Three Steps for Becoming AI Ready

A report in February found that nonprofits have been much slower than their corporate counterparts when it comes to adoption of generative AI. Jamie Notter, founder and CEO of the consultancy Propel, which works with associations, isn’t surprised.

“AI is going to be so iterative—we do not know what direction this is going to go in until we actually start playing around with it,” he said. “And that’s not in the wheelhouse of associations—trying it and learning and adjusting. They want to spend 18 weeks building the perfect thing and then release it.”

Part of the sluggishness in associations around AI, Notter said, is the siloed nature of many of them. He raised a hypothetical example of an association where multiple departments begin to develop AI-based apps, only to hit a roadblock. “They realize they need content from the journal, so they go to the journal and the journal people freak out,” he said. “‘Then no one will subscribe to our journal!’ So everybody wasted ten weeks developing something.”

Getting ahead when it comes to AI, then, means resolving some cultural disconnects within your organization. Notter made three general suggestions about how to do that. 

1. Bring in partners. Instead of treating AI like a skunkworks project that needs to be developed within a particular department, associations should leverage the goodwill and enthusiasm of membership to help with brainstorming and development. “One place where associations have a leg up on others is the idea of co-creation,” he said. “If you want to be agile, you’ve got to be good at creating something with someone, and that is completely in an association’s DNA. You’re co-creating with volunteers. So bring in users to show them what you have, get feedback, and incorporate it on a regular basis.”

2. Develop internal transparency. Silos occur when organizations fail to communicate their efforts across departments. This often happens because some stakeholders want to avoid talking about failures and missteps, Notter says. But that attitude is itself is a misstep. “Organizations struggle with this because they don’t want to look bad—they want to control information,” he said. “But we have to build systems where everyone can see what everyone else is working on, so you don’t have three different silos developing apps at the same time. You can have a Slack channel about AI where people say, ‘Hey, I’m going in this direction.’ Putting stuff out there where you don’t know who needs to see it. We still haven’t built that habit well enough.”
3. Experiment aggressively. Think of AI as an efficiency project, Notter says—think of the kinds of ways you might be able to serve members better if you had more time and resources, and design AI pilots that can help clear space for them. But doing that requires taking AI tools out for a spin, sometimes uncertainly at first. So spend less time standing up ad hoc committees or task forces and get to work. “We need to get more pilots started right now, and that’s not a committee-led thing,” he said. “Sometimes we don’t know what we’re going to develop as we’re developing it. You might wind up something that needs to be managed and that you’ll have to build some structures around. But at this stage, I would rather that more activity be done.”

[gorodenkoff/IStock]