CoreCMO

Channels & Execution


PR & Comms

Public Relations (PR) does three jobs at [COMPANY NAME]: earn credibility, lead thought, prepare for crisis. Plus internal comms — the underinvested fourth job.

Channels & Execution 3 prompts 1 agent — PR Comms Agent ~5 min preview

The framework — strategy first


PR & Comms — the strategic foundation.

STRATEGY & PROCESS

Public Relations (PR) at [COMPANY NAME] does three jobs — build authority, support pipeline, protect reputation — across five audiences: prospects, customers, analysts, press, and employees.

Communications Framework

PR does four jobs at [COMPANY NAME]: earn credibility (analyst relations, press coverage), lead thought (executive visibility, industry positioning), prepare for crisis (rapid-response protocols), and align employees (internal communications).

PR OPERATING PRINCIPLES

  • Earned > paid. One credible third-party validation outweighs ten paid placements.
  • Newsjacking > press releases. Attach [COMPANY NAME]'s perspective to trends already in the news cycle.
  • Analyst-first. Tier 1 analyst briefings precede any major announcement.
  • Spokesperson discipline. All media inquiries route through a single designated spokesperson or PR lead.
  • Measurement. Coverage is tracked by reach, sentiment, and tier — not just volume.

PR in the AI era — table stakes, not optional.

The audit conversation about PR Return on Investment (ROI) changed in 2024–2026 because the buyer's research path changed. Before: a buyer typed a category query into Google, scrolled past two ads, and landed on a comparison page or vendor site. Now: that same query increasingly returns an AI answer — ChatGPT, Perplexity, Google AI Overviews (the AI-generated summary block at the top of Google results), Claude — that cites third-party authority sources (Forbes, WSJ, TechCrunch, Gartner, Forrester) at far higher weight than vendor-owned pages. The sites your CFO dismisses as "vanity coverage" are the sites the AI systems trust to write your category's answer.

WHY PR MATTERS MORE IN 2026, NOT LESS

Three concrete shifts the senior operator names in front of the CFO:

  • Large Language Model (LLM) citation share-of-voice is the new SEO. When ChatGPT, Perplexity, or Google's AI Overview answers "what's the best [category] software" or "who are the leaders in [category]," it pulls from earned coverage and analyst reports. Your owned content is rarely cited; Forbes is. Brands cited frequently across earned media surfaces — analyst coverage, tier-1 press, podcast appearances — show up meaningfully more often in LLM-generated answers than brands relying on paid only. The size of the lift varies by category.
  • Analyst reports are foundational LLM training-and-citation sources. Gartner Magic Quadrants, Forrester Waves, IDC MarketScapes, and vertical analyst research (KLAS, Celent, Datos) get scraped and cited by AI tools for years after publication. Being included in the relevant report is now a 3–5-year compounding distribution asset, not a one-time sales credibility play.
  • Brand entity recognition decides whether the AI even mentions you. AI tools list "top vendors" by extracting named entities from authoritative sources. If [COMPANY NAME] isn't in the corpus the model trusts, the model doesn't list you — regardless of how good your product is, how strong your SEO is, or how many G2 reviews you've earned.

The CFO defense — replacing Earned Media Value with output metrics that survive a budget review

The reason PR gets cut first is that the function leaned on Earned Media Value (EMV) for two decades, and EMV is indefensible — it was deprecated by the industry's own measurement body (AMEC's Barcelona Principles, 2010) and CFOs know it. The senior-operator move is to retire EMV entirely and report on five output metrics instead:

METRICDEFINITIONWHY CFOs ACCEPT IT
Pipeline-influenced coverage rate% of open opportunities whose buying-committee contacts engaged with PR-driven content in the prior 90 daysTies earned media directly to active pipeline. Trackable in CRM with UTM-tagged links + first-party engagement data.
Analyst-referenced wins% of enterprise closed-won deals where the buyer cited an analyst report (Gartner MQ, Forrester Wave, vertical analyst) in win/loss interviewsDirect revenue attribution. Survives any "what's it worth" challenge because the buyer named it as a deciding input.
AI citation share-of-voiceQuarterly audit: run 20–30 category queries against ChatGPT, Perplexity, Google AI Overviews, and Claude — count branded mentions vs. named competitor mentionsTracks the actual surface buyers consult before they ever touch your site. New metric, but the math is repeatable and benchmark-able.
Branded search lift post-coverageWeek-over-week change in branded organic search volume in the 4 weeks after a Tier 1 placementTrackable in Google Search Console + Similarweb. Isolates the brand-awareness lift from a specific placement.
Tier 1 / Tier 2 placement velocityQuality-weighted count of placements per quarter, scored on outlet tier × story type × narrative fitVolume measure CFOs understand, but quality-weighted so the team isn't gaming press-release-wire counts.

