Associations are buying the same AI tools, but the ones seeing returns started by looking at how their people work.
This technology cycle is different from ones before it. Previously, tools got buy-in at the top and were then pushed down into the workforce. With AI, employees are already experimenting on their own and the ideas are bubbling up. The challenge for leaders is to corral staff usage and funnel those ideas so the organization is aligned and learning together.
At one 900-person organization we work with, leadership identified 20 people they thought would use AI once training rolled out. At the same time, we did a technology assessment and talked to about 100 people. Ninety-five of them were already using an AI tool the company hadn't approved for their work.
Most of these AI users thought they were being careful, and some leaders thought the same. One said she was comfortable with her team’s ad hoc usage because they were using paid versions. And while there was a time when paying $20 a month kept your data out of the model, that isn’t true anymore. Every one of those licenses was for a consumer plan with training turned on by default—meaning the platform integrated all of the organization’s information into its AI knowledge base.
One association we work with uses AI to understand member profiles based on role, email clicks, and site searches, so members see relevant content when they log in to the portal. That idea came out of a training. In one of the early sessions, someone asked whether AI could help personalize the website. The answer was yes, but not until the team had worked through the basics. Those ideas kept stacking up, and now that association has a roadmap of AI projects.
If you can capture what your staff is already doing and surface it to leadership, that's the right way to start. Assess which tools employees are using and why, then use the organization’s AI policy to define an approval process based on business fit, necessity, security, and data use. An AI committee can maintain the approved tool list as a matrix, updated quarterly, which shows each tool’s permitted uses (including whether it can handle confidential data) and notes the tools the association has rejected. Once a tool is approved for general use, train staff in person by department, using current work, and follow that training with drop-in office hours.
Eventually staff ideas will outrun your approved tool list. That's the point. Each time you’ll need to vet each tool, and it comes down to four questions: what data it touches, whether it trains on your information, who signs off, and how you'll judge it a year from now.
A CFO always asks whether they're getting value for all these licenses. It’s a fair question, but hard to answer. Using AI for the sake of using AI won’t change how anyone works. Some large organizations now measure how much AI each person used in a week, which almost incentivizes people to overuse it so they show up on a leaderboard.
What to look for instead is the volume of ideas going up. Most of our clients form an AI committee after the initial training to keep evaluating new ideas, and the measure is whether the requests to that committee keep growing. Staff start to ask whether AI could handle a task they do manually or suggest a way to improve the member experience based on something they tried. You can't measure that on a dashboard, but when it's working, you can feel it.
When people hear AI and member experience in the same sentence, they picture a bot where a human used to be and members arguing with it to get an answer. The personalized member portal described earlier is the counter example: AI that enhances the experience instead of degrading it or replacing a person.
No one in the association world is asking us how to use AI to cut staffing. These teams are already lean, and they want AI to free up hours in the day and get to the things no one ever has time to do. For associations with a growing membership, AI can help existing staff serve more members, so back-office hiring does not have to rise at the same rate. An association projecting 10 percent membership growth may need to add one or two employees instead of three. In associations with a stable membership, AI helps staff respond faster and deliver more value, which can make it easier to justify higher dues at renewal.
The other thing worth pushing back on is the idea that AI is the only thing on your technology roadmap right now. Data strategy, security, the systems running your membership and your finances all still need attention and budget. None of that became less important this year.
The associations that helped their people become AI fluent first are now asking what they could do for members with the hours they got back. More than one has asked us why they waited three years longer than they had to.
