Mostly for editing, research and drafting, almost never for the parts of PR that actually earn coverage. Muck Rack's State of AI in PR 2026 report, based on a survey of 564 PR professionals fielded December 5 to 24, 2025, found that 86% of PR pros who use generative AI use it for editing and refinement, and 76% for research, but only 30% use it for media list building and outreach, the part of the job that actually puts a pitch in front of a journalist. That gap between where PR teams let AI in and where they keep it out is the real story in this year's numbers, more than the headline adoption rate itself.
Source: Muck Rack's State of AI in PR 2026 report, linked in the article.
How many PR professionals are actually using AI now?
Adoption has stopped climbing. Muck Rack's survey found 76% of PR professionals use generative AI in their workflow, which the report describes as having "peaked," essentially flat against 75% the year before. Cision's separately conducted Inside PR 2026 report, based on nearly 600 PR professionals across the US and UK and also released in January 2026, found a higher figure, 91%. The two surveys use different populations and question wording, so the gap between 76% and 91% is not a contradiction so much as a reminder that "AI adoption" depends heavily on how a survey asks the question. Read together, the honest range is that somewhere between three-quarters and nine in ten working PR professionals now use AI in some part of their job, and that range has stabilized rather than kept growing. The 7% of holdouts in Muck Rack's survey mostly cite ethical or environmental concerns, not a lack of access, and 80% of them say there is nothing short-term that would change their mind.
What are PR teams actually using it for?
Writing-adjacent tasks, overwhelmingly. Muck Rack's task breakdown, in order: editing and refinement at 86%, research and insights at 76%, writing and content creation at 74%, strategy and planning at 68%, administrative or operational tasks at 51%, media list building and outreach at 30%, and measurement and reporting plus creative asset production tied at 19% each. Cision's figures point the same direction: idea generation at 73%, writing and content refinement at 68%, AI-driven media monitoring at 40%. Both surveys land on the same shape, PR professionals use AI heavily to speed up writing and research, and use it far less for the relationship-dependent parts of the job: pitching, list building, and judging what will actually land with a specific reporter.
Where PR pros say AI genuinely saves them time tracks the same pattern. Muck Rack found 59% save the most time on editing and refinement, 53% on research, 46% on writing. Only 8% cite media list building and outreach as where AI saves meaningful time, which lines up with how little that task has been handed to AI in the first place.
Does PR actually trust the output, or just use it anyway?
Mostly trust it, with real limits. Muck Rack's survey found 82% say AI has made the quality of their work somewhat or much better, and 93% say it helps them complete projects more quickly, with 49% saying "much more quickly." But that trust is not unconditional. Editing habits are loosening: in 2024, 61% of PR pros said they had to make edits to most of the AI-generated text they used. By 2026 that fell to 45%, while the share who only touch a few paragraphs rose from 30% to 44%. PR professionals are reviewing AI output less thoroughly than they were two years ago, even as the same survey shows 98% still say they always or often edit it at all. Nobody has stopped checking the work. People are checking less of it.
There are clear lines PR pros will not cross. 77% say they avoid entering financial data into an AI system, 69% avoid personally identifiable information, and another 69% avoid proprietary or strategic materials. Fewer draw the same line around client or brand names, 36%, or internal communications, 33%, which suggests the caution is about data sensitivity specifically, not AI use in general.
What do PR teams think the risks actually are?
The number one worry is not about accuracy. 77% of Muck Rack's respondents said their biggest concern is that newer PR professionals will not learn the fundamentals of the profession because they rely too heavily on AI tools. That outranks 63% who worry about unscrutinized AI output lowering the quality of their work, and 61% who worry clients will decide they no longer need human content creators at all. Those figures are essentially unchanged from the prior year's survey, which suggests PR as an industry has settled into a stable, if uneasy, view of where the real risk sits: not in the tool itself, but in what newer professionals stop learning because the tool is available.
Agentic AI, systems that can take PR actions with minimal human input, has not caught on at all. Only 12% of AI-using PR pros have adopted AI agents, even though 59% say they have heard of them. And 90% say they are not comfortable letting an AI agent send messages or publish updates without a human reviewing it first, with 59% saying they are "not at all comfortable" with that. Earned coverage still runs on a human deciding what is worth pitching, and PR professionals, by their own numbers, are not ready to hand that decision to software.
What does this mean if you are hiring a PR agency?
That an agency using AI tools well is now closer to table stakes than a differentiator, and the question worth asking is which parts of the work they have automated. An agency that uses AI to speed up research, drafting and editing, the same 86%, 76% and 74% of tasks most of the industry already has, is simply working at the pace the rest of the field has settled into. An agency claiming AI has automated its pitching, its media list building, or its judgment about which journalist to approach is claiming something only 30% of the industry has even attempted, and something 90% of PR professionals say they would not trust unsupervised. Choosing a PR agency in 2026 means asking what AI touches and what a person still owns, not whether AI is involved at all. By this point, it almost always is.
The skills gap data backs this up from the inside. 85% of PR professionals say prompt writing is a skill the industry now needs, and 71% say ethical decision-making is too, well ahead of AI tool evaluation at 55% or data literacy at 54%. The industry itself is telling you that judgment, not tool access, is what separates agencies now.
The direct answer, for anyone who read this far
PR teams in 2026 use AI heavily for editing, research and first drafts, somewhere between 76% and 91% of professionals depending on which survey you read, and lightly or not at all for pitching, media list building and anything that requires a human relationship with a specific journalist. Trust in the output is high and growing less cautious, with review depth dropping even as usage stays near universal. The one place the industry draws a hard line is autonomous action: 90% will not let an AI agent publish or send anything without a person checking it first. If you are evaluating a PR partner, the useful question is not whether they use AI, almost every serious one does, it is whether they still have a human deciding what gets pitched and to whom. Talk to us if you want a straight answer on where that line sits for your own program.