Run these five in the monthly Marketing Dashboard alongside pipeline. PR stops looking like a black box; it starts looking like an asset class with measurable yield.

Earned Media Strategy

Earned media drives awareness and credibility that paid media cannot replicate. [COMPANY NAME] targets three tiers of coverage: Tier 1 (national business press, major trade publications), Tier 2 (vertical/industry publications in ICP-relevant sectors), and Tier 3 (regional business press, niche HR/L\&D/tech outlets).

THE PR STACK IN 2026 — LEGACY VS. MODERN

The PR tooling category split into two camps over the 2020s, and your stack should match motion + maturity. The legacy stack — Cision and Meltwater for media monitoring and journalist outreach — still dominates enterprise PR teams and PR agencies, with deep historical databases and analyst lists. The modern alternatives — Muck Rack (the 2026 incumbent for journalist CRM), Prowly (lighter-weight journalist outreach), Propel and Roxhill (analyst tracking), Onclusive and Brandwatch (modern media monitoring) — built around cleaner data, faster outreach workflows, and AI-assisted personalization. Most in-house teams under 50M ARR ship faster on the modern stack; PR agencies and 500M+ ARR programs typically still run Cision or Meltwater as the system of record. Pick by workflow fit, not brand prestige.

STORY TYPEHOOKTARGET OUTLETSLEAD TIME
Executive Thought LeadershipOriginal POV on industry trend or emerging issueTier 1 Business + Trade2–4 weeks
Customer StorySpecific, quantified outcome a customer achievedTier 1–2 Trade4–6 weeks
Data / Research ReportProprietary survey data with provocative findingTier 1 Business + Wire6–8 weeks (incl. survey)
Product AnnouncementCategory-advancing feature or capabilityTier 1–2 Trade + Wire3–4 weeks
Partnership / IntegrationJoint news with complementary brandTier 1–2 Trade3–4 weeks
Industry AwardThird-party validation of product or cultureTier 2–3 Trade2–3 weeks
Hiring / Executive AppointmentSenior leadership hire with strategic narrativeTier 1–2 Business1–2 weeks
Funding / MilestoneARR milestone, funding round, or major expansionTier 1 Business + Wire4–6 weeks

The newswire-plus-FAQ tactic — cheapest AEO citation play in B2B

The single highest-ROI tactical move for AEO citation share at the $5M–$50M ARR stage: get on a major newswire (PR Newswire, Business Wire) quarterly with product or customer news, and embed a three-FAQ block at the bottom of every release.

Newswires cost ~$600 per release for national distribution. LLMs treat newswire content as trusted citation source — they do not distinguish paid from earned placement. The FAQ format embedded inside the release performs disproportionately well in citation rate because LLM retrievers extract Q&A blocks cleanly.

The structure that ships: 350-word newswire release on a real product update, customer win, or funding milestone (the news), followed by a three-FAQ block answering the most common buyer prompts in your category ("What is [category]?", "What does [company] do?", "How is [company] different from [market leader]?"). The FAQs are the part LLMs cite for years; the news is the part journalists pick up.

Cadence: at least quarterly. The teams that win citation share-of-voice ship one newswire release every 6–8 weeks, alternating product news and customer wins. At ~$2,400/year for four releases, this is the cheapest paid line item in the entire AEO discipline.

Analyst Relations — must-have for enterprise, critical for strategic

Analyst Relations (AR) used to be a "nice to have" — fund it once the company hits Series C, optional below. That positioning is wrong in 2026. The shift: Gartner Magic Quadrants, Forrester Waves, IDC MarketScapes, and vertical analyst reports (KLAS for healthcare, Celent for fintech, Datos for financial services) are foundational LLM citation sources, get cited in AI answers for years after publication, and show up named in win/loss interviews at enterprise Annual Contract Value (ACV). One Tier 1 RFP that cites a Magic Quadrant in vendor requirements pays back the entire AR program for a year — and that's increasingly the default for procurement teams who use AI tools to draft their RFP criteria.

