Editor’s note: Cameron Burford is Managing Director of SaaS at Growers Edge, a company providing modern financial products and data-driven tools for agricultural retailers, manufacturers, and lenders.
The views expressed in this article are the author’s own and do not necessarily represent those of AgFunderNews.
Already, AI technology is transforming the work that ag lenders do every day. In my conversations, I hear about lenders using applications to identify loan opportunities based on crop patterns, streamline compliance workflows (with explainable decision logic, of course), and spotlight borrower needs following weather events.
Successful pilots have two things in common. First, they automate back-end workflows so that loan officers can spend even more time face to face with their customers. Second, they’re built, tested, and evaluated on a 90-day decision cycle.
Both are crucial components of successful AI deployments in ag lending, and can assist ag lenders in getting buy-in for AI pilots on a condensed timeline.
Trust is your moat. AI can deepen it
One of the most important things to keep in mind when adopting AI as an ag lender is that it shouldn’t replace human interactions or human judgment. Ag lenders are pillars of their communities. They have their customers’ trust, sometimes going back generations, which is something that can never be commoditized.
With the wrong AI product, a lender could quickly lose that trust. For example, a poorly implemented chatbot could give conflicting or incorrect answers about loan terms or approvals. That would create confusion and undermine confidence in the lender’s reliability.
The key is to find jobs for AI that let loan officers and other employees double down on the activities that build trust with customers. Consider farm visits. With the right AI agent (by which I mean an AI-powered bot), loan officers could get real-time reminders about the person they’re talking to: current loan products, family member names, past crop performance, etc.
Imagine the impact of farm visits when every loan officer suddenly has perfect recall. Farmers will feel seen, heard, and cared for.
How does all this work on the back end? One possibility is that an AI agent pulls data from a CRM and feeds it via smart glasses to the loan officers. It sounds out there, right? So let’s take a look at how something like this could be built and tested. (After that, we’ll talk about how to get the budget to do so.)
The 90-day rule: how AI decision cycles differ from SaaS
In the 1990s, an investment in smart glasses would have required a 12–24-month implementation cycle, followed by three to five years of evaluation. Seven years down the road, businesses would decide whether to keep the tech or not.
Then came SaaS cycles. An executive might deploy smart glasses in three to six months, then make a renewal decision at the end of the year. That’s where most ag lenders are comfortable today.
AI works on a different timeline: 90-day cycles. Build and deploy an application in weeks, test for 60 to 75 days, and make a decision at day 90 to scale, refine, or retire.
That can feel like a breakneck pace to organizations accustomed to the pace of SaaS. But it’s this pace where the real gains lie, since 90 days is about the length of AI’s economic cycle and every few months, infrastructure costs decrease as capabilities rise.
Right now, most AI deployments in ag lending are untested ground. The most valuable data is fast data: learning what doesn’t work (and why) from failures is just as valuable as finding an application that increases efficiency or productivity.
Everyone needs to get reps. Half the battle right now for ag lenders is getting familiar with what AI can do. While the first few apps might all get retired, they’ll equip lenders with invaluable knowledge that will guide the development of their next 10.
Organizations are just beginning to picture what this might look like in practice. The next step is to focus on how to secure buy-in and budget for initial deployments.
Getting buy-in from your board for the 90-day AI cycle
To get buy-in for AI projects and the AI investment cycle from your board, you’ll need to do two things.
The first is to explain that we’re currently at the end of the “installation” phase of AI. Per Carlota Perez’s theory of technological revolutions and financial capital, in the next three to five years we’re likely to experience a crisis driven by over-investment, inflated expectations, and unsustainable business models. This will push us into the “deployment” phase. At that point, AI stops being speculative and becomes table stakes. Organizations that were ahead of the game will suddenly become very difficult to catch up with.
Secondly, getting buy-in requires breaking down the practical reality of what 90-day AI cycles might look like in your organization. Here are a few frameworks to try:
1. Economic model shift framework
On-prem systems were all about big upfront bets and multi-year commitments. SaaS meant paying less up front and evaluating effectiveness annually. With AI, deployment, measurement, and decisions happen in 90-day cycles. This means you’re always keeping pace with the speed of AI’s evolution.
2. Budget and risk framework
Thanks to the 90-day framework, all AI investments come with limited, contained risk. At $25,000 to $75,000 per application, plus monthly inference costs of $500 to $2,000, you can experiment for far less than was possible with traditional platform purchases.
Further, without this investment, the organization risks falling behind in institutional knowledge and AI capabilities.
3. Strategic focus framework
Rather than chasing commoditized features everyone will have, an organization can focus on domain-specific applications that leverage its expertise. Fueling AI applications with unique insights from lending data, operations knowledge, and relationships will help ag lenders further dominate in their areas of strength and get stronger in current areas of weakness.
4. Regulatory framework
Every deployment will include explainable decision logic and audit trails that meet regulatory requirements. Governance will be core to every app built. By prioritizing edge deployments with human-in-the-loop validation, our technology will be easier to explain to examiners. Security protocols will be non-negotiable.
Now is the time to embrace AI—even with imperfect data
One final (but important) point illustrates the urgency of embracing AI right now: unlike with past technologies, ag lenders don’t need “perfect” or perfectly unified data to develop AI pilots that drive real gains.
That agent-in-the-smart-glasses deployment, for example, might only need CRM data to work well. This is what I mean by AI “at the edge.” Instead of a centralized system, you can deploy agents at the edges of your operations, where the data lives.
As you learn and iterate, you can continuously clean up your data infrastructure so that it’s capable of fueling more and more autonomous workflows.
Leading ag lenders are already experimenting with and learning from AI. They’re building smarter back ends that let them have more targeted, helpful, meaningful conversations with their customers and identifying likely prospects from their competitors.
If similar experimentation isn’t in your roadmap this year, you may lose your chance to join them.
The post Guest article: The ag lender’s guide to AI investment appeared first on AgFunderNews.
