Credit Connect Technology Think Tank 2025 – Detailed Session Summaries

Interesting discussion across the 4 sessions, building on the regulatory and data theme.

Regulatory pressure intensifying, with smaller lenders feeling the greatest strain.
AI adoption accelerating, but only working well where data discipline is strong.
Growing financial stress among customers, especially those already struggling.
Transparency demand is surging, with DSAR volumes rising sharply.
Increased reliance on behavioural insights to drive engagement and outcomes.
Firms shifting towards personalised, data-driven communication.
Omnichannel flexibility becoming essential – customer preferences are changing more frequently.

Key takeaways from each session

Session 1: Addressing Industry Challenges

Key Take Aways

Regulatory change and data demands (e.g. product sales data and consumer duty) are creating sustained, existential pressure for smaller and specialist lenders.
The volume and ambiguity of new reporting requirements, particularly product sales data, is high and difficult to operationalise, even for firms with strong data capabilities.
Consumer duty has sharpened thinking and led to some positive improvements, but has also significantly tightened affordability and reduced access to credit in some markets.
Lending volumes in certain specialist markets have fallen dramatically, leaving some communities without access to needed credit.
Financially vulnerable customers are struggling more deeply, even where overall arrears and delinquency volumes appear broadly stable.
There is a growing wave of digitally driven disputes, DSARs and complaints, fuelled by online content, CMCs and changing consumer expectations.
Smaller firms face a disproportionate burden from regulatory change and technology investment, particularly given static FCA fee levels since 2014.
The opacity of the FCA’s “real” expectations (e.g. around affordability) and retrospective reinterpretation of rules create material risk and cost for firms.
Technology and AI are increasingly used to monitor compliance, identify vulnerability and support customer education, but raise new risks around definitions, nuance and data volumes.
Enforcement and collections sectors are capturing far more data (recorded calls, body-worn video, case data), which can both protect firms and expose them to more regulatory challenge.
Perceptions of the collections and enforcement industry remain a major issue; better data, oversight and public education are seen as critical to shifting this.
Large-scale regulatory data collection (product sales data and other datasets) may drive blunt, outlier-based supervisory action, with particular concern for sub-prime and specialist lenders.

Innovatation

Use of AI and automation to interpret complex product sales data requirements and streamline regulatory reporting.
Application of AI to detect intent and vulnerability from speech and interaction data, enabling smarter routing and tailored questioning.
Voice analytics and similar technologies to help collections firms identify vulnerable customers consistently at scale.
Behavioural and educational interventions, such as reward programmes focused on “using credit responsibly” and explaining the impact of missed payments.
Rules-based and workflow systems in enforcement that automatically flag potentially vulnerable cases (e.g. repeat tickets, long histories) for human review.
Sector-wide data gathering by voluntary oversight bodies (e.g. Enforcement Conduct Board) to evidence actual industry behaviour and challenge misconceptions.
Use of comprehensive recording (all calls, three-way calls, video evidence) to protect both firms and customers, while supporting complaint and DSAR handling.

Key Statistics

One member firm received over 2,000 DSARs in a single day.
Product sales data reporting involved jumping between questions such as 168, 7 and 19, highlighting complexity.
The FCA’s affordability rules are described as 1.5 sides of regulation, versus around 10 pages of “true interpretation” used in practice.
In some markets, lending volumes have reduced by 80–90% following regulatory and affordability changes.
The government has spoken about cutting regulation by 25% across regulators.
Industry feedback notes no FCA fee increase since 2014, despite growing requirements.
TSB’s example involved issues going back 10 years that were only picked up recently.
The motor finance redress scheme is referenced as going back to 2007.
One member’s customer base is described as 100% vulnerable customers.
That firm answers customer calls within one minute, with all call centres in the UK.
FCA compensation expectations are described as £700 per case, compared to some CMC claims suggesting £19,000 for 3 or 4 agreements.
It is suggested that some issues can take up to 15 years to be identified and remediated.