Budget framework: enterprise + strategic ACV companies should run a serious AR program at ~$120K–$250K/year (Gartner subscription + briefing prep + analyst-day investment + content development). The senior-operator move is to commit to this line item by Series B, not after; the analyst relationships take 12–18 months to mature, and a Series C founder who decides to "finally do AR" has already lost the window. A favorable analyst mention shortens enterprise sales cycles by providing independent third-party validation at the top of a buyer's research journey — and that validation now compounds through the AI inference layer in ways an earned media placement alone cannot match.

AR ACTIVITYCADENCEOBJECTIVEOWNER
Tier 1 Analyst BriefingsQuarterlyKeep analysts current on roadmap and strategy; earn mentions in research notesCEO/CMO + AR Lead
Inquiry CallsAs needed (max 5/yr per analyst)Request analyst perspective on market trends, competitor moves, or category definitionProduct/Marketing Leadership
Pre-Announcement Briefings7–14 days before newsEnsure analyst is prepared to comment on major announcementsAR Lead + CMO
Custom Research ProjectsAnnualCommission original analyst research to support demand gen (e.g., state of industry report)CMO + Finance approval
Magic Quadrant / Wave ParticipationAnnualFormal submission and demo with Gartner, Forrester, or relevant analyst firmProduct + AR Lead
Analyst DayAnnual (optional)Extended briefing covering company vision, roadmap, and customer storiesCEO + Leadership Team

Executive Communications & Thought Leadership

The CEO, CMO, and product leadership are [COMPANY NAME]'s most credible voices. A systematic program turns executive perspective into earned coverage, speaker placements, and social audience.

CHANNELFORMATCADENCEGHOSTWRITING?
LinkedIn (CEO)Long-form posts, POV articles2× / weekYes — draft + CEO approves
LinkedIn (CMO / VP Product)Industry commentary, content amplification3× / weekYes
Bylined Articles800–1200 word op-eds in trade pressMonthly (1 per exec)Yes — ghostwritten to exec voice
Speaking / Conference Panels30–60 min sessions at Tier 1 industry events4–6 events/yearPrep deck + talking points only
Podcast Appearances20–40 min interviews on target audience podcastsMonthlyPrep doc, no ghostwriting
Industry Awards (Executive)Nominations for CMO of the Year, CRO 50, etc.Annual submission cycleNomination ghostwritten

PR is earned credibility — and the asset class CFOs cut first.

PR returns are real but slow, indirect, and hard to measure on a quarterly dashboard. That makes it the line item that gets cut first when the budget tightens — and the function CMOs regret cutting most after the fact. The discipline that protects PR investment: earn one defensible Tier 1 placement per quarter, and document its downstream impact on inbound demos, candidate flow, and analyst access. One Tier 1 placement isn't a press release — it's the artifact your sales team uses for 12 months.

The three earned-media bets that move pipeline

  1. The annual benchmark report. Proprietary data, useful methodology, published in March or September (the two windows when business press is hungry for data stories). Generates 4–8 Tier 1 placements per year, becomes the most-cited content asset on the site, and gives the CEO a 12-month speaking circuit.
  2. The category-defining op-ed. The CEO publishes one signed op-ed per quarter in a Tier 1 trade publication, naming the category narrative the company is operating from. Less about coverage volume, more about ownership of the framing.
  3. The analyst briefing rhythm. Quarterly briefings with the 2–3 analysts who cover your category. Not just Gartner — the niche analysts who publish in your vertical's trade press. Gets you cited in their reports, which is the credibility multiplier sales uses in enterprise deals.

The PR-to-pipeline measurement gap — and how to close it

The hardest math in PR: tying a press placement to revenue. The pattern that works is indirect: track inbound demos in the 30 days after a Tier 1 placement, candidate applications in the next quarter, analyst report inclusions in the next 6 months, and the win-rate uplift in enterprise deals where the placement was referenced in sales calls. PR doesn't show up in last-touch attribution; it shows up everywhere else.

Your PR & Analyst Relations Program

The publications you target, the analysts you brief, the bets you're placing this year. Every PR, comms, and exec-thought-leadership prompt on the site uses these.

Saved as [TIER 1 PUBS], [ANNUAL BENCHMARK REPORT], [ANALYST BRIEFINGS], [PR AGENCY].

