Debt-shaming is the risk that defines Nigerian consumer lending compliance, and AI agents help by reviewing every collection message, auditing loan apps' data access, reconciling loan costs and testing recovery mandates. Compliance keeps every finding and regulator contact. CBN, the federal consumer protection regulator and the Nigeria Data Protection Act 2023 set the boundaries.
What does consumer lending compliance look like in Nigeria?
Collections conduct dominates. Digital lenders grew fast, and some used aggressive recovery methods: messaging a borrower's contacts, publishing defaulters' details, threatening calls. Regulators have treated this as both a consumer protection and a data protection problem. Compliance at any reputable lender now spends real time proving its own collections, and those of any outsourced agency, stay within the rules.
Loan apps raise data questions. What an app accesses on the borrower's phone, what the borrower agreed to, and what the lender does with that data must line up. Compliance checks permissions, consent records and data retention.
Banks and microfinance banks add their own tools. Salary-backed loans and recovery mandates that let a lender recover from a borrower's accounts at other banks are powerful, and compliance checks that they are used only as agreed and only after proper notice. Credit bureau reporting, cost disclosure and complaints complete the week.
Where can agents help?
- Collections message review (depth 1). The agent reviews every outbound collection message, call transcript and messaging-app chat, in English, Pidgin, Yoruba, Hausa or Igbo, and flags threats, contact outside allowed hours, and any contact with someone other than the borrower or their guarantor.
- App data access audit (depth 1). The agent compares the permissions the app requests with the borrower's consent records and the lender's data policy, and flags any access to contacts or media.
- Cost reconciliation (depth 1). The agent compares the interest, fees and total repayable shown at application with what the loan system charged, and lists differences.
- Recovery mandate testing (depth 2). For loans where the lender used an account-wide recovery mandate, the agent checks the mandate, the default, the notice and the amount recovered, and drafts a finding where any step is missing.
Collections review comes first because harm happens there and spreads fast. A single debt-shaming message to a borrower's employer can trigger a complaint, a regulator's attention and lasting reputational damage.
What must stay with people?
Decisions on remediating borrowers, sanctioning a collections agency, or reporting an incident to a regulator stay with named compliance staff. These involve judgement, money and relationships.
Vulnerable borrowers need human attention. Nigeria has no direct counterpart to the UK's detailed rules on vulnerability, but the idea carries: a borrower who has lost a job or faces illness needs a person, not a script. Our piece on FCA Consumer Duty, vulnerable customers and AI agents sets out a method Nigerian lenders can borrow.
Use of recovery mandates stays with authorised staff. Taking money from a borrower's other accounts is hard to reverse and has serious effects. The compliance agent tests it; it never triggers it. That is the principle behind AI compliance agents as second line, never the last word.
Which rules apply in Nigeria?
The Central Bank of Nigeria regulates banks, microfinance banks and finance companies, including their consumer protection obligations, disclosure of loan costs and use of recovery mechanisms. The federal consumer protection regulator registers and oversees digital lenders that are not CBN-licensed and has set rules on their conduct, including collections. Lenders report to licensed credit bureaus.
The Nigeria Data Protection Act 2023, enforced by the Nigeria Data Protection Commission, governs borrowers' personal data: lawful basis, purpose limitation, data minimisation and security. Accessing a borrower's contacts to pressure them, or disclosing their debt to others, conflicts with these principles.
There is no AI-specific law. The UK debate on whether the FCA Consumer Duty applies to AI agents is a useful reference: the answer there is that duties attach to the firm, whatever tool it uses, and the same logic holds under Nigerian law.
This is an explanation of the rules, not legal advice; confirm how they apply to you with counsel.
How should a lender start in the first 90 days?
Days 1 to 30: run collections message review at depth 1 over the last month of messages and calls, including outsourced agencies. Share agency-level results with collections leadership.
Days 31 to 60: audit app data access and consent records. Fix anything that accesses contacts or media before going further.
Days 61 to 90: add cost reconciliation and, for banks, recovery mandate testing at depth 2, with each finding held for compliance. Our piece on Consumer Duty and AI agent communications with customers offers a useful test for clarity of borrower messages.
Where Pryme fits
Pryme's minimum chain includes platform guardrails that no customer setting can lift. For lending, a guardrail can make contact with a borrower's phone contacts a withheld action: it exists technically, but is refused for every agent at every depth. Each review capability is certified by Attest on the lender's own messages, in each language.
Questions people ask
Can an AI agent send collection reminders?
That is an operations use rather than compliance, and it can be done at depth 3 within strict limits on timing, wording and recipients. Every message should be reviewable by compliance.
Does the data protection law stop lenders using phone data for credit scoring?
Not outright, but the lender needs a lawful basis, must limit data to what is necessary and must be transparent. Contacts and media are rarely justifiable.
How does Nigeria compare with Singapore for consumer lending compliance?
Singapore has a separate regime for moneylenders and MAS rules for banks. See our brief on compliance agents in Singapore consumer lending.