Good Transformer
AI for PR and comms agencies: drafting fast, checking faster
Three in four PR professionals already use generative AI. Where it speeds drafting, monitoring and research, and why, in PR, disclosure and a human check matter more, not less.
If you run a PR or comms agency and want to know where AI fits, start from the fact that three in four PR professionals already use generative AI, mostly for drafting, research and monitoring. Use it there, and keep a person on the accuracy and the disclosure. In PR, trust is the product you sell.
The pull is obvious and the risk is specific. AI writes a passable first draft in seconds, but it also invents facts, and the same technology fakes quotes and images. In a business built on being believed, that is a sharper problem than in most sectors.
Where AI speeds PR
The gains cluster in the research-heavy and draft-heavy parts of the work. Muck Rack's State of AI in PR 2026, a survey of more than 500 PR professionals, found 76% now use generative AI, up from 75% the year before and levelling off as the tools become standard. Of those who use it, 82% say it improves the quality of their work and 93% say it speeds it up. That is where to point it: drafting, monitoring, research and reporting.
Drafting. First drafts of releases, backgrounders, social posts and pitch emails come faster with AI in the loop. It removes the blank-page cost. It does not remove the need for a person to shape the message and check every claim.
Media monitoring. Scanning coverage, clustering mentions and flagging what shifted overnight is exactly the kind of volume work AI handles well. Treat its summary as a first read to verify, not a report to forward.
Research. Pulling together a first view of a sector, a journalist's beat, or a stakeholder map is quicker with AI. The facts still need checking against the primary source before they inform a plan.
Reporting. Turning campaign data into a first-cut summary saves time on the monthly report. A person still decides what the numbers mean.
Where it is dangerous
Here is the honest part. The same speed that helps you draft is the speed that ships a mistake to a client or a journalist.
Accuracy. AI states wrong facts with total confidence. In PR a single invented statistic or misattributed quote in a release is a reputational hit for the client and the agency. Every fact an AI tool produces needs verifying before it leaves the building. Our note on how to stop AI mistakes reaching clients sets out a workflow for exactly this.
Disclosure. PR has a professional code, and honesty is the core of it. Members of the CIPR are bound by a Code of Conduct that turns on integrity, competence, transparency and confidentiality. Passing off AI-generated material as human work, or presenting AI-fabricated endorsement as real, cuts straight against it. Be transparent about how the agency uses AI, and never fake what looks like an independent voice.
Deepfakes and disinformation. The technology that drafts your copy also fabricates convincing quotes, audio and images. That is a threat to your clients' reputations and to public trust, and it raises the value of a profession that can be trusted to tell the truth. Handle it as a risk to plan for, not a tool to use.
The trust premium
There is an opportunity hidden in the risk. When anyone can generate a plausible press release, a plausible press release is worth less. What holds its value is the judgement, the relationship and the credibility a machine cannot fake.
The CIPR's own members see this. Its State of the Profession 2024 named AI the biggest skills shortage in the profession, at 36%, and the top challenge facing the industry, at 38%. The agencies that treat that as a reason to build judgement, not just speed, are the ones clients will keep. Your understanding of a client's world is the part a general model will never hold. It is the same reason your brand is read by machines matters: the machines draft, but the meaning is yours to protect.
A responsible starter workflow
You do not need a programme to begin. You need one bounded task and a firm rule about checking.
- Pick one low-risk, repetitive task. First-draft social posts or media-monitoring summaries are the natural first step, because they save real time and are easy to check.
- Use a business-grade account. One with a data agreement that keeps client and embargoed information out of model training, not a free consumer login.
- Verify every fact before it leaves. Names, figures, quotes and dates get checked against the primary source, every time, by a person.
- Disclose plainly. Agree with each client how the agency uses AI, and be able to say what a tool touched and what a person checked.
- Never fake a voice. No AI-generated testimonial, fake expert or synthetic endorsement presented as real. Ever.
The skills gap and how to close it
The 36% skills-shortage figure is the real story for agency owners. The bottleneck is not access to tools, it is people who can use them with judgement. And the guardrails lag the tools: the same Muck Rack 2026 survey found only 51% of PR professionals work somewhere with an AI use policy, though that has more than doubled since 2024. Closing both gaps is a management task, not a software purchase.
Start by naming who in the agency is your AI champion. Give them time to build a simple playbook of what the team may and may not do, and run short, real sessions on live work rather than generic training. Confidence with AI comes from using it on the actual job, with a person who knows the guardrails standing beside it.
What to do next
Pick one recurring task your team does every week that is easy to check, such as first-draft social posts or a weekly monitoring summary. Run it through a business-grade AI tool for a fortnight, with a person verifying every fact before anything ships, and write down the time saved against any corrections. That bounded test teaches you where AI helps your agency, without putting a client's trust at risk.
If it would help to map where AI fits across your agency, and where a person must stay in charge, book a call and we will work it through with you. If you are weighing outside help first, our note on how to choose an AI consultant is a good place to start.
Common questions
Should PR agencies use AI?
Yes, on the inputs. AI is well suited to first drafts, media monitoring, research and reporting, where it saves real time. It is a poor fit for anything that ships without checking, because it states wrong facts confidently, and for anything that fakes an independent voice, which cuts against the profession's code. Use it to prepare the work, and keep a person on the accuracy and the disclosure.
Do PR professionals have to disclose their use of AI?
Transparency and honesty are core principles of the CIPR Code of Conduct, so passing AI-generated material off as human work, or faking endorsement, runs against professional standards. The safe default is to agree with each client how the agency uses AI and to be able to say plainly what a tool touched and what a person checked. Disclosure protects the trust the agency is selling.
What is the biggest AI risk for PR agencies?
Accuracy. AI invents facts, quotes and statistics and presents them with confidence, and a single wrong fact in a release damages both the client and the agency. Close behind it are faked content, deepfakes and synthetic endorsements, which threaten client reputation and public trust. Both are managed by verifying every fact against its source and never presenting fabricated material as real.
This is general information, not legal or professional-conduct advice. Check your own obligations with a qualified professional or with the CIPR.