Reimagining asset finance with AI: practical wins now, bigger shifts ahead

Summary

Asset finance is still catching up with other financial services sectors in adopting artificial intelligence. A live poll during the latest Asset Finance Connect webcast, sponsored by Solifi, revealed that more than a quarter of delegates have not yet started their AI journey – a striking indicator of how early-stage the sector still is. While document processing and automation are the most common entry points, the picture across the industry remains fragmented, with barriers around integration, data quality and skills slowing progress.

Against this backdrop, the webcast – moderated by Richard Huston, managing director and co-founder of Vamos, and featuring Karan Oberoi, chief product officer at Solifi; Christian Brough, head of platform solutions at Haydock Finance; and Ed Ockleford, head of operations at Liberty Leasing – set out to separate hype from reality.

The focus was less on futuristic promises and more on what firms are deploying today, how they are measuring impact, where they are encountering obstacles, and what might shift most significantly over the next two to four years.

What are businesses doing today?

Much of the early traction, the panel agreed, has come in low-risk, high-value use cases. Oberoi pointed to two clear front-runners: chatbots and document intelligence. “The obvious ones we’ve seen for sure are chatbots,” he said. “You see those plugged into different applications or portals so you can interact and get quick support. And the other area that’s really picking up is document intelligence – faster approvals on one side, but also cheaper operations.”

While many companies have started internally with knowledge assistants, Oberoi noted that external-facing chatbots are becoming more viable. “Beyond the initial hallucinations we saw back in 2022 and 2023, there’s now enough technology, tooling and best practices to really lock down what your chatbot can do.”

At Haydock Finance, Brough described a deliberate progression from exploratory pilots to embedded operations. “We started looking at this around two years ago. The very first use case was a standalone solution, really just to see what the power of the technology was. Now we’re moving into embedding it into our processes.” The company has focused on proposal loading and fraud detection, but Brough stressed that the human role remains central. “Our approach is very much about keeping the human in the loop. The technology is all around supporting that, doing the heavy lifting. We’re still 100 percent manually in control.”

Liberty Leasing has taken a slightly different tack. Ockleford explained that the first priority has been to digitise workflows and connect siloed systems, while experimenting with contained use cases such as automated bank statement analysis. “We don’t really want to commit to anything too quickly and get bogged down in systems we can’t get out of,” he said. “With AI agents coming out, it’s really important to define what you want those agents to do. For us, a good work case would be internal audits, where the AI can focus on a small, well-trained task. Again, it’s all about assist, don’t bypass.”

Audience polls reinforced the sense that the asset finance industry is still at the beginning of the journey. Document processing was the most common use case followed by credit and risk assessment, while a quarter of respondents admitted they had not yet started implementing AI into their processes.

Delegates agreed that system integration challenges and skills gaps were the biggest barriers to implementing AI, with compliance concerns close behind.

For Oberoi, this underlines why firms must focus on the fundamentals. “It’s about data and disparate systems. You want to bring that together, feed all the context you can, to unlock value. You may not need to make big choices on advanced AI today, but you definitely want to get your data ready today.”

Huston described this as “context engineering” – the work of connecting the right data to models at the right time.

Standalone vs modular platforms

The webcast also highlighted different strategies around platform choice. Haydock’s initial implementation was a standalone AI system, kept separate from its core operations. This gave the business a safe way to test functionality before moving towards embedded tools directly within proposal workflows and fraud detection. This progression reflects a cautious but confident approach: learn from controlled pilots, then expand into mission-critical processes with manual fallbacks in place.

Liberty, on the other hand, is deliberately adopting a modular approach. Ockleford emphasised that Liberty looks for systems that can be quickly implemented, easily integrated and, crucially, swapped out if they do not deliver. Shorter contracts, flexible commercial terms and robust APIs are central to their evaluation process. This flexibility ensures Liberty remains agile and avoids being locked into technologies that may become obsolete as AI capabilities evolve.

ROI and value creation

Looking at return on AI investment, the panel was quick to push back on headlines suggesting most firms are seeing no benefit. Oberoi argued that the reality is more nuanced: “I would hesitate with a zero return. If you’ve really thought through what you want to achieve, there’s definitely value. But it’s not instant – you don’t get instant value from everything. The ROI is usually over a period of time.”

Brough agreed, noting that Haydock Finance has created an AI Working Group with members from different departments across the business, who are maintaining a roadmap and building business cases for every AI project. “We have 12 to 15 use cases on our roadmap. Each one only goes ahead once we’ve looked at the benefits case and ensured there’s ROI. I don’t see how we would be getting zero return from those.”

For Ockleford at Liberty Leasing, even softer outcomes matter. “It seems to me too early to try and put a hard ROI on it. Winning hearts and minds of staff, making sure they’re comfortable and see AI as assisting rather than replacing them – that’s still a return.”

Future of AI in asset finance

Looking ahead, the panel foresaw a mix of continuity and change over the next two to four years, while webcast delegates saw AI having the biggest future impact in operational cost reduction, speed of decision-making (days to minutes), and enhanced customer experience.

Brough predicted that expectations will shift dramatically: “The genie is out of the bottle. In 1986 people were willing to wait two hours for an eight-bit game to load. Now they get frustrated if a hyper-reality game takes five minutes. The same will happen in asset finance. We’ll have to get quicker, and technology will have to support that, because we can’t just keep hiring tens and tens of people.”

Ockleford, however, warned against overreach: “Do we want an AI agent making that first collections call to find out why a payment’s been missed? Probably not. That’s a good opportunity to learn from the customer, and can we trust a bot to judge if they’re vulnerable? Not yet.”

Solifi’s Oberoi offered a broader vision of augmentation rather than replacement: “The real beauty lies in converting your humans into super-humans – empowering them with context and insights to serve customers better through every stage of their journey.”

The webcast provided a measured and realistic snapshot of where the sector stands. AI is already delivering tangible value in knowledge management and document processing, but the transformational potential will only be unlocked through integration, careful scoping and organisational readiness.

As Ockleford neatly summed up, “It’s not about being reliant on AI. It’s there to assist us, not replace us.”

Watch the full webcast on-demand here.