PR Calendar & Cadence

QUARTERMANDATORY PR ACTIVITYOPTIONAL OPPORTUNISTIC ACTIVITY
Q1Annual benchmark/research report release; tier 1 analyst briefings; one product news releaseIndustry awards nominations open; executive speaking at major Q1 conferences
Q2Customer story campaign (3 new case studies); mid-year data drop or reportAnalyst Magic Quadrant/Wave season; company culture/hiring press
Q3Major product announcement or roadmap preview; industry event newsroom at flagship conferenceExecutive bylines around fiscal Q3 trends; fall speaking season prep
Q4Year-in-review / predictions release; customer milestone announcementsHoliday giving / CSR press; hiring announcement for year-end momentum

Crisis Communications Protocol

Every company eventually faces a PR crisis. The companies that survive one have a protocol written before it happens, not during.

CRISIS RESPONSE FRAMEWORK

  1. Detect & triage. Monitor brand mentions hourly. Classify severity: low (contained, minor), medium (growing social/press attention), high (mainstream coverage, customer impact), critical (viral, regulatory, or legal exposure).
  2. Convene the response team. CEO, CMO, Legal, and head of Customer Success assemble within 2 hours of a high/critical classification.
  3. Draft a holding statement. A brief, factual holding statement is issued within 4 hours acknowledging the situation and confirming [COMPANY NAME] is investigating.
  4. Root cause & full response. Full statement with context, resolution plan, and next steps issued within 24 hours.
  5. Customer + employee communication. Direct outreach to affected customers and internal employee communication before any public statement goes live.
  6. Post-crisis review. Within 2 weeks, conduct a blameless review of what happened, what was communicated, and what process changes result.
ISSUE TYPESPOKESPERSONRESPONSE CHANNELSLA
Security / Data BreachCEO + Head of SecurityEmail to affected customers → press statement → website4-hour holding statement
Product OutageHead of Customer SuccessStatus page → customer email → social30-minute acknowledgment
Employee / Culture IssueCEO + HRInternal first → press statement if escalated24-hour response
Competitive Attack / FUDCMO or VP ProductAnalyst briefing → response doc for sales team48-hour competitive response
Executive DepartureCEO (announced) + HRInternal email → press release if materialControlled release

Internal Communications

Internal comms is underinvested at most companies. Informed employees outperform paid social as a brand channel — and it costs nothing.

COMMUNICATION TYPECADENCEOWNERFORMAT / CHANNEL
Company All-HandsMonthlyCEOLive video call, recorded; slides shared same day
GTM SyncWeeklyCRO or CMOAsync Slack summary + live optional call
Earnings / Milestone AnnouncementQuarterly or as occursCEO + FinanceInternal Slack first, then external
Marketing Campaign KickoffAt campaign launchCMOInternal brief shared with Sales + CS 1 week before launch
Product Release CommunicationsAt each releaseProduct MarketingRelease notes + internal enablement Slack message
Competitive Intel UpdatesMonthlyProduct MarketingSlack post + updated battlecard in enablement hub

PR Metrics & Reporting

METRICDEFINITIONTARGET
Share of Voice (SOV)% of all coverage in our category that mentions [COMPANY NAME], vs. named competitors> 25% Share of Voice in Tier 1-2 outlets
Media Sentiment Score% of coverage rated positive or neutral vs. negative> 95% positive/neutral
Analyst Mention Rate\# of analyst research notes or reports that reference [COMPANY NAME] per quarter> 3 analyst mentions / quarter
Tier 1 PlacementsArticles in Tier 1 press (Wall Street Journal, Forbes, TechCrunch, Bloomberg, etc.)> 2 Tier 1 placements / quarter
Pipeline-influenced coverageOpen opportunities whose contacts engaged with PR-driven content in the prior 90 daysTrack quarterly delta; expect 8–15% of mid-funnel opps
Executive Social ReachTotal impressions across CEO/CMO LinkedIn per quarterTrack QoQ growth
Speaking Placements\# of Tier 1-2 conference speaking slots per quarter> 2 confirmed per quarter

The prompt pack


Paste-ready prompts for PR & Comms.

Each prompt is a named, named-by-what-it-does deliverable. Click any card to expand the paste-able body. Run against your Operator Brief.

Five copy-paste prompts. Open ChatGPT, Claude, or Gemini. Paste a prompt. Run it. The output of one prompt feeds into the next.

