What the Agency AI Survey results mean for PR and marketing firms

The SAGA Agency AI Survey results are in, and small agency owners are feeling great about AI. Maybe too great. In this episode, Chip and Gini dig into the numbers and find the gap between how owners think they're using AI and the reality of what's happening inside their businesses. The headline figures look impressive: 89% of respondents report regular or widespread AI use, 74% use it daily, and 88% say they've seen productivity gains. But Chip isn't buying it. He questions whether the sample skews toward early adopters, or more likely, whether agency owners simply don't have a clear enough picture of what "good" AI use looks like elsewhere. When 53% say they're ahead of their peers but only 13% say they're behind, the math doesn't work. As Gini puts it, they're probably grading themselves on usage habits, not operational depth. Next, Chip and Gini look at what agencies are actually doing with AI. Most activity falls squarely into what Chip calls "generative AI 101" — drafting emails, writing social posts, generating blog content. The more interesting stuff is largely absent. AI-assisted design work barely registers. Only 74% are even using AI to revise or edit content, a number both hosts find inexplicable given how easy and useful that is. Gini's own example of running an article through an AP style agent before sending it to a notoriously precise editor at PR Daily illustrates exactly the kind of practical, low-friction habit that should be universal by now. Another data point they discuss is the disconnect between productivity gains and revenue. Agencies report getting faster, but their top-line numbers are flat or down. Gini's read is that AI efficiency is getting absorbed into existing scope rather than converted into new value. Agencies are over-servicing clients at the same fees, filling freed time with more of the same work instead of building something new. On the pricing side, almost no one reported clients pushing for discounts tied to AI use. Instead of a reduction in cost, the larger enterprise clients are asking about data governance, usage policies, and procurement compliance. Chip advises unless your agency has the infrastructure to manage those requirements consistently, that's a market best left to someone else.