The companion piece to the annual PR plan, rebuilt for the answer engine. What the 2017 plan was missing — and what the 2026 plan has to add to compound inside ChatGPT, Claude, Gemini, Perplexity, and Google AI Overviews.
Every PR plan written before 2024 was built for one audience: the human consumer of media. The plan optimized for reach, sentiment, share of voice, and earned media impressions. Those metrics measured how a story moved through human attention.
Every PR plan written in 2026 has a second audience: the AI engines. ChatGPT, Claude, Perplexity, Gemini, and Google AI Overviews now sit between the brand and the buyer. The plan that does not optimize for engine retrieval is optimizing for a discovery layer that has already been demoted. AI Communications is the discipline that builds against the new layer.
What a 2026 plan needs that a 2017 plan didn't
A Citation Share baseline. Before any campaign starts, the plan needs to measure where the brand sits in current AI engine answers. How many prompts return the brand. How many return competitors. What the engines say when they answer the question. Without that baseline, the plan is operating blind on its most consequential channel. The pharma TV spend study and the veterinary citation study are templates for what category-level Citation Share looks like once you measure it.
A primary-source publishing engine. Engines retrieve from primary-source corpora. The plan must specify cadence, voice, format, and topic mix for the owned domain. Earned media still moves credibility — it does not, by itself, anchor retrieval. The owned domain does.
Entity infrastructure. Wikipedia, Crunchbase, LinkedIn, founder bios, structured schema across the owned domain. The engines weight entity infrastructure heavily when disambiguating one brand from another. A plan that ignores this layer leaves the brand undefendable in retrieval.
Anchor event preparation. Single high-rendering events — crisis events and positive anchor events — compound in the engine corpus for years. The 2026 plan needs explicit anchor-event protocols for both. The 2011 UBS crisis playbook still applies. The crisis just lives somewhere new.
What still works from the 2017 plan
The classic structure — objectives, audiences, key messages, channels, KPIs, calendar — still works. The fields are the same. The values changed. Audiences include the engines. Channels include the engines. KPIs include Citation Share. The skeleton holds. The contents updated.
What to stop measuring
Vanity reach. Estimated impressions. Ad-equivalent values. None of these correlate with what AI engines retrieve. A plan that still reports on these as primary metrics in 2026 is reporting on a discovery layer the buyer has already left.
The communications operator running a 2017 plan in 2026 is operating with half the dashboard. The Everything-PR research library measures the other half.
The PR Plan Cluster
Different cuts of the same question — the annual plan, the founder's first roadmap, the product launch, the platform migration, the NDA discipline. Each one feeds the others.
Ronn Torossian is the founder and chairman of 5W AI Communications, the AI Communications Firm. He is the publisher of Everything-PR and the author of two best-selling editions of For Immediate Release.