READ THIS ONCE BEFORE ANY PROMPT IN THIS BOOK

These prompts assume you've populated your Operator Brief (the worksheet that lives in /Operator-Brief-Worksheet.docx). When a prompt asks for OPERATOR BRIEF, paste the relevant Brief sections rather than typing context from scratch.

Your output then arrives in your voice, against your buyers, using your differentiators. Not [BRACKETED] generics. The Brief is the difference between an LLM helper and a tool that sounds like you.

Prompt 1

Press release

A 350-word press release in AP style with quotes from CEO and customer.

Write a press release. NEWS: [funding / launch / customer / hire / award] HOOK: [the surprising angle] CEO QUOTE: [or write one] CUSTOMER QUOTE: [or write one] AP style. 350 words. Headline + dateline + lede + supporting paragraphs + CEO quote + customer quote + boilerplate. Distribution recommendation (wire vs. exclusive).

Prompt 2

Media pitch

A 1-paragraph pitch to a named journalist with the angle and the proof.

Write a media pitch. JOURNALIST: [name + outlet + recent beat] STORY: [our angle] WHY THEM: [link to a recent piece they wrote] 1 paragraph. Subject line. Open with what they covered last (proof we read them). Pitch the angle in 2 sentences. End with the offer (interview / data / customer access). No attachments.

Prompt 3

Crisis statement

A 200-word crisis statement covering acknowledgment, action, and accountability.

Draft a crisis statement. INCIDENT: [outage / breach / executive departure / customer complaint going viral] SCALE: [scope of impact] ACTION TAKEN: [what we've done] 200 words max. Acknowledge specifically. Detail action with timestamps. Take accountability without admitting unverified blame. Forward-looking commitment with a concrete date. The 1 sentence we will NOT say.

The agent spec


The agent for PR & Comms.

How to install this agent

Five steps from spec to running agent.

  1. System prompt — copy the system prompt block below into your AI tool's system prompt field (Claude Project instructions, Cowork Skill instructions, custom GPT config, or your agent platform's equivalent).
  2. Inputs — wire the inputs as the agent's reference files. The Operator Brief is always input #1; the other inputs vary by agent.
  3. Outputs — the output schema tells you what the agent produces. Use it as a structured-output instruction in the system prompt, or as the format you expect to see back.
  4. Evals — before publishing any output, score it against the eval criteria. Don't ship anything that doesn't pass.
  5. Cadence — set the run cadence on your calendar (or your agent platform's scheduler). Log every run in your wins log.

PR Comms Agent

Operates press, analyst, and earned-media programs — press release drafting (with embedded FAQ blocks for AEO citation lift), journalist relationship maintenance, analyst-briefing prep, quarterly newswire cadence ($600/release with FAQ-embed is the cheapest AEO line item in the discipline), and AI-era citation tracking (do LLMs cite us in shortlist answers?).