Key Discussion Points

The cumulative burden and pace of regulatory change, and its disproportionate impact on smaller and specialist lenders.
The operational and interpretive challenges of product sales data reporting and similar granular data submissions.
Consumer duty’s impact on affordability expectations, market tightening and exclusion of some communities from credit.
The deepening financial strain on already struggling customers, even where headline arrears levels appear stable.
The surge in online-driven complaints, DSARs and “no-win, no-fee” style activity, and the pressure this places on operational capacity.
The tension between outcomes-based, principles-based regulation and the need for clear, transparent standards and guidance.
Questions over whether promised regulatory burden reductions (e.g. 25% cut) have materialised in financial services.
The role of technology and AI in monitoring compliance, evidencing good outcomes and managing large volumes of regulatory data.
How technology, recordings and data-rich oversight are reshaping the enforcement sector, including ethics, privacy and complaint dynamics.
Ongoing reforms at FOS and their impact on CMC economics, complaint volumes and claims culture.
The evolution of vulnerability management, including definitions, identification at scale and consistency of treatment across sectors.
Future risks from large-scale regulatory data pools, including outlier-driven supervision, sub-prime targeting and potential “big brother” concerns.

Session 2 : Effectively using Data & Analytics

Key Take Aways

Firms face increasing complexity in managing and interpreting large volumes of data across legacy systems, digital channels and third-party environments.
Data quality, catalogue completeness and consistent definitions remain major barriers, especially for product sales data.
Historical data is not always reliable for forecasting, given changing customer behaviour, new products and evolving regulatory expectations.
Lenders need to become more data-savvy, with staff trained to understand new data types, behavioural indicators and AI outputs.
Credit bureau and transactional data increasingly reveal early signs of stress, such as shifts in spending, reduced buffers and changes in shopping patterns.
AI adoption is growing, but customers react negatively when AI-driven insights feel intrusive or “Big Brother-like.”
Behavioural data and psychology are becoming core to collections strategies, especially in predicting engagement and tailoring channels.
Product sales data implementation has been difficult due to unclear definitions, missing data, data quality challenges and differing interpretations across firms.
Firms must integrate data into change management, ensuring future systems are designed for interoperability, measurement and regulatory reporting.
Operational teams need systems that capture the right data automatically so frontline staff can focus on conversations, not manual recording.
FCA outcomes testing requires firms to demonstrate not only operational improvements but also improved customer outcomes.
Open banking remains widely used and is becoming business-as-usual, offering easier and more accurate affordability assessment than traditional statements.

Innovatation

Using AI to detect behavioural indicators such as form-hovering, hesitation and navigation difficulty during digital journeys.
Keystroke replication and automation to accelerate data analysis and reduce manual effort.
AI-driven scorecards to route customers dynamically based on patterns rather than static rules.
Behavioural prompts and nudges embedded into collection journeys, derived from data signals.
Integrated models linking bank interaction history, DCA strategies and digital engagement to personalise collections.
APIs embedded throughout change-management processes to ensure future data capture and measurement.

Key Statistics

200–300 people working for 18 months on migration and data-related transformation at a major bank.
Some products require dual processing due to differences between legacy and new systems.
24 business areas involved in digitisation and decisioning transformation.
Behavioural programme claims an uplift achievable in four days.

Key Discussion Points

How firms choose which data to focus on amid overwhelming volumes.
The challenge of creating consistent data definitions for product sales data.
The impact of missing or unclear data on regulatory submissions and analytics.
The growing importance of behavioural insights in collections decisioning.
The tension between using AI for efficiency and maintaining customer trust.
Integrating data across legacy environments, third-party DCAs and disparate systems.
The need for data-driven change-management frameworks tied to KPIs and customer outcomes.
Ensuring frontline staff capture data consistently while maintaining high-quality customer conversations.
Future readiness: designing systems to support regulatory reporting, API connectivity and measurement.
The shift from manual affordability assessments to open banking-driven insights.
How firms evidence that digital changes improve customer outcomes, not just KPIs.
Preparing for increased FCA data use, including thematic analysis of product sales data and new reporting.

Session 3 : Value of Communications

Key Take Aways

Firms must deeply understand customer mindsets, especially stress behaviours, avoidance tendencies, and short-term thinking, to design effective communications.
“Customer first” operating models, with roles dedicated to representing the customer voice, support clearer and more empathetic engagement.
Written communication must use low reading ages, simple design and clear calls to action, reflecting modern expectations and reducing misunderstanding.
Omni-channel strategies remain essential, with customers shifting channel preference during different stages of their financial journey.
Telephony continues to play a critical role, especially for vulnerable customers and for detecting needs that digital cannot surface.
AI can personalise content, adjust tone, tailor reading levels and recommend the most effective communication channel based on behavioural signals.
Customer communication preferences can be both stated and inferred; firms must combine data with human judgement to determine true needs.
AI and machine-learning models require continuous monitoring to avoid drift, bias and unintended outcomes, with human-in-the-loop models recommended.
Increasing automation can free staff for more sensitive customer interactions, but requires strong controls and ongoing performance review.
Regulation is shifting towards outcomes-based expectations; firms must evidence consumer understanding and fair outcomes across channels.
Explainability in decisioning is essential, particularly as AI-assisted credit and collections processes expand.
Measuring communication effectiveness requires behavioural metrics, customer actions, complaints, sentiment and channel-level data.

