PR teams have a measurement problem, and it's not an obvious one.
Many communications professionals now use AI to draft pitches, assemble media lists and summarize coverage— 77%, according to Muck Rack's State of PR 2025 report. The tools are fast, capable of generating full drafts in seconds.
But any PR pro knows a first draft is far from usable. Generated releases and pitches still have to be edited, fact-checked and aligned with journalists before they can be used.
That gap between generation and usable output is worth thinking about. It mirrors a broader pattern across industries: More than 80% of companies saw no noticeable productivity gains despite widespread AI adoption, per a recent National Bureau of Economic Research survey.
In PR, it comes down to how the work is measured.
Are We Tracking the Right Number?
Most evaluations of AI implementations measure speed. In PR, that may involve tracking how quickly AI generates a draft or puts together media lists.
What is rarely measured is the correction cost: the time required to improve the output to the point it’s actually usable.
Consider a pitch. With proper prompting, AI can produce a serviceable first draft in under three minutes. But it will need editing for tone, checking against the reporter’s beat and revision. It also needs verification: AI tools can fabricate details, so you’ll have to fact-check claims, confirm sources and verify references.
That review can add 10 to 15 minutes, sometimes more. The total time saved is not so impressive once you account for this.
This formula clarifies the picture:
Net Productivity Gain = Time Saved on Generation – Correction Cost
Most teams track only the left side of that equation, which is why perceived efficiency and actual gains diverge significantly.
A Quick Test Any PR Team Can Run
Pick a task your team already performs with AI, like drafting a pitch, building a media list or preparing a briefing note.
First, time how long it takes to get the first output. Then time the correction phase: every edit, fact-check, and revision performed before the output is usable. Finally, compare the total with how long the task takes without AI tools.
In my experience, there’s a consistent pattern. The time taken to deliver a final pitch might drop from 25 minutes to roughly 20, and media research might fall from 45 minutes to 30.
Those are real time savings, but they're far more modest than the "AI wrote this in seconds" narrative. Moreover, when AI output requires significant revision—say, pitches to journalists covering niche beats—the correction cost can nearly erase the time gained.
Where AI Earns Its Keep
Calculating the time-cost of correction lets PR teams realize the most productive applications of AI tools, which tend to involve inputs rather than outputs.
Think: identifying journalists covering a specific narrative; uncovering new storylines that haven’t yet made a splash; or building background to inform a pitch. Such applications reduce the revision burden simply because better research produces fewer drafts.
When AI improves inputs, the correction cost on outputs shrinks. This is where productivity gains are realized.
Moving from Activity to Evidence
The productivity conversation around AI has stalled partly because we’re talking about the wrong metrics. Speed of generation is a measure of activity. What comms leaders need are outcome metrics: How much did correction time decrease? Did journalist response rates improve as inputs were refined with AI? Are time savings being redirected to strategic work?
Most PR teams already have the data to answer these questions. They just require a different perspective.
Tim Gray is a communications leader and Strategic Communications Adviser at Intelligent Relations.