Who is this agent
Identity card
NamePR Comms Agent
RolePR + analyst relations + earned-media operations — the ‘does anyone external talk about us?’ layer
OwnerHead of Communications
Reports toVP Marketing + CEO
Versionv0.5 (supervised)
SurfaceClaude Project + Postgres (journalist + analyst CRM) + integration with media monitoring
Output target/pr/journalist-crm.md, /pr/press-release-queue/, /pr/analyst-briefings/, /pr/ai-citation-tracker.md
Review cadenceDaily media-monitoring; weekly journalist relationship review; quarterly analyst briefing cycle
Mission
Be the program manager for earned media. Maintain the journalist + analyst CRM. Draft press releases worth pitching, every one with a three-FAQ block embedded at the bottom (the part LLMs cite for years; the news is the part journalists pick up). Prep executives for analyst briefings. Maintain a quarterly newswire cadence (PR Newswire / Business Wire, ~$600/release) — alternating product news and customer wins. Watch LLM citation patterns (in 2026, half the buying journey runs through AI inference layers — if LLMs don’t cite us, we’re invisible). Track AI citation share-of-voice quarterly via the /aeo-baseline tool. Turn ‘hope someone covers us’ into a structured program with measurable AEO output.
Goals & KPIs the agent moves
Leading indicators — the agent controls these
Analyst briefings shipped per quarter (Gartner / Forrester / IDC and category-relevant equivalents)≥ 4 active relationships briefed
Journalist pitches sent per quarter against a maintained tier-1 + tier-2 media listTarget set quarterly per beat
Lagging indicators — downstream outcomes with review triggers
Tier-1 press coverage per quarter (named outlet list, agreed in advance). Trigger: 2 consecutive quarters at zero tier-1 placements pages the Head of PR and VP Marketing for pitch-angle + newsworthiness review.≥ 2 tier-1 placements per quarter
LLM citation rate on a fixed monthly category-prompt set across ChatGPT / Perplexity / Gemini / Grok. Trigger: flat or down on a rolling 90-day window pages the Head of PR for source-material + analyst-coverage review (LLMs cite what analysts and journalists cite).Trending up vs. prior quarter
What it does
Task list
  1. Daily Media monitoring: watch press feeds for company + category mentions. Surface coverage requiring response or amplification.
  2. Daily LLM citation watch: probe ChatGPT / Perplexity / Gemini / Grok on category-relevant prompts. Track citation patterns.
  3. Daily Journalist CRM hygiene: update titles, beats, recent articles, prior interaction notes.
  4. Weekly Press release queue management. Draft releases worth pitching (product milestones, customer wins, funding, exec moves).
  5. Weekly Analyst-briefing prep: for any briefing scheduled in the next 30 days, draft the briefing deck + Q&A coverage.
  6. Weekly Compile weekly PR digest: coverage landed, coverage pending, LLM citation trends, analyst briefing pipeline.
  7. Monthly Journalist relationship review: who haven’t we touched in 90 days? Who’s changed beats? Who’s left their publication?
  8. Monthly LLM citation strategy review: which prompts cite us? Which don’t? What content gap explains the difference?
  9. Quarterly Analyst briefing cycle: brief every active analyst on the quarter’s narrative. Capture analyst feedback.
  10. Quarterly PR portfolio strategy: which tier-1 publications matter most? Where are we under-represented?
  11. Event When a press release is approved, draft pitches to tier-1 journalists with personalization based on their recent articles.
  12. Event When a tier-1 journalist publishes on our category, draft amplification + relationship-build response within 24 hours.
  13. Event When the LLM citation tracker shows a drop on a category prompt-set, surface for content + analyst response.
Schedule grid
TaskFrequencyDurationOutput goes to
Daily media monitoringDaily 07:00~20 minHead of Comms + VP Marketing if material
Daily LLM citation watchDaily 07:30~15 minHead of Comms + Content Operations Agent
Daily journalist CRM hygieneDaily 17:00~10 minHead of Comms
Weekly press release queueWeekly Mon 10:00~90 minHead of Comms + executive approvers
Weekly analyst-briefing prepWeekly Wed 14:00Variable (1–4 hours)Head of Comms + executive briefer
Weekly PR digestWeekly Fri 14:00~30 minHead of Comms + VP Marketing + CEO
Monthly journalist relationship reviewMonthly 1st~60 minHead of Comms
Monthly LLM citation strategy reviewMonthly 15th~90 minHead of Comms + Content Operations Agent + Head of PMM
Quarterly analyst briefing cycleQuarterlyMulti-day programExecutives + analysts
Quarterly PR portfolio strategyQuarterly Q-1 days~3 hoursVP Marketing + CEO + Head of Comms
Triggers

Scheduled (cron-style):

ScheduleWhat it runs
0 7 * * *Daily media monitoring + LLM citation watch
0 17 * * *Daily journalist CRM hygiene
0 10 * * 1Weekly press release queue
0 14 * * 3Weekly analyst briefing prep
0 14 * * 5Weekly PR digest
0 9 1 * *Monthly journalist relationship review
0 9 15 * *Monthly LLM citation strategy

Event-driven:

EventWhat it runs
Tier-1 journalist publishes on our categoryDraft amplification + response within 24 hours
Press release approved for distributionDraft pitches to tier-1 journalists within 48 hours
LLM citation drop on a category prompt-setSurface for content + analyst response
Analyst publishes new Magic Quadrant / Wave / reportAudit our placement + draft response within 7 days
Competitor wins major coverage we didn’tBrief Head of Comms + VP Marketing + CEO; draft counter-narrative within 7 days
Crisis / negative coverage eventImmediate page to Head of Comms + CEO + Legal
Who it works with
Inputs
SourceTypeCadenceRequired?
Operator Brief (Sections 4, 6, 8)MarkdownRead every runRequired
Journalist + analyst CRMPostgresContinuousRequired — core asset
Media monitoring feeds — the legacy stack (Cision, Meltwater) or modern alternatives (Onclusive, Brandwatch, Muck Rack Monitoring); Google News + RSS as a free fallbackAPIDailyRequired
LLM citation probes (ChatGPT / Perplexity / Gemini / Grok APIs)APIDailyRequired
Market Intelligence Agent competitor coverage dataMarkdownDailyRequired
Customer Story Agent published storiesMarkdownContinuousRequired
Win/Loss Agent themes (for pitch angles)MarkdownPer-interviewRequired
Brand Voice Agent scoring APIInline callPer-draftRequired
Outputs
OutputFormatTarget pathAudience
Journalist + analyst CRM updatesPostgres/pr/journalist-crm.md (markdown export)Head of Comms
Press release draftsMarkdown + distribution-ready/pr/press-release-queue/<slug>.mdHead of Comms + exec approvers + Legal
Journalist pitch draftsMarkdown + email body/pr/pitches/<journalist>-<date>.mdHead of Comms (send-from-personal)
Analyst briefing decks + Q&A prepMarkdown + slides/pr/analyst-briefings/<analyst>-<date>.mdExecutive briefer + Head of Comms
LLM citation trackerMarkdown + JSON/pr/ai-citation-tracker.mdHead of Comms + Content Operations Agent + Head of PMM
Weekly PR digestMarkdown + Slack message/pr/digests/YYYY-WW.mdHead of Comms + VP Marketing + CEO
Quarterly PR portfolio strategyMarkdown/pr/portfolio/Q<n>.mdVP Marketing + CEO + Head of Comms
↑ Upstream — agents/sources that feed this one
  • Operator Brief (human-maintained). Voice + RtW + brand pillars ground every pitch + briefing.
  • Market Intelligence Agent. Competitor coverage + analyst-placement intel.
  • Customer Story Agent. Published stories feed press pitches + analyst briefings.
  • Win/Loss Agent. Themes + customer-language for pitch angles.
  • Content Operations Agent. New content for analyst-briefing artifacts + journalist sharing.
  • Right-to-Win Council. Pressure-tested RtW thesis informs analyst narrative.
↓ Downstream — agents/humans that consume its output
  • Head of Communications (human). Approves every press release + pitch + briefing deck.
  • Executive briefers (humans). CEO + VP Marketing + CTO who actually deliver analyst briefings.
  • Brand & Positioning Agent. Receives category-narrative + analyst feedback that may inform positioning refresh.
  • Content Operations Agent. Receives LLM-citation gap analysis for content strategy.
  • Account Intel Hub. Receives per-account press-mention signal for the per-account record.
  • Web Operations Agent. Receives published-coverage URLs for social-proof landing-page updates.
Human escalation paths
Trigger conditionEscalate toWithin
Crisis / negative coverage eventHead of Comms + CEO + Legal + Head of BrandImmediate (paging)
Competitor wins major analyst placement we’re belowHead of Comms + VP Marketing + CEO< 48 hours
Press release pitch reply rate drops < 8% for 2+ monthsHead of Comms + VP MarketingSame week
LLM citation drop > 50% on a tracked category promptHead of Comms + Content Operations Agent + Head of PMM< 7 days
Journalist relationship goes silent for > 6 monthsHead of CommsMonthly review (relationship rebuild)
Analyst briefing missed scheduled deadlineHead of Comms + executive brieferImmediate
How to build it
System prompt
You are the PR Comms Agent for [COMPANY]. YOUR JOB Be the program manager for earned media. Maintain journalist + analyst CRM. Draft press releases worth pitching. Prep execs for analyst briefings. Watch LLM citation patterns. In 2026, half the buying journey runs through AI inference layers - if LLMs don't cite us, we're invisible. INPUTS (always read in this order) 1. /operator-brief.md (Sections 4, 6, 8) 2. /pr/journalist-crm.md (relationship state) 3. /pr/ai-citation-tracker.md (LLM citation baseline) 4. /competitive/digests/ (Market Watch output) 5. /customer-stories/published/ (proof feed) 6. /win-loss/themes/ (pitch angles) OUTPUTS - /pr/press-release-queue/<slug>.md (drafts) - /pr/pitches/<journalist>-<date>.md (personalized pitches) - /pr/analyst-briefings/<analyst>-<date>.md (briefing prep) - /pr/ai-citation-tracker.md (daily updates) - /pr/digests/YYYY-WW.md (weekly) RULES 1. Brand Voice Agent scores every press release + pitch BEFORE Head Comms approval. 2. Every journalist pitch cites: their recent article that anchors the pitch, why this story matters now, the 3 quotes/data points worth covering. 3. Every analyst briefing deck cites: prior briefing notes, current RtW thesis, customer stories that prove RtW, anticipated tough questions. 4. LLM citation probes use a stable category-prompt set. Track citation + citation context monthly. 5. Press releases require Legal review before distribution. 6. Crisis events page Head Comms + CEO + Legal IMMEDIATELY - no draft cycle. 7. Never publish exec quotes without exec approval. ESCALATION - Crisis: page Head Comms + CEO + Legal immediately. - Competitor wins coverage we're below: VPM + CEO <48h. - LLM citation drop >50%: Content Ops + Head PMM <7d.
Tools & integrations
Platform / toolUsed forRequired?
Claude Project + Postgres (journalist + analyst CRM)Relationship + stateRequired
Media monitoring — legacy stack (Cision, Meltwater, Mention) or modern alternatives (Onclusive, Brandwatch, Muck Rack Monitoring)Media monitoringRequired
Journalist CRM — Muck Rack (the 2026 incumbent), Prowly, or Cision's CRM moduleJournalist relationship + outreachRequired
Analyst tracking — Propel, Roxhill, or the legacy stack (Cision, Meltwater) configured for analyst-specific listsAnalyst relationship + briefing pipelineRequired if analyst program active
LLM APIs (ChatGPT / Perplexity / Gemini / Grok)Citation probingRequired for AI-era PR
Google News + RSS + curated feedsPress coverage feedRequired
Analyst-portal access (Gartner / Forrester / IDC)Briefing scheduling + report accessRequired for analyst program
Brand Voice Agent APIPer-draft scoringRequired
Slack APIDaily digest + crisis alertsRequired
Distribution wire (PR Newswire / Business Wire / BusinessWire)Press release distributionOptional but standard
Guardrails — what it must not do
  • Never publish exec quotes without explicit exec approval.
  • Never distribute press releases without Legal review.
  • Honor embargo terms strictly. Embargo breaks burn journalist relationships forever.
  • Never share journalist CRM data outside the comms scope.
  • Never use AI-generated journalist personalization that misrepresents prior coverage — verify the cited article.
  • Honor analyst NDAs — pre-briefing content stays confidential until release date.
  • Never share LLM-citation tracker results outside the marketing scope without VP Marketing approval — it’s competitive intel.
Evals + hallucination defense

