visa's 2026 Stay Secure study found a trust gap among Nigerian consumers, with 88% having used AI for shopping but only 34% trusting AI to handle payment checkout.
  • •To manage agent-driven payment risks, Nigerian financial institutions must authenticate automated systems, limit their spending power, control permitted transaction types and make permission revocation easy for customers.
AI Agents Raise New Risks for Banks and Fintechs in Nigeria

An AI assistant that books a flight, reserves a hotel and completes the payment from a customer's bank account — all without the customer opening a website, comparing prices or entering card details — captures both the promise and the peril of a new class of software now drawing scrutiny across Nigeria's financial sector. Unlike conventional chatbots, which only answer questions, these autonomous agents can execute tasks, make decisions in real time and connect to multiple online services with minimal human involvement.

For banks and payment providers, that capability introduces an uncomfortable problem: a transaction can look entirely legitimate, originating from a genuine customer's account with credentials that pass every security check, even though the account owner never intended to make it.

From Decision Support to Autonomous Action

The issue has become a major topic for the global payments industry, and Nigerian banks and fintechs are confronting it directly. At Nigeria Fintech Week 2026, it featured prominently on the agenda. During a panel discussion titled "Artificial Intelligence at Work", executives from VFD Technologies, CBI Technologies and Advansio examined how fintech companies are deploying the technology for fraud detection, credit decisions and customer service, while cautioning against the pitfalls of acting on inaccurate data and accepting machine-generated answers without verification.

A key trend highlighted at the event is the shift from software that merely supports decision-making at financial institutions toward software that begins to carry out transactions on its own.

The distinction between traditional systems and AI agents matters. A conventional fraud-detection tool might flag a suspicious transaction for a bank employee to review. An AI agent, by contrast, can be set up to pay a supplier every month, buy groceries within a specified budget, or find the cheapest way to send money abroad. The customer issues a task, and the agent handles the details until it is done.

An International Monetary Fund (IMF) note published in April 2026 concluded that this could change how payments are made: rather than a person approving every transaction, software could follow instructions and interact with payment services independently.

Federal Reserve Governor Christopher Waller addressed the subject in September 2026, outlining two models — one in which a person shops and pays themselves, and another in which an agent shops and pays on the person's behalf. The second model poses greater risks.

Consider a customer who allows an agent to use a virtual card to buy groceries for under ₦50,000. If the agent misunderstands the instruction and buys something more expensive, or follows a bad instruction entirely, the bank may see a valid transaction without realizing the customer never approved it.

Waller identified authentication, fraud and liability as the key challenges. Banks will need to establish not just who owns an account, but whether a piece of software has permission to make a particular payment. The central question is whether the agent is acting in line with the customer's wishes.

Nigeria's Payments Scale

The debate is especially relevant in Nigeria, where digital payments are widespread. According to the Central Bank of Nigeria's (CBN) 2025 fintech report, the country processed nearly 11 billion transactions through real-time payment channels in 2024, up from five billion in 2022, with more than a quarter of electronic transactions conducted through those channels. The same report showed that 87.5% of fintech companies use the technology for fraud detection, 62.5% for customer service, and 37.5% for credit scoring and risk assessment.

The numbers confirm that the technology is already embedded in Nigerian finance, but allowing an agent to handle payments requires new levels of control.

Fraud exposure remains substantial. The CBN's Nigeria Payments System Vision 2028 puts fraud losses across banking and payment channels at ₦25.85 billion in 2025, down from ₦52.26 billion in 2024. That is a significant improvement, yet billions of naira are still being lost.

Security pressure is also intensifying. Lendsqr CEO Adedeji Olowe recently warned that many banks and fintech companies in Nigeria are being targeted by cyberattacks. In an advisory published on November 6, he described various attempts to steal information and break into systems — including cases in which criminals used accounts containing only zeros to receive stolen money and passed them off as legitimate. These attacks expose vulnerabilities that could worsen if technology is allowed to directly handle money transfers.

Security Breaches and Consumer Trust

The global cryptocurrency sector has recorded several costly security breaches this year. In a major incident on October 29, attackers compromised about $6.34 billion in digital assets from Binance, which managed to recover around $2.85 billion. Other episodes, including the KelpDAO exploit and the Bitget breach, also resulted in significant losses.

The risks extend beyond hackers seizing control. An automated payment system can also be tricked into making incorrect payments through hidden instructions embedded in emails or on websites. There is likewise the unresolved question of who bears responsibility when an unauthorized payment occurs: the customer, or the institution that created the system?

Consumer trust lags well behind adoption. According to Visa's 2026 Stay Secure study, while 88% of Nigerian consumers have used AI for shopping, only 34% trust AI to handle payment checkout. That gap suggests people may be comfortable using AI for simple tasks while remaining wary of letting it manage their money. That wariness weighs heavily on the agent-driven model Waller described, since an agent that shops and pays on a customer's behalf demands precisely the level of trust most consumers have not yet extended.

What Institutions Must Do

For Nigerian financial institutions, adjusting to these changes will involve more than implementing fraud-detection systems. They will need to verify the identity of automated systems, limit their spending capabilities, control the types of transactions they can perform, and allow customers to revoke permissions easily. Those safeguards line up with the challenges Waller named: verifying automated systems speaks to authentication, spending limits and transaction controls curb fraud, and clear rules on unauthorized payments settle the liability question.

The CBN emphasizes that adopting the technology does not replace a financial institution's responsibility to maintain strong controls and oversight. Automated systems may make payments faster and easier, but it is essential that customers do not find out too late when something goes wrong with their money. Nigeria has worked hard to build an efficient payments system; the next step is ensuring that when technology is put in charge, it operates within clear boundaries.

Source: TechNext24 — AI agents raising new risks for banks and fintechs in Nigeria