What AI Actually Changes in comms: Operations, Not Creativity

Part 5 of 6: The Structural Crisis in Public Relations

The PR Industry's AI Moment Is About Operations, Not Creativity

There are two conversations happening about AI in public relations right now. One is loud, speculative, and mostly wrong. The other is quiet, operational, and actually consequential.

The loud conversation is about whether AI will replace PR professionals. Whether it will write better press releases. Whether "AI-generated content" will flood the zone and make human communicators obsolete. It's the kind of existential hand-wringing that makes for good conference panels and bad strategy.

The quiet conversation is about something far more important: whether AI can fix the structural problem described in Part 4. The agency model's built-in ceiling, where human judgment is consumed by coordination work that doesn't create value. Not whether AI can do the creative work better, but whether it can absorb the operational burden that prevents experts from doing the work clients actually pay for.

That distinction changes everything.

What generative AI actually did

First, the part most people get wrong.

Generative AI made writing, research, and synthesis instantly accessible. It raised the baseline of what "acceptable" looks like. Tasks that once required specialized expertise now appear trivial: draft a pitch, write a press release, summarize a market landscape, all achievable in minutes. This has a demystifying effect. Work that once justified fees now feels inexpensive. For entry-level or undifferentiated PR services, this is deeply disruptive.

But the effect is polarizing, not equalizing.

AI can make mediocre work passable. It cannot make unskilled practitioners exceptional. It cannot turn surface-level understanding into mastery. The gap between baseline competence and real expertise becomes clearer, not smaller. If AI can generate something that looks professional, then looking professional is no longer the differentiator. What differentiates is the judgment behind the work: knowing what matters, what to prioritize, when to act, and when restraint is the better move.

AI compresses the middle. It exposes which parts of PR were never scarce, and which still are.

The hardest parts of communications were never about producing words. They were about knowing which words to produce, for whom, in what context, at what moment. That judgment is not being replaced. It's becoming more visible as the commodity layers get automated away.

The messy reality of AI adoption in PR

Here's the uncomfortable truth about where the industry actually stands: 91% of PR professionals now use generative AI. But 50% cite "AI and automation" as a top challenge, according to Cision's 2026 Inside PR survey.

Read those two numbers together. Nearly everyone is using it. Half the industry considers it a problem. That's not resistance. It's fragmentation.

Junior practitioners use AI to keep pace with volume and speed. Senior leaders recognize its importance but lack a practical way to deploy it across the organization with consistency and control. The result is isolated usage at the edges (individual practitioners bolting ChatGPT onto their workflow) with little impact on the core operating model.

Most agencies face a practical constraint. There are few AI systems designed specifically around PR workflows. Firms assemble stacks of generic tools, each strong at a narrow task, none aware of the broader context in which communications work unfolds. Context must be reloaded between tools. Insights don't travel cleanly. Judgment remains siloed. What appears to be adoption often adds complexity rather than reducing it.

This is why AI can feel simultaneously indispensable and underwhelming: it accelerates individual tasks but leaves the structural inefficiencies of the operating model completely untouched. You can write a pitch draft in thirty seconds. You still spend three hours managing the coordination, approvals, context assembly, and reporting rituals around it.

Tools versus infrastructure

This is where the distinction between tools and infrastructure becomes decisive.

Tools optimize fragments of work. Infrastructure reshapes how work moves.

A better media database is a tool. An AI writing assistant is a tool. A sentiment analysis dashboard is a tool. Each one reduces effort locally (at one specific step in the workflow) but leaves coherence as a human responsibility. Someone still has to stitch the pieces together. Someone still has to carry context from one step to the next. Someone still has to coordinate the handoffs, manage the pipeline, maintain the institutional knowledge, and ensure nothing falls between the cracks.

That "someone" is currently a room full of people whose time would be better spent on judgment, strategy, and client relationships: the work that actually creates value.

Infrastructure operates differently. Infrastructure understands processes end-to-end and moves work forward unless judgment is truly required. It maintains context continuously rather than requiring it to be reloaded at every step. It handles onboarding, workflow progression, internal handoffs, pipeline management, and reporting mechanics: the coordination burden that Part 4 identified as the source of the agency model's structural ceiling.

AI's role in PR is not to generate better copy or automate creativity. It is to absorb the coordination burden that prevents experts from doing the work clients actually value. By standardizing and automating the non-judgment layers (context continuity, workflow progression, internal mechanics) AI infrastructure allows human teams to operate at the level of awareness and responsiveness the environment now demands.

This cannot be achieved through disconnected platforms. Monitoring tools, CRMs, and analytics dashboards reduce effort locally but leave integration as a human cost. What's required is AI that understands agency processes end-to-end and moves work forward autonomously unless human judgment is truly needed.

Redrawing the boundary

The demands now placed on public relations exceed what traditional agency structures were built to sustain.

PR is expected to operate with the awareness, speed, and accountability of an internal function: continuously tracking discourse, interpreting weak signals, responding in real time, and connecting communications activity to business outcomes. These are no longer exceptional expectations. They are baseline. The challenge is not capability; PR professionals are fully capable of meeting these demands. The challenge is structural: the operating model was never designed to support this level of persistence, velocity, and contextual continuity through human effort alone.

At the heart of the problem is a widening expectation gap. Clients want internal-level performance without internal-level headcount. They want judgment without administrative drag. No amount of incremental staffing or process refinement can close that gap sustainably. What must change is not execution, but how human effort is allocated.

Public relations is not being replaced by machines. But it is being forced to confront a reality long deferred: judgment is scarce, expensive, and valuable, and it should not be consumed by work that exists only to keep the machine running.

The firms that adapt will not be those that adopt AI most aggressively. They will be those that use it deliberately, to redraw the boundary between what humans must do and what they never should have been doing in the first place.

The diagnosis is complete. The structural forces are clear. The question that remains is the one that matters most: what does the model that comes next actually look like?