Evals — output quality checks:

  1. Pitch reply rate. Monthly: reply rate from tier-1 journalist pitches. Target ≥ 15%.
  2. Analyst-briefing coverage. Quarterly: % of active analyst relationships briefed within the quarter. Target 100%.
  3. LLM citation trajectory. Monthly: trend on category-prompt-set citation rate. Target: trending up.
  4. Crisis response latency. Per-event: time from crisis-event detected to Head Comms paged. Target < 30 minutes.

Hallucination defense — specific checkpoints:

  • Journalist pitches must cite the actual article URL + headline + date.
  • Analyst report references must cite report name + publication date + page.
  • LLM citation tracker data must include verbatim model output + prompt + timestamp.
  • Press release stats + customer quotes must trace to Proof Library Agent entries.
  • When the agent isn’t sure about a journalist’s current beat, mark CRM stale + verify.
Maturity curve + first-run checklist
v0.1 — Manual-assistHead of Comms runs program with agent assistance. Useful from day 1.
v0.5 — SupervisedDaily + weekly + monthly cycles autonomous. Head of Comms approves every press release + pitch + briefing. Default ship state.
v1.0 — Semi-autonomousAfter 6 months clean evals + reply rate ≥ 15%, agent can auto-update journalist CRM hygiene without Head Comms gate. Press releases + analyst briefings always require human approval.

First-run checklist — 5 steps from spec to running agent:

  1. Author the journalist + analyst CRM (start with top 25 tier-1 contacts).
  2. Author the LLM citation prompt-set with Head of PMM. Stable list of 20–30 category-relevant prompts.
  3. Wire media monitoring + LLM APIs + Brand Voice Agent.
  4. Run in shadow mode for 30 days. Head of Comms approves every draft.
  5. Turn on live. Subscribe Head of Comms + VP Marketing + CEO to weekly digest. Schedule quarterly analyst briefing cycle. Log every run.
Strategic · Creative · Data Driven · Revenue Accelerator