AInora
United Kingdom - 2026

AI Debt Collection for UK Lenders

Voice AI calibrated to FCA Consumer Duty (PRIN 2A), the CONC Handbook, ICO data protection and SMCR. Consumer understanding, vulnerability detection and full transcripts on every call.

AI debt collection for UK lenders is the use of voice AI agents to contact customers in arrears, run affordability conversations, propose forbearance, escalate vulnerable cases to a human specialist, and produce a fully transcribed audit trail that satisfies FCA, ICO and FOS evidencing requirements.

Hear the live demo +1 (332) 241-0221Speak to the team
VendorVoice approachUK conduct fitVulnerability handlingDeploymentBest for
AINORABespoke per portfolioCONC + Consumer Duty nativeReal-time sentiment + escalationInside firm environmentRegulated UK lenders that want auditability
Skit.aiGeneric SaaS voiceUS-first, UK retrofitStandard scriptsVendor cloudHigh-volume US portfolios
InDebtedDigital-first messagingActive in UK, FCA-awareHardship workflowsVendor platformBNPL and digital-native portfolios
ProdigalAgent assist plus voiceUS-firstQA scoring on humansVendor cloudAugmenting US human teams
TrueAccordEmail/SMS-ledUS-onlyDecision-engine routingVendor platformUS digital-only collections
VodexOutbound voiceLimited UK referencesBasic sentimentVendor cloudOutbound campaigns at scale

Transparency note: AINORA appears in this comparison. We have written each row evenly. Verify each criterion with the vendor before deciding.

FCA Consumer Duty: four outcomes for collections

PRIN 2A applies in full to collections journeys. The voice AI is configured to evidence each of the four outcomes on every call, every day, on every customer.

Consumer Duty outcome

Products and services

Collections journeys must be designed for the target customer base, not the easiest debtor cohort. The voice AI is configured per portfolio so that vulnerable customers in arrears receive a different journey than commercial debtors. The journey is documented, version-controlled and reviewable by the firm's board.

Consumer Duty outcome

Price and value

Under PRIN 2A.4, the firm must show that fees and charges applied during collection deliver fair value. The voice AI logs every fee disclosure verbatim, including any default sums and contractual interest, so the audit trail evidences fair-value reasoning rather than relying on humans to remember what they said on the call.

Consumer Duty outcome

Consumer understanding

The customer must understand the communication. The AI uses plain English, slows the pace when sentiment indicates confusion, and offers to repeat the affordability question, the arrangement and the consequences of non-payment. Every disclosure is captured in the transcript and can be sampled by Compliance Monitoring.

Consumer Duty outcome

Consumer support

A firm must not put unreasonable barriers between the customer and a human specialist. The voice AI offers to transfer to a human at any point, and always escalates when it detects vulnerability indicators. Hold times, drop rates and human-handoff success rates are reported back to senior managers under SMCR.

Why Consumer Duty raises the bar for collections

Consumer Duty went live for open products on 31 July 2023 and for closed products on 31 July 2024. It replaces the older treating-customers-fairly principle with a sharper, outcomes-based standard and gives the FCA stronger supervisory tools. For collections, this means a firm can no longer rely on a thin script and a quality-monitoring sample to demonstrate fair customer treatment.

The board, the senior manager accountable under SMCR, and Compliance Monitoring all need management information that ties customer outcomes to specific call behaviours. Voice AI provides this natively: every word is captured, every disclosure is logged, every forbearance offer is time-stamped, and every vulnerability flag is attached to a case record.

For more detail, the FCA publishes the Consumer Duty rules and guidance at fca.org.uk/firms/consumer-duty.

Sources: FCA PRIN 2A, FCA Consumer Duty rules and guidance, FCA Consumer Duty supervision strategy 2024.

CONC Handbook: arrears, default and recovery

CONC 7 is the chapter that governs UK collections. The voice AI is configured against each relevant section, with the dialler, the dialogue tree and the data flow all enforcing the rules at platform level.

CONC 7.3 - Treating customers in default or arrears with forbearance and due consideration

CONC 7.3.4R requires lenders to treat customers in default fairly and consider forbearance options. The voice AI is configured to offer affordability assessment, payment holidays, reduced payment plans, and signposting to free debt advice (Citizens Advice, StepChange, National Debtline) before applying any escalation. These options are scripted into the dialogue tree, not left to the discretion of an individual collector.

CONC 7.9 - Contact with customers

CONC 7.9 limits how often and when a firm may contact a customer. The voice AI honours dialler caps, suppression flags, do-not-contact preferences and quiet hours by default. The cap configuration is owned by the SMF holder for collections, not by the agent on the floor, which removes the historic compliance gap where humans dialled outside the window.

CONC 7.13 - Data accuracy and outsourcing

CONC 7.13 requires collections data to be accurate and any outsourcing to be supervised. The voice AI streams call outcomes back to the firm's system of record in real time, eliminating the lag between conversation and CRM update where data drift typically appears. When the firm uses a debt purchaser or DCA, the AI can be deployed inside the supervisor's environment so the regulated firm retains line of sight.

CONC 7.14 - Settlements

CONC 7.14 governs settlement offers and the disclosure required when a discount is offered. The voice AI generates settlement offers from a pre-approved matrix only, never on the fly, and reads the standard settlement disclosure verbatim. The transcript becomes the evidence pack for any subsequent customer complaint or Financial Ombudsman Service review.

Source: FCA CONC Handbook, handbook.fca.org.uk/handbook/CONC.pdf.

Powered by industry-leading technology

OpenAIAnthropicAWSGoogleElevenLabs●TelnyxOpenAIAnthropicAWSGoogleElevenLabs●TelnyxOpenAIAnthropicAWSGoogleElevenLabs●Telnyx

Four UK collections use cases for voice AI

Where voice AI delivers the strongest economics and the strongest regulatory evidencing in the UK lending market.

UK BNPL portfolios under FCA regulation (2026)

HM Treasury confirmed in October 2024 that Buy Now Pay Later providers including Klarna, Clearpay and similar platforms become FCA-regulated in 2026. Once authorised, BNPL arrears must be handled under CONC 7. The voice AI carries out the affordability check on missed payments, offers forbearance, signposts to free debt advice, and only escalates to human collectors when vulnerability is detected or the customer requests it.

Credit card and unsecured personal loan arrears

Credit card arrears are the largest unsecured collections segment in the UK by value. The voice AI handles the early-stage reminder, captures the reason for missed payment, runs an income and expenditure conversation aligned to the StepChange common financial statement, and proposes a CONC-compliant repayment plan. Human collectors only see the cases that genuinely need human judgement.

Motor finance after the DCA scandal

After the Court of Appeal's 2024 ruling on undisclosed motor finance commissions and the FCA's ongoing review, motor finance lenders face heightened scrutiny on every collections call. The voice AI provides a fully recorded, fully transcribed audit trail for every customer interaction, addressing the historic problem that DCA-era cases lacked recordings or had inconsistent collector behaviour. Disclosures, complaint handling pointers and FOS referral language are scripted in.

Utility arrears and Ofgem ability-to-pay rules

Under Ofgem's standard licence conditions, energy suppliers must take ability to pay into account when recovering arrears. The voice AI runs a structured ability-to-pay conversation, proposes Fuel Direct or Priority Services Register signposting where indicated, and never threatens disconnection without the regulated process. Every conversation is captured against the customer reference for Ofgem audit.

Try before you brief

Hear the live demo

Phone the AI, push it on vulnerability scenarios, ask for a forbearance plan, request a human handoff. Test what your customers will actually experience.

+1 (332) 241-0221

Available 24/7. No registration required.

Speak to the team

Vulnerability detection: the four FCA drivers

The FCA Financial Lives Survey reports that around 58 per cent of customers showing vulnerability characteristics did not disclose them to their financial provider. Voice AI surfaces signals that human collectors miss.

Vulnerability driver

Health vulnerability

The customer mentions a serious illness, mental health crisis, recent diagnosis, or care responsibility for a family member. The voice AI lowers pace, expresses empathy, offers to pause the conversation, and either schedules a callback or transfers to a specialist trained in vulnerability handling.

Vulnerability driver

Life-event vulnerability

Bereavement, separation, redundancy, or a sudden drop in income. These trigger a forbearance branch in the dialogue tree: hardship plan, payment holiday, signpost to StepChange or Citizens Advice, and a temporary suspension of arrears charges where the firm's policy allows.

Vulnerability driver

Capability vulnerability

Difficulty understanding the call, low literacy or numeracy, English as a second language, sensory impairment. The AI slows pace, repeats key points, offers a translated callback, and surfaces the option of nominating a third party to speak on the customer's behalf under the firm's power-of-attorney policy.

Vulnerability driver

Resilience vulnerability

No savings buffer, sole earner, dependent children. The AI runs the income and expenditure conversation gently, never pushes a settlement that would leave the customer worse off, and signposts to free debt advice charities. Resilience signals are flagged to senior management dashboards weekly.

Why machine sentiment beats the human baseline

A human collector handling several hundred calls a week will not remember every voice tremor, every long pause, every drop in fluency. The voice AI runs sentiment analysis on every utterance and never tires. When a vulnerability driver is detected, the conversation routes into a forbearance branch, not the standard arrears branch. The handoff to a human specialist comes with full context, so the customer does not have to repeat their story.

This is the operational case for AI in UK collections that resonates with the senior manager accountable under SMCR: a higher floor on conduct quality, captured on every call, with management information that holds up to FCA supervision and FOS scrutiny.

ICO data protection and SMCR senior manager controls

UK GDPR and the Data Protection Act 2018

Under UK GDPR, the lawful basis for collections contact is typically legitimate interest under Article 6(1)(f) or contractual necessity under Article 6(1)(b). Special category data, such as a health disclosure during a vulnerability conversation, is processed under the relevant Article 9 condition, usually substantial public interest with the appropriate safeguard policy.

The voice AI never makes a solely automated decision producing legal effects on the customer. Recommendations for forbearance, settlement, or escalation are surfaced for human review before action. This sits within Article 22 UK GDPR cleanly. Recordings, transcripts and metadata follow the firm's retention schedule. Right-to-erasure requests are routed through the firm's standard subject access process.

For the ICO's guidance on automated decision making and AI, see ico.org.uk.

Senior Managers and Certification Regime

Under SMCR, a senior manager holds the prescribed responsibility for the firm's collections operation. That responsibility does not move when the firm deploys voice AI; it remains with the human at the top of the function. What changes is the quality of the management information available to that senior manager.

Instead of monthly QA samples on a fraction of calls, the senior manager receives complete-population MI: every disclosure, every forbearance offer, every vulnerability flag, every handoff to a human, every customer outcome. This is what the FCA expects when it asks how a senior manager has reasonable steps in place. With AI, the reasonable-steps file practically writes itself.

More on AI collections

Related resources across the AINORA collections cluster.

AI Debt Collection in Europe (GDPR)

EU-wide overview: GDPR, ePrivacy, member-state collections rules from Germany to Spain.

Read more

UK FCA Compliance Guide for AI Collections

Long-form guide to building a Consumer Duty and CONC compliant AI collections programme.

Read more

AI Debt Collection (Pillar)

Global overview of AI voice agents in the collections lifecycle, all verticals.

Read more

Skit.ai Alternatives

Ten Skit.ai alternatives compared, including UK-active vendors.

Read more

InDebted Alternatives

Comparison of InDebted alternatives for digital-first collections.

Read more

Frequently asked questions

AI debt collection for UK lenders is the use of voice AI agents to contact customers in arrears, run identity verification, conduct income and expenditure conversations, propose repayment plans, log outcomes to the lender's system of record, and escalate vulnerable cases to a human specialist. The AI operates inside the FCA Consumer Duty framework (PRIN 2A) and the CONC Handbook, with full transcripts retained for ICO data protection and Financial Ombudsman Service evidencing.
Consumer Duty introduced four outcomes: products and services, price and value, consumer understanding, and consumer support. The voice AI evidences each one. Products and services: portfolio-specific journeys configured per cohort. Price and value: every fee disclosure logged verbatim. Consumer understanding: plain English, paced delivery, on-demand repetition. Consumer support: human handoff offered at any point and always triggered on vulnerability detection. Senior managers under SMCR receive monthly Consumer Duty MI showing how each outcome is performing.
CONC 7 covers arrears, default and recovery and is the primary chapter. CONC 7.3 requires forbearance and due consideration for customers in default. CONC 7.9 caps contact frequency. CONC 7.13 requires accurate data and supervised outsourcing. CONC 7.14 governs settlements. The AI is configured against each section: forbearance options scripted in, dialler caps enforced at platform level, real-time CRM updates to keep data accurate, and pre-approved settlement matrices read verbatim. The full CONC Handbook is at handbook.fca.org.uk/handbook/CONC.pdf.
The AI runs real-time sentiment and content analysis on every utterance. It looks for the four FCA vulnerability drivers: health, life events, resilience, and capability. The FCA Financial Lives Survey reports that around 58 per cent of customers showing vulnerability characteristics did not disclose them to their financial provider. The AI surfaces signals that humans miss: voice tremor, long pauses, mentions of health or bereavement, drops in fluency. Once detected, the conversation routes into a forbearance branch and a human specialist is offered. Detected vulnerability flags are written to the case record.
Yes, provided the firm meets its obligations under PRIN 2A (Consumer Duty), the relevant CONC chapters, SMCR senior management responsibility, and ICO data protection law. There is no FCA prohibition on automated voice agents in collections. The FCA expects firms to evidence fair customer outcomes regardless of channel. AI delivers stronger evidencing than humans because every word is transcribed and time-stamped. The senior manager accountable under SMCR retains overall responsibility, the same as for any third-party-supported collections operation.
Under UK GDPR and the Data Protection Act 2018, the lawful basis for collections contact is usually legitimate interest or contractual necessity, with a fallback to consent for marketing-flavoured outreach. The voice AI processes special category data (for example, health information disclosed during a vulnerability conversation) only when the customer raises it voluntarily, and stores it under the firm's Article 9 condition. Recordings, transcripts and metadata are retained per the firm's retention schedule, with right-to-erasure handled through the firm's standard process. The AI never makes a solely automated decision producing a legal effect on the customer; recommendations are surfaced to human reviewers for final action.
HM Treasury confirmed in October 2024 that BNPL providers including Klarna and Clearpay become FCA-regulated. From the moment the regime is in force, BNPL arrears must be handled under CONC 7 like any other unsecured credit product. That means affordability assessment, forbearance, signposting to free debt advice, dialler caps, and Consumer Duty outcomes evidencing. Voice AI is well placed to absorb the volume because BNPL portfolios are large, low average balance, and rule-driven, which is exactly where AI delivers cost-per-contact economics that humans cannot match while still meeting the conduct rules.
The team builds a bespoke API integration with the lender's system of record, whether that is a tier-one collections platform, a homegrown CRM, or a debt purchaser's case management tool. The AI streams call outcomes, payment promises, vulnerability flags, and full transcripts back in real time. Telephony connects through the lender's existing carrier or through a regulated route the team configures. Deployment can be inside the lender's VPC if the firm requires that for ICO and SMCR control reasons.
Yes. CONC 7.13 makes the regulated firm responsible for outsourced collections, so when a DCA or debt purchaser deploys the AI, the originating lender retains line of sight through the same MI dashboards, transcripts and outcome reports. This is one of the strongest reasons to move from a black-box dialler to an evidencing-first AI: the principal can review every disclosure, every forbearance offer, and every escalation, rather than trusting QA samples on a small fraction of calls.
When a customer indicates dissatisfaction, the AI offers the firm's complaints route, captures the complaint reason, and creates a complaint case in the system of record. It reads the FOS referral wording verbatim where the firm's policy requires it (eight-week final response, six-month FOS window). The transcript becomes the primary evidence document if the case reaches FOS. Because the AI does not improvise or paraphrase regulatory disclosures, FOS evidence packs are stronger and more consistent than the human-collector baseline.
Skit.ai is a US-first SaaS platform optimised for high-volume American collections environments where TCPA and Reg F dominate; UK conduct rules are a retrofit. InDebted is a UK-active digital-first platform strong on email, SMS and self-serve flows. AINORA builds bespoke voice agents inside the firm's environment, calibrated to FCA Consumer Duty and CONC from the first deployment, with full transcripts and SMCR-grade evidencing. For pure-volume US-style dialling, Skit wins. For digital-channel BNPL, InDebted is strong. For UK regulated lenders that need a compliance-grade voice channel deployable inside their own perimeter, AINORA is the fit.
The fastest route is to phone the live demo: +1 (332) 241-0221. You will speak to a voice AI configured as a UK collections agent. Test the income and expenditure conversation, push it on vulnerability scenarios, ask for forbearance, and request a human handoff. After the demo, book a session with the team at https://ainora.lt/contact and we will share a deployment plan calibrated to your portfolio mix, your collections platform, and your SMCR control framework.
JB
Justas Butkus

Founder & CEO, AInora

Building AI digital administrators that replace front-desk overhead for service businesses across Europe. Previously built voice AI systems for dental clinics, hotels, and restaurants.

View all articles
Bespoke configuration

Deploy a UK-conduct-grade voice AI in your collections operation

Phone the demo or book a session with the team. We will scope to your portfolio mix, your collections platform, your SMCR control framework, and your supervisory expectations.

+1 (332) 241-0221

Live demo. Available 24/7.

Book a scoping call