.svg)
Year one: Let's experiment with AI! Education staff use AI for course descriptions. The membership department uses AI to draft renewal emails.
Year two brings more experiments. Your marketing team builds custom GPTs for content creation. The finance team uses ChatGPT for some basic analysis. Small wins accumulate.
Year three: Your association launches a chatbot for basic member questions. Success! Response times drop dramatically.
But zoom out: You're still fundamentally the same organization, just with some AI sprinkled on top. Revenue hasn't dramatically shifted. Member experience hasn't been revolutionized. You haven't reimagined what an association can be. Welcome to the innovation plateau—where successful experiments never quite become transformation.
Understanding the Bullets and Cannonballs Framework
In his book Great by Choice, Jim Collins shares a powerful metaphor that explains exactly what's happening in associations right now. Picture yourself as a ship's captain in thick fog. Somewhere in the darkness is an enemy vessel. You have limited gunpowder—enough for lots of small bullets or a few massive cannonballs. The smart strategy seems obvious: fire bullets first. Small, cheap shots to find your target. Listen for the ping of metal on metal. Adjust your aim. Fire again. Once you hear that consistent ping—once you know exactly where to aim—that's when you load the cannonball.
But here's what Collins discovered in his research, and it's counterintuitive: Companies rarely fail from firing cannonballs too early. Instead, they fail from never firing them at all. They perfect their aim with endless bullets, hear the ping repeatedly, validate their target over and over, but never take the big shot. They get so comfortable in the experimentation phase that they forget the whole point was to prepare for transformation.
This pattern shows up everywhere, but it's particularly acute in associations right now with AI adoption. You've been firing bullets for two or three years now. That chatbot handling routine member questions? That's a direct hit. The custom GPTs your departments built for their specific workflows? Another ping. The AI tools that draft emails, summarize meetings, and create first drafts of content? You're consistently hitting your target. The calibration phase is complete. You know AI works because you've proven it in your own organization repeatedly.
Why Associations Get Stuck in Perpetual Experimentation
The innovation plateau happens when organizations confuse activity with progress. Yes, you're doing more with AI each year. Yes, each new tool succeeds. But adding more AI experiments isn't the same as AI transformation. If you have ten successful AI tools each serving a different department or function, you don't have transformation—you have ten experiments running in parallel.
Part of this stems from how associations typically approach change. It's easier to get approval for a small experiment than a transformational investment. Testing with a small group feels responsible. Fundamentally changing how you deliver value feels risky. So you stay in experiment mode, where success is measured by time saved in individual departments rather than organizational impact.
There's also a resource allocation problem. When your AI initiatives are scattered across departments, each getting a small budget and part-time attention, you're guaranteed to stay small. True transformation requires concentrated resources—the kind that might make finance nervous and requires hard choices about what to stop funding. It's easier to find $25,000 for another experiment than to reallocate $250,000 from traditional programs to AI transformation.
Recognizing Your Cannonball Moment
So how do you know when it's time to stop experimenting and start transforming? The signals are often subtle but consistent. When the same AI solution keeps proving valuable across different departments, that's a signal. When your experiments consistently save significant time or improve member satisfaction, that's a signal. When staff who were skeptical become advocates after using the tools, that's a signal. When your board asks about AI strategy and you have a long list of successful experiments but no unified vision, that's definitely a signal.
The truth is, most associations are standing at their cannonball moment right now. You're not too early—the bullets have validated the approach. You're not too late—the opportunity is still massive. You're right in the thick of the decision point. The additional experiments aren't adding new information—they're just delaying the choice to commit. We tell ourselves we're being prudent, gathering more data, ensuring success. But what we're really doing is avoiding the discomfort of large-scale change.
What Loading the Cannonball Actually Means
Moving from bullets to cannonball isn't just about scaling up existing tools, though that may be part of it. It's about fundamentally reimagining how your association creates and delivers value using AI as a core capability rather than an add-on. This means hard choices and real commitments.
First, it means stopping some things. That AMS upgrade you've been planning? If AI tools could deliver better member experience for a fraction of the cost, should you proceed? Those traditional programs that consume huge resources but deliver marginal value? Maybe it's time to sunset them and redirect resources to AI-powered alternatives that members actually use and love.
It also means changing how you think about risk. Right now, you might see scaling AI as risky and continuing experiments as safe. But consider the opposite view: In a world where member expectations are shaped by AI experiences everywhere else in their lives, isn't staying in experiment mode the real risk? While you're perfecting your bullets, members are wondering why their professional association feels technologically behind their coffee shop.
Loading the cannonball means setting a date by which AI tools will be available to all members, not just test groups. It means budgeting for infrastructure and support at scale. It means training all staff, not just the early adopters. It means telling members that this is how you operate now, not what you're testing for the future. It means taking those custom GPTs that three departments love and making them standard across the organization.
>> Related: If you're looking for a concrete exercise to help you reimagine how you deliver value to your members, check out our recent blog on building your own competitor.
From Calibration to Commitment
Jim Collins found that great companies fire their cannonballs after their bullets confirm the target. Good companies keep firing bullets forever, perfecting their aim while missing their opportunity. The question for your association is simple: Which kind of organization will you be?
Your AI experiments have given you something precious: proof of what works. You've validated the technology, demonstrated the value, and shown the ROI. The bullets have done their job. Continuing to fire more bullets may not be necessary.
The cannonball moment isn't when you have perfect information or zero risk. It's when you have enough validation to act and the courage to do so. For most associations reading this, that moment is now. Your successful experiments are calling for transformation. Your members are ready. Your staff have proven they can adapt.
Time to load the cannonball.
