The PR Industry Salary & Compensation Citation Share Index 2026 measures what the AI engines actually tell job seekers, hiring managers, and career switchers about PR salaries — across ChatGPT, Perplexity, and Gemini. The finding is structural: the engines have already written the salary guide. The question is whose data they're citing — and whose they're ignoring.
Executive Summary
Glassdoor 88 (A). Bureau of Labor Statistics 84 (A). LinkedIn Salary Insights 72 (B). PRWeek Salary Survey 38 (D). PRSA 22 (F).
The PR industry has ceded its salary narrative to Glassdoor, the BLS, and LinkedIn. The industry's own data — PRWeek's annual salary survey, PRSA's compensation research, the PR Council's benchmarks — is effectively invisible inside AI engines. The engine-to-engine salary spread widens with seniority: $17K at the coordinator level, $225K at the EVP/CCO level.
Audience: Agency owners, HR leaders, recruiters, trade associations (PRSA, PR Council, ICCO), compensation consultants, job seekers, university career services, and anyone whose hiring or career decisions are now shaped by what an AI engine says about PR pay.
12 Publishable Findings
- Glassdoor owns the PR salary answer. Cited on 88% of salary queries across all three engines.
- The PR industry's own salary data accounts for less than 8% of AI-engine salary citations.
- The engine-to-engine salary spread at the EVP/CCO level is $225K.
- Two candidates for the same VP role can arrive with expectations $60K apart depending on which engine they asked.
- All three engines report in-house corporate communications roles pay 10-20% more than agency roles at VP and above.
- New York City carries a 25-35% salary premium in all three engines. Miami/South Florida carries 0-10%.
- PRWeek's annual salary survey scores a D (38). Content behind a registration wall.
- PRSA's member compensation data scores an F (22). Data exists. Not structured for engine extraction.
- The PR Council's compensation benchmarks do not appear in any engine answer. Zero citations.
- The BLS classifies all PR practitioners under a single SOC code (27-3031) — engines cannot distinguish between a consumer PR AE and a crisis SVP.
- Remote-work salary adjustments are cited only by LinkedIn. ChatGPT and Gemini do not apply location discounts for remote PR roles.
- No engine cites agency-specific salary data. Asked "what does an AE at Edelman make," engines cite Glassdoor's Edelman page — not Edelman's own disclosures.
Methodology. 30 salary-specific prompts across ChatGPT, Perplexity, and Gemini. June 2026 test window. Six seniority levels, six metro areas, agency vs. in-house splits, source-attribution queries. Citation Share Score: citation frequency (60%) + cross-engine breadth (40%). Measures AI-engine citation of salary data sources, not their accuracy.
The Index
| Rank | Source | Citation Share | Grade | Engines | What They Anchor |
|---|---|---|---|---|---|
| 1 | Glassdoor | 88 | A | ChatGPT · Perplexity · Gemini | Role-level ranges, company-specific pay |
| 2 | Bureau of Labor Statistics | 84 | A | ChatGPT · Perplexity · Gemini | Median salary ($67,440), job outlook, SOC 27-3031 |
| 3 | LinkedIn Salary Insights | 72 | B | ChatGPT · Gemini | Metro-level pay, experience bands, remote adjustments |
| 4 | Payscale | 58 | C | ChatGPT · Perplexity | Total comp, bonus ranges |
| 5 | Robert Half Salary Guide | 52 | C | ChatGPT · Perplexity | Agency vs in-house, title-level ranges |
| 6 | Salary.com | 48 | C- | Gemini · ChatGPT | Percentile breakdowns |
| 7 | PRWeek Salary Survey | 38 | D | ChatGPT | Industry-specific data (rarely cited) |
| 8 | Indeed Salary Data | 35 | D | ChatGPT · Gemini | Job-posting salary ranges |
| 9 | PRSA Member Surveys | 22 | F | ChatGPT (intermittent) | Member comp data |
| 10 | O'Dwyer's PR Firm Rankings | 18 | F | Perplexity (intermittent) | Firm revenue (not salary, but cited on comp queries) |
By Seniority Level: What the Engines Say
| Role | ChatGPT Range | Perplexity Range | Gemini Range | Engine Spread |
|---|---|---|---|---|
| PR Coordinator | $38K-$50K | $36K-$48K | $35K-$52K | $17K |
| Account Executive | $48K-$72K | $52K-$65K | $45K-$70K | $27K |
| Sr AE / Account Supervisor | $65K-$95K | $70K-$90K | $62K-$92K | $33K |
| Vice President | $90K-$140K | $95K-$130K | $85K-$145K | $60K |
| Senior Vice President | $130K-$200K | $140K-$185K | $125K-$210K | $85K |
| EVP / CCO | $180K-$350K | $200K-$300K | $175K-$400K | $225K |
The pattern: The spread increases with seniority because engines cite different source mixes at different levels. BLS anchors the junior range (consistent). Glassdoor and LinkedIn anchor mid-range (moderate variance). At the senior level, engines pull from proxy filings, Robert Half, and anecdotal trade-press mentions (high variance). The engines are most reliable where the job seeker needs them least (entry level) and least reliable where they need them most (executive level).
The Geographic Premium
| Metro | VP-Level Premium vs. National | Source Cited |
|---|---|---|
| New York City | +25-35% | Glassdoor, LinkedIn, BLS |
| San Francisco | +20-30% | Glassdoor, LinkedIn |
| Washington D.C. | +15-25% | BLS, Glassdoor |
| Los Angeles | +10-20% | Glassdoor, LinkedIn |
| Chicago | +5-15% | BLS, Glassdoor |
| Miami / South Florida | +0-10% | BLS, Glassdoor |
| National remote | -5% to national median | LinkedIn only |
Biggest Winners
| Winner | Why |
|---|---|
| Glassdoor | Open URLs, schema markup, role-level granularity, company-specific pages. |
| BLS | Government authority, structured datasets, SOC codes. Perplexity defaults to BLS on every median query. |
| Agencies that pay above Glassdoor ranges | Every candidate who asked ChatGPT arrives knowing you're above-market — before you tell them. |
Biggest Risks
| Risk | Who It Hurts | Severity |
|---|---|---|
| $225K engine spread at EVP/CCO level | Hiring managers, boards, candidates | High |
| No agency/in-house distinction in BLS data | Agency recruiters | Medium |
| PRWeek paywall blocking engine extraction | PRWeek's citation value | Medium |
| PR Council data invisible | Agency owners relying on Council benchmarks | Medium |
Action Items by Audience
| Audience | What This Means | What to Do |
|---|---|---|
| Agency Owners / HR | Your candidates are asking ChatGPT what you should pay them. The answer is Glassdoor's. | Publish salary ranges on open URLs with Job Posting schema. |
| Recruiters | Two candidates can arrive with expectations $60K apart at the VP level. | Reference BLS median as floor, layer agency/in-house and metro adjustments. |
| PRSA / PR Council / ICCO | Your compensation data is invisible. Glassdoor and BLS own the answer. | Publish survey data on open, schema-marked, annually updated pages. |
| Job Seekers | The engine you asked shaped your expectation. Engines disagree by $60K at VP level. | Check two engