TUESDAY, MARCH 25, 2026

THE LEDE

Gartner says double PR budgets by 2027. One researcher says the data doesn't exist.

A Gartner prediction is dividing the PR industry. In its "Top Predictions to Inform 2026 Comms Strategies," Gartner stated: "By 2027, mass adoption of public LLMs as a replacement for traditional search will drive a 2x increase in PR and earned media budgets." The claim is everywhere -- keynotes, agency pitches, LinkedIn threads. The numbers sound impressive: more than 95% of links cited in AI-generated answers are nonpaid mentions. AI visitors convert at 4.4x the rate of traditional organic search visitors. ChatGPT traffic grew 608% year-over-year.

Then came the pushback. PR strategist Stephen Waddington published a direct challenge on March 16. His findings: the prediction emerged from a Gartner lead-gen blog post for a webinar -- no methodology, no data disclosure. "The prediction has escaped its marketing context," he wrote, "and now appears repeatedly as established fact." Frank McFarland at Sword and the Script added nuance in a 5-part analysis: PR spends 2.9% of budgets on measurement versus marketing's 8%. Any budget doubling without measurement reform is a wasted opportunity.

The comms angle: this is the most-cited data point being used to sell AI PR services right now. Your clients may have already heard it. The sharper take: yes, earned media visibility in AI matters -- but the industry should not oversell an unverified prediction. If LLMs cite fresh content at the highest rates, velocity and recency of coverage matter more than cumulative volume.

THE ROUNDUP

Intelligence

UK Government releases AI copyright report -- no concrete proposals yet

On March 18, the UK Government published its long-awaited Report on Copyright and Artificial Intelligence. The headline: no firm reform yet. The government says there is "no consensus" on how to meet its objectives. Meanwhile, the House of Lords Communications Committee released its own report calling for a "licensing-first" approach, including mandatory labeling of AI-generated content and transparency obligations around training data. The EU's second draft Code of Practice on AI content labeling was also published March 5, expected to be finalized by June.

Why it matters: The UK and EU are moving -- slowly -- toward mandatory AI content disclosure. Communications teams generating AI-assisted content should start building disclosure practices now, before regulations force an awkward retrofit.

Cision "Inside PR 2026" report: 91% using GenAI, but a leadership gap is widening

Cision released "Inside PR 2026," drawing on nearly 600 PR professionals across the US and UK. Key findings: 91% of professionals use generative AI as part of their workflow. 73% apply it to idea generation; 68% to writing and content refinement. Storytelling was named the most in-demand skill for 2026 (59%). But a perception gap emerged: executives see their teams as highly agile, while staff report rigid hierarchies, limited data access, and operational barriers.

Why it matters: The gap between AI adoption rates (very high) and strategic integration (still fragmented) is the real story. Everyone's using AI; few have built the systems to use it well.

From the Field

AI misuse can backfire on clients' PR goals

I've been testing Claude's context window for something most teams would never try: loading an entire quarter's worth of client coverage -- 40 to 50 articles -- into a single session and asking it to identify the three narratives that actually moved based on coverage patterns. Not what we pitched. What landed. The output isn't perfect. But it surfaces patterns you cannot see when you're reviewing clips one at a time. One client had been described as a "challenger" in every top-tier piece, not the "market leader" we'd been positioning them as. We'd been pitching the wrong frame for three months.

Try it: Export your last 30 days of client coverage as plain text. Ask Claude: "What three narratives does this coverage suggest this company actually owns -- based on how media is describing them, not how they describe themselves?" Compare the answer to your messaging doc. The gap is your actual communications problem.

Ask AI what journalists will skip, not what to improve

I started asking AI to flag which parts of a press release a journalist is most likely to skip, rather than asking it to improve the writing. The results are different and more useful. When you ask for improvements, you get better prose. When you ask what gets skipped, you get structure feedback: the boilerplate quote in paragraph three, the third consecutive sentence starting with "The company," the stat buried after the call to action. Those are the real problems.

Try it: Paste your next release into Claude and ask: "You are a journalist receiving this release in a busy inbox. Which sections will you skip, and why?" Cut or restructure whatever it flags before you send.

THE WORKFLOW

How to Use AI to Draft a Crisis Holding Statement in Under 30 Minutes

A crisis breaks at 4:47pm on a Friday. By 5:15, your CEO's phone is ringing. You have no statement. You have incomplete facts. And your communications counsel is on a plane.

This workflow produces a defensible first holding statement using AI -- fast, without fabricating facts, and without handing a journalist something that will embarrass you.

Time: 25 to 30 minutes | Tools: Claude or ChatGPT, a browser, existing messaging docs

Step 1: Lock the facts you actually have (5 min)

Before you touch AI, write down what you know for certain: what happened (confirmed), when it happened, who is affected, what action is being taken, and what you do NOT yet know. This list is your input. If you skip this step, you'll get a confident-sounding draft full of things you can't confirm. That's worse than silence.

Step 2: Feed the facts to Claude with a structured prompt (10 min)

Use this prompt: "You are a crisis communications specialist. Draft a holding statement based only on the following confirmed facts. Do not infer, assume, or add anything not in this list. [paste facts] The statement should: acknowledge the situation without admitting fault not yet established, express concern for those affected, state what action is underway, avoid speculation about cause, scale, or timeline, be 75 to 100 words, use plain English."

Step 3: Run the four-question test (5 min)

Before you show it to anyone: (1) Does it say anything you cannot confirm right now? Delete it. (2) Does it close off any door you may need open later? Soften it. (3) Would it make a reasonable person angrier? Rewrite the opening. (4) Is there a clear next step? If no, add one.

Step 4: Generate the internal version (5 min)

Run a second prompt for a 150-word internal employee communication. Tone: calm, factual, direct. Acknowledge what is happening and what leadership is doing. Direct employees to communications team for any press inquiries. You now have two documents: external holding statement and internal employee notice.

Step 5: What AI cannot do for you

AI will not tell you whether to say something. That is a judgment call. It also will not write the second statement -- the one you release once you have full facts. Use this workflow to stop the clock on the "we have no response" window.

Failure modes to watch:

Letting AI add facts to fill gaps (it will do this confidently and incorrectly). Using the draft without legal review on anything involving litigation risk. Sending the external statement before the internal one -- your employees should never read it on the news first.

TOOL SPOTLIGHT

Ruder Finn rf.Voices

Ruder Finn launched rf.Voices on March 9 -- an integrated influence platform designed for brands being discovered through AI systems. It combines AI-powered influencer marketing, customer advocacy, and paid amplification under a single framework. Core capabilities: audience intelligence via behavioral modeling, AI-assisted influence mapping across micro/nano/macro influencers plus employees and partners, precision amplification through targeted paid activations, and risk intelligence monitoring.

Who it's for:

Brands that want to build human-credible authority in an era where AI-generated content is ubiquitous and authentic voices carry disproportionate weight in AI discovery.

Pricing:

Not publicly disclosed. Enterprise positioning.

Limitation:

The announcement emphasizes moving beyond "vanity metrics" to "behavioral and reputational signals" -- but no independent benchmarks or case studies have been published yet. The tool's claim to AI-era influence optimization remains unverified by third parties.

THE ANGLE

The 91% Problem

Cision's Inside PR 2026 report found that 91% of PR professionals now use generative AI in their workflow. That sounds like a success story. It isn't.

Using AI is not the same as using it strategically. Most of that 91% is using AI the way a mid-level associate used to use a thesaurus -- to make sentences sound better. They're not changing how they measure, how they plan, or how they present their value to leadership.

The quiet crisis is in the 91%. Across organizations, the pattern is the same: prompts with no strategy behind them, AI outputs treated as research, and measurement frameworks that were already broken now buried under a layer of AI-generated activity metrics.

PR Daily called it this month: the most dangerous AI gap in communications isn't technical. It's operational. Teams know how to generate content with AI. They don't know what to stop doing.

Until the profession figures out what AI replaces -- not just what it supplements -- the 91% number will look impressive and mean almost nothing.

PROMPT LAB

The Client Briefing Generator

Before every client call, someone needs to synthesize everything -- recent coverage, competitive moves, pending messages, stakeholder concerns -- into a single document the client will actually read. Most teams do this manually. It takes two hours. Here's the shortcut.

The prompt:

You are a senior communications strategist preparing a pre-meeting client brief. Using only the information I provide, create a structured briefing document for [client name] covering:

1. Recent earned media: key headlines, tone, and what they signal
2. Competitive activity: announcements or coverage from [list 2-3 competitors]
3. Pending messages: topics we need to address proactively
4. Risk flags: developing stories or sensitivities to monitor
5. Recommended talking points: 3 to 5 bullets for the call

Source material: [paste recent coverage or monitoring report]

Keep each section to 3 to 4 bullets. Flag anything uncertain with "[VERIFY]".

Why it works:

The [VERIFY] flag discipline matters. It forces the AI to surface its uncertainty rather than paper over gaps with confident-sounding filler. You review it faster because you know exactly where to focus.

Before:

A 90-minute manual synthesis that misses half the recent coverage because you're copy-pasting from three different monitoring tools.

After:

A structured 4-section brief in 8 minutes that you edit rather than write from scratch -- and a consistent format clients start to expect and rely on.

THE READ

What the industry is reading this week

Harvard Business Review, March 9, 2026

John Winsor argues that as generative AI makes it effortless to sound authoritative, the gap between those who talk about the future and those who actually build it has become a credibility crisis. What creates real progress: hands-on experimentation, messy real-world conditions, unvarnished results. Essential reading for any comms team trying to stand out with client content.

AMEC, March 18, 2026

Kelly Costello of Meltwater makes the case that having more data has not made PR measurement better -- it's made the gap between activity and business impact more visible. Essential reading if you're being asked to prove ROI this quarter and don't yet have a framework for doing it.

TRENDING IN THE FIELD

The Gartner debate is playing out in public:

"94% of AI citations come from non-paid sources. Gartner is telling CMOs to double their PR budgets by 2027."

-- Jan-Willem Bobbink on X (widely shared, drove multiple blog responses)

"The prediction has escaped its marketing context and now appears repeatedly as established fact."

-- Stephen Waddington on Substack (picked up by Daily PR Brief, Sword and the Script)

"PR spends 2.9% of budgets on measurement vs. marketing's 8%. The measurement gap is the buried lead."

-- Frank McFarland, Sword and the Script (featured in multiple PR newsletters)

You're receiving this because you subscribed to The AI Comms Desk.
Tuesday, March 25, 2026