AI spending is rising fast across auto finance, but results aren’t. Cox Automotive found 60% of dealers are still “testing the waters” with AI, while only about 15% have embedded it into workflows and decision-making. At the same time, dealers are signaling real caution: 74% worry about AI accuracy and errors, 60% cite concerns about data and algorithms and 66% want more education and training.
The gap between interest and impact shouldn’t be surprising. Across industries, many AI programs look promising in pilot mode but struggle to produce repeatable business value. MIT NANDA’s State of AI in Business 2025 found only 5% of custom AI tools reach production with real operational or financial impact. Ultimately, sustained impact comes down to infrastructure choices and the organizations unlocking lasting value are those treating AI as a long term capability to be built deliberately, not a quick win to be bolted on.
Generic AI breaks down in real operations
Auto finance feels this problem more sharply than many sectors because the work is complex, regulated and deeply interconnected. Credit, underwriting, servicing, collections, recovery, fraud, compliance and dealer operations all depend on data moving cleanly across systems. But the average dealership relies on more than 40 different software systems, many of which weren’t built for real time data sharing, flexible integrations, or AI enabled decisioning. The result is a fragmented environment where AI becomes just another point solution, adding complexity instead of improving performance.
This is especially risky in collections and recovery, where decisions must be consistent, auditable and sensitive to customer circumstances. Generic AI can look compelling in a demo, but in live operations it often struggles with policy nuance, disconnected account data and the need for clear explainability. In a function where compliance, accuracy and judgment all matter, this gap becomes impossible to ignore.
The real issue is the foundation
The organizations making AI work aren’t starting with the flashiest use cases. They’re building a foundation supporting multiple use cases on the same data, governance and workflow layer. A connected, AI-native framework provides a shared base for multiple agents and applications without forcing teams to rebuild core infrastructure every time a new capability is introduced. Instead of accumulating tools, organizations invest in a foundation that becomes more capable and coherent over time.
The strategic value of this approach shows up in the flexibility it creates. Teams gain a controlled environment where new ideas can be tested, validated and scaled without adding governance complexity or integration debt at every step. When regulations shift or business priorities change, the framework moves along with it.
Start small, then scale with confidence
Most institutions are sensibly starting with use cases keeping humans in the loop and carrying a well-understood risk profile. An AI agent might surface next-best-action recommendations during a collections call or deliver real time guidance to a collector navigating a sensitive conversation. This drives immediate productivity gains helping compliance teams and senior stakeholders build confidence in the technology before supporting broader adoption. This trust-building phase is often underestimated, yet it frequently determines whether a capability scales or quietly stalls after a proof of concept.
From there, the same framework can extend into workflow automation, quality assurance, compliance monitoring and increasingly autonomous customer interactions. Each new capability builds on infrastructure that’s already been validated, rather than triggering a fresh implementation cycle. The result is a system where teams can experiment under a shared governance model consistently managing risk.
Durable AI is built, not installed
The lesson for auto finance teams is simple: most AI investments fail not because the ambition is wrong, but because the foundation is weak.
The organizations pulling ahead aren’t the ones chasing every new model or vendor. They’re investing in durable foundations that enable intelligence to compound over time. By treating AI as a long term capability, grounded in the realities of collections and designed to stand up to real world scrutiny, these leaders move beyond experimentation and toward sustained, defensible value.
About C&R Software
Trusted by leading creditors in over 60 countries and more than 20 industries, C&R Software’s Debt Manager is the preferred collections and recovery solution for auto lenders. Its AI native framework provides the credibility to deliver immediate results and the flexibility to scale over time. Learn more at www.crsoftware.com.
