Deploying AI in auto and equipment finance

Summary

The latest AFC webcast, “When, how, and why should auto and equipment finance lenders deploy AI?” sponsored by NETSOL Technologies, offered a grounded and insightful conversation on the real-world application of AI across the asset finance sector.

Moderated by AFC’s AI advisor and founder of VAMOS, Richard Huston, the session brought together industry leaders with hands-on AI experience: Dario Morelli (VP of AI, NETSOL Technologies), Tony Lynch (founder and CEO of carpass.ai, and entrepreneur in residence at Toyota Financial Services), and Andy Trimmer (Head of Technology at Simply).

The central theme throughout the discussion was clarity over hype. AI, the panellists agreed, should not be viewed as a radical transformation but as a practical tool – one that must align with business goals, integrate into existing systems, and be rolled out in manageable, iterative steps.

NETSOL’s Dario Morelli set the tone by emphasizing that AI deployment only works when anchored to three pillars: business alignment, data usability, and a clear talent strategy. Without direct impact on metrics like revenue, cost, or efficiency, he warned, AI risks becoming a solution in search of a problem. “If you can’t explain how your AI initiative improves a business metric, you’re probably not ready,” he said. His message was reinforced by poll data showing that nearly 46% of organisations are only beginning to explore AI.

Andy Trimmer offered a more grounded view from Simply, where AI isn’t part of an overhaul, but a natural extension of existing processes. “We didn’t go to the business and say it’s time for an AI revolution,” he said. “We just asked: where can we make people’s lives easier?” At Simply, AI has been introduced quietly, in the background, e.g. automating document extraction, categorising emails, and providing support agents with contextual data. These are small interventions with big time savings.

Tony Lynch took a more evangelistic stance. “A day doesn’t go by where I’m not staggered by how good AI is,” he said. “Launch something – even if basic. Just get it live.” His experience working with car dealers showed how quickly AI can be deployed as a conversational sales agent. By connecting large language models to stock databases, they created systems that outperform standard website filters in minutes, not months. For Lynch, the most surprising power of AI isn’t just in answering questions, but in asking the right ones – something fundamental to selling.

Integration, however, remains a key challenge. Morelli noted that while spinning up an AI tool can be fast, connecting it to a legacy tech stack is a fundamentally different, often harder, task. This problem is especially acute in auto and equipment finance, where systems are fragmented and workflows vary. “Interoperability and integration are still among the hardest problems,” he said. Even as agentic AI and new APIs evolve to simplify connections, much of the work still falls on engineering teams rather than AI teams.

The panel widely agreed on the importance of keeping humans in the loop – particularly in risk-sensitive areas like credit underwriting. Morelli cautioned against full autonomy at this stage, citing current reliability issues with language models. Instead, augmenting human expertise with AI-generated insights offers the best of both worlds: speed and accountability. Trimmer added that organisations should treat AI missteps like any other operational risk. “As long as your team can still operate without it, you’re not going to break the business.”

One of the most telling discussions came around the poll on motivations for AI. While only 2.6% cited new revenue streams as their top priority, the panel believed this will shift. AI’s ability to serve customers around the clock, ask better questions, and extract new insights from existing data sets presents a clear pathway to growth – not just efficiency. Morelli described an advanced use case he’s exploring: AI-powered asset lifecycle optimisation across networks of OEMs, captives, and customers. It’s a complex system, requiring graph tech, machine learning, and agentic AI. But the potential to unlock entirely new business models is real.

Governance also emerged as a crucial focus. When asked who should be accountable for AI systems, the largest share of participants (37.5%) favoured a dedicated AI governance committee, followed closely by executive leadership. This reflects a growing trend in the sector toward forming AI councils that bring together legal, IT, compliance, and business leaders to jointly manage risks and opportunities. As Morelli noted, this kind of cross-functional oversight is essential for navigating the trade-offs that come with advanced deployments.

The panel concluded with a clear message: don’t delay. Powerful tools are already at your fingertips, many of them low-code or no-code. While there are genuine risks, they can be addressed with a thoughtful approach. As Lynch encouraged, “Be brave. Start basic. Learn fast.” Andy Trimmer echoed this sentiment, adding, “As long as your organisation sets clear expectations, the sky’s the limit.”

In short, AI is no longer a far-off goal – it’s a present-day tool. And as expectations from customers, brokers, and staff continue to rise, organisations that start small and build iteratively will be the ones best positioned to deliver lasting value.

Watch the webcast in full here.