How Prescriptive AI Can Transform the Future of Second-Hand Car Finance

Addressing margin compression, market volatility, and consumer affordability

The UK market has seen average increases in new car finance payments of around 20–55 per cent since 2019, driven by sharp rises in vehicle prices and interest rates. Today, typical PCP APRs range from below 4 per cent up to 6–8 per cent, further inflating the cost of finance. Across Europe, Germany exhibits similar trends, while the Nordic countries remain the most expensive places to own and maintain a vehicle.

This shift is reshaping the car-buying landscape. Many middle-income households and first-time buyers are feeling the pinch, causing them to delay purchases or abandon the option of a new car altogether, thereby fuelling robust demand for second-hand vehicles.

According to AutoTrader, the second-hand automobile market across most of Europe entered the second half of 2025 in good health, with consumer demand, sales velocity, and transaction volumes all remaining stable. However, Cox Automotive’s Insight Quarterly forecasts 7,458,356 pre-owned vehicle transactions in the UK for 2025—a modest 0.3 per cent rise on 2024—indicating a market that has plateaued after its post-pandemic surge.

Overall, the UK pre-owned vehicle sector in the first half of 2025 has demonstrated admirable resilience, with stable volumes, firm pricing, and an emergence of electrified stock. Nevertheless, lenders are contending with margin headwinds arising from regulatory liabilities, tighter credit spreads, and elevated holding costs:

• Higher stock levels and slower turnover. Many car supermarkets and franchised dealers began 2025 with larger inventories than in January 2024, yet vehicles remained on forecourts longer—average days-to-sell increased from 30 to 35 days—and sales conversion rates fell by 11 per cent for car supermarkets and 13 per cent for franchised outlets year-on-year.

• Regulatory and litigation risk. A looming Supreme Court ruling on undisclosed “discretionary commission arrangements” (DCAs) could impose over £10 billion in historic compensation liabilities, compelling lenders to set aside substantial reserves.

• Margin compression. Competitive pressures to reduce finance rates, while provisioning for possible compensation claims and bolstering delinquency reserves, are squeezing net interest margins.

• Capital tied up in ageing stock. Prolonged days-to-sell translate into higher floor plan and holding costs. Coupled with thinner margins, this erodes returns on capital for captive finance arms.

• Credit-risk volatility. Although aggregate default rates remain subdued, lenders report rising delinquencies among sub-prime and fringe customers, driven by cost-of-living pressures and uneven wage growth.

Looking ahead, modest transaction growth and potential relief on interest rates should support incremental expansion of finance business through 2026—provided lenders navigate evolving regulatory outcomes and ongoing cost-of-living challenges. The availability and affordability of finance will underpin this growth, yet lenders face turbulent times. In this blog post, we will examine how AI-based pricing analytics can help overcome these challenges and pave the way to a more profitable second-hand vehicle finance operation.

Used Vehicle Value Prediction and Loan Profitability

Having spent years immersed in the pricing analytics and motor finance industry, I know that forecasting the cost of risk is a perennial headache for any lender. As is estimating the price sensitivity of each customer. For those specialising in pre-owned vehicles, recent swings in interest rates and DTI ratios, and global trade tensions have compounded the challenge. Lenders also grapple with valuing ageing cars—predicting the future value of a two-year-old EV with 40,000 kilometers on the clock is far more complex than forecasting the depreciation of an older model with an established second-hand market. Moreover, customer risk profiles in Europe remain blunt instruments, seldom reflecting an individual’s true likelihood of default.

Prescriptive AI-based pricing analytics can change that. Where lenders once set prices on a per-vehicle basis, prescriptive models will guide them towards bespoke deals tailored to each customer. Many platforms and lenders are already rolling out these solutions; those who lag risk irrelevance in an increasingly automated and competitive market.

Predicting Residual Values

In the auto-loan business, the expected value of the vehicle at contract maturity directly influences loss severity and loan pricing. Machine-learning models can ingest vehicle attributes, macroeconomic and local indicators, plus auction and market data, to determine the optimal financing terms for each cohort. This enables motor-finance firms to:

• Produce more accurate residual forecasts

• Maintain tighter spreads with competitive APRs without sacrificing profitability

• Reduce reserve buffers

• Offer differentiated pricing to encourage high-value retention

Benefits of an AI-Driven Pricing Analytics Platform

Integrating this data into an end-to-end Pricing Analytics Platform—such as Earnix Price-It™—unifies cost-of-funds and residual predictions with price sensitivity within a single optimisation engine, allowing lenders to:

• Balance profit and risk trade-offs with confidence

• Implement real-time underwriting: as credit-bureau data and internal scores arrive, the engine instantly calculates the optimal rate for each scenario

• Personalise loan offers within regulatory guidelines: adjusting price for customer segment, deposit level and loyalty tier, all within predefined guardrails

• Run A/B tests on pricing strategies in small cohorts: reinforcement-learning agents discover which price points maximise acceptance without eroding margins

• Scale effortlessly: once trained, the system can price thousands of applications per minute with minimal manual input

By embedding AI-driven analytics into your pricing stack, you can maintain tighter, more competitive margins—and pivot swiftly as market conditions evolve.

From Prediction to Prescription

Predictive AI has already made strides in forecasting residual values and borrower behaviour. The next generation of prescriptive AI will not only anticipate outcomes but also recommend actions. Lenders are beginning to deploy prescriptive models to refine pricing strategies in real time.

• Accurate market-value estimations. Rather than relying on patchy guidebooks or intuition, lenders can leverage AI models trained on vast datasets—auctions, dealer listings and vehicle histories—to pinpoint current values. Future values can be forecast with unprecedented precision, ensuring lease-end and buyout options rest on solid financial ground.

• Customer-level insights. By integrating behavioural data, prescriptive AI can identify customers most likely to default or switch to competitors. Models can suggest targeted discounts for loyal, low-risk clients, while holding firm on rates for riskier buyers—gradually replacing the sector’s one-size-fits-all pricing approach.

Looking further ahead, some dealer networks are experimenting with conversational agents that guide customers to the ideal car—and accompanying finance package—based on lifestyle, driving habits and budget. Likewise, discerning buyers are already using AI concierges that scour local and cross-border inventories, pinpoint the perfect vehicle and arrange finance and delivery in one seamless transaction.

Finance operations, too, will be transformed. Rather than exchanging payslips and bank statements via email, customers might grant temporary access to open-banking data, enabling AI agents to assess income, spending and risk in an instant. What now takes days of paperwork could shrink to a ten-minute online journey.

On the Frontiers of an AI-Assisted Future

Far from science fiction, many leaders in vehicle finance are embedding prescriptive AI for immediate customer benefit. Earnix has recently introduced a Co-Pilot AI agent within its pricing and personalisation platform—a tool that empowers users to work more efficiently by automating complex pricing decisions, optimising loan structures and accelerating service. Drawing on over 20 years of global implementation experience, this Co-Pilot acts as an advanced internal assistant, surfacing insights and streamlining tasks.

But Earnix isn’t stopping there. In the coming years, generative AI will be woven into predictive models, enabling relationship managers to query real-time customer profiles, default probabilities and bespoke discount strategies at the point of sale.

The goal is clear: insulate lenders from today’s uncertainties and future-proof them against tomorrow’s evolving customer expectations. In a world where instant, personalised service is becoming the norm, those relying on yesterday’s tools will be left behind.

With prescriptive AI set to revolutionise second-hand motor finance, lenders must recalibrate their strategies now. AI-powered Pricing Analytics offers the way forward.

For more information about Earnix Price-It™ and its capabilities in the used-vehicle finance market, schedule a discovery call with our SME today.