Innovatation

AI-driven personalisation of tone, reading age and message style for each customer.
Behavioural analytics highlighting hesitation, navigation patterns and disengagement points in digital journeys.
GenAI optimisation (“GEO”) for marketing and acquisition, aligned to new search behaviours.
Automated recommendation engines suggesting optimal contact time, channel and approach based on live customer data.
Large-scale speech and call analytics for vulnerability detection and sentiment analysis.
Integrated omni-channel data synchronisation enabling seamless movement between channels.

Key Statistics

HMRC previously reported 300,000 customers receiving communication at one stage in a major digital shift.
One firm reports 40% of its customers have at least one present vulnerability, with an average of three vulnerabilities.
Klarna analysis showed younger customers (25–30) had lowest chatbot satisfaction, while those 31–35 had the highest.
Call-centre communication reading level targeted at around age 11 for clarity.

Key Discussion Points

How to identify and influence customer mindset during periods of financial stress.
The limits of customer journey mapping when customers behave differently under different circumstances.
Importance of low-friction omni-channel journeys with customer-driven channel switching.
Balancing channel complexity with cost, operational capacity and customer expectations.
Managing AI-driven personalisation ethically, avoiding bias and ensuring explainability.
The evolution of channel preferences across a customer’s lifecycle and changing financial circumstances.
How to infer communication preferences from historical behaviour rather than relying only on stated preferences.
Ensuring transparency and auditability of AI-supported decisions under growing regulatory scrutiny.
How to design communication journeys that still preserve human support at critical points.
Using communications to gather actionable data for segmentation, personas and behavioural modelling.
Ensuring customer understanding of key information, especially costs, product mechanics and outcomes.
Preparing communication strategies for 2026, focusing on data-driven personalisation, AI literacy and clear value measurement.

Session 4 : Future of Technology in Credit & Collections

Key Take Aways

Investment in technology and AI has accelerated across the credit and collections sector in 2025
Organisations are shifting focus from “AI hype” to solving concrete business problems
Foundational data quality and modern data infrastructure are now seen as essential enablers
Legacy systems remain a barrier, but legacy policy and legacy thinking are often bigger blockers
Generative AI is driving interest, but its value depends on strong context, governance and business alignment
Human-in-the-loop decision-making remains critical and is unlikely to disappear
Personalised journeys, personalised communications and dynamic context windows are emerging priorities
Data security, model safety and clarity on how models use data are rising concerns
The industry must prepare for AI-driven consumer behaviour, including automated consumer requests
Fraud risk is increasing as criminals adopt AI more quickly than regulated firms
Firms must plan for business continuity when AI platforms or cloud infrastructure fail
The next stage of AI adoption will move from hype to demonstrable business value

Innovatation

Use of dynamic context windows to tailor LLM outputs to customer, business and industry rules
Embedding multiple small language models and large language models within platforms
Personalised video communication using new generative media tools
AI-driven triage for contact centres, reserving human agents for specialist and high-value cases
Development of hyper-specific models built on industry-specific context
Automated consumer-initiated interactions driven by AI personal finance tools

Key Statistics

85% average automation rate referenced by a vendor across client implementations
300 DSARs received by a firm in a previous example, with predictions of thousands driven by AI
70% potential contact-centre reduction discussed as a hypothetical business objective
5–10% expected error rates if humans were fully removed from certain AI-driven decisions

Key Discussion Points

AI hype versus real business value and the need for problem-first design
Foundational importance of modern data architecture before deploying AI
Challenges integrating AI with legacy systems, processes and policy frameworks
Security, data governance and misconceptions about “safe” versions of public AI tools
Potential regulatory constraints on removing humans from decision-making
The likelihood that customers will adopt AI automation faster than creditors
Risks of AI-generated consumer complaints and mass automated requests
Business continuity implications of AI or cloud outages
Fraud amplification through AI tools
The need for AI literacy and fluency across organisations
Emerging personalisation technologies in collections communication

RO-AR insider newsletter

Receive notifications of new RO-AR content notifications: Also subscribe here - unsubscribe anytime