Over the past two months, I've been fortunate to attend three key equipment finance conferences: ELFA's Innovation Lab in Nashville, NEFA's Spring Conference in Huntington Beach, and ELFA's Funding Conference in Chicago. Across those events I spoke with more than 200 lenders, each with distinct challenges, and goals they need technology to achieve. Nearly all of them asked me the same question: How should we use AI?

For most equipment finance companies, getting real results from AI solutions has been a struggle. They see the outcomes AI companies promise, but they can't achieve them in their own businesses. The issue isn't awareness; I found that most leaders clearly understand their processes and where the bottlenecks are. What they struggle with is how to break those processes down into decisions and actions that can be automated safely. 

The result is a sharp understanding of what needs to change, but almost no clarity on the solution.

Spackle Won't Save You

It reminded me of remodeling my house in St. Paul. I could handle basic maintenance (thanks, YouTube!), but for more complex tasks, I needed an expert to diagnose the problem and to take action where I couldn't. Not just someone to do the work correctly, but someone who could identify the problems my inexperience would miss and understand how every room, wall, and system connects and supports everything else.

That expert's real value wasn't knowing how to do things right. It was knowing how to call out what was wrong. Because if a building needs structural repair and doesn't get it, the problem festers. And it only gets worse the closer it gets to the foundation.

The same is true in our industry. Businesses built on manual processes and legacy systems can't simply apply AI to what they're already doing and expect better results. The outcome produces the same failures, just faster.
If we apply AI like spackle over a cracked foundation, the underlying issues remain – often worsening over time. But if we fix the foundation first and then build AI into a structure designed for it, we'll unlock those efficiency gains in our own businesses. 

Equipment finance has yet to truly realize the potential of interconnected systems and plug-and-play solutions, for a number of reasons. Chief among them being that the industry's technology foundation simply can't support its ambitions.

The Uncomfortable Truth

The uncomfortable truth under most AI conversations in equipment finance is that the data those tools require doesn't exist in usable form. Not from neglect, but because the architecture was never built to produce it cleanly.

It's an open secret that technology in equipment finance is lagging well behind consumer lending. Platforms common today were built long before the advent of AI with architecture that can't support the real-time data access needed to maximize such tools. 

Building an integration today is like laying out a new dirt road: a manual implementation that meets the bare minimum requirements, but you'll be doing constant repairs and will find some nasty potholes that send tools flying out of the truck bed.

Those potholes force teams to create workarounds to overcome inaccurate data handling, errors in handoffs, and consistent manual reconciliation. The dirt road itself isn't the problem; the integrity of the architecture and the maintenance required to keep it functional is.

The answer isn't more patches. No matter how many times you patch a dirt road, it's still a dirt road. What the industry needs is a platform built for where we're going, not where we were in 2005.

Build for the Future, Not the Past

It also means we need to evolve our collective understanding of the ideal tech stack for equipment finance. Lenders have been chasing the dream of a single platform that is a central source of truth and hub for everything from originations through servicing and end-of-life. However, no technology platform exists to handle that whole lifecycle that isn’t rife with legacy tech and expensive integrations that require frequent maintenance like those dirt roads.

AI-enabled solutions built on API-first technology can be evaluated, piloted and deployed in weeks, and integrate easily with other data sources. New AI tools are brought to market daily that could provide immediate benefits to equipment finance teams, but most lack the infrastructure and access to the data schema needed to make those AI tools useful.

For organizations stuck on legacy platforms, taking advantage of any of these new tools means going back to your vendor, negotiating a workaround, and paying for the privilege. And when the integration is ready, there’s a new solution that does the job better for cheaper, starting the process over from scratch.

Where to Start

You don't have to rebuild everything at once, but you do have to be honest about what you're working with. Here are a few practical places to start:

• Diagnose and catalogue your challenges: Before you evaluate any new AI tool or platform, ask yourself what data that tool will need access to, and learn if that’s possible today. If your team is spending real time on manual reconciliation, you already know the answer.

• Involve your team before you commit to a solution: The people doing manual reconciliation every day know exactly where the architecture is breaking down. Bring that input into vendor conversations – it’s the only way to know whether a demo reflects your reality.

• Require proof, not promises: On-time and under-budget implementations happen when the foundation is validated before construction starts. Insist on a structured pilot with real data; not a reference call, not a sandbox environment that doesn't reflect your actual portfolio.

• Stop investing in maintenance on a road that needs to be replaced: Every integration layer, every reconciliation workaround, every manual process filling a gap is money spent preserving a foundation that's holding you back. At some point the repair costs more than the rebuild.

The lenders who come out of this cycle ahead will be the ones who paused, assessed honestly, and fixed what needed fixing before building on top of it.

You wouldn’t build your house on a cracked foundation. Don’t build your AI strategy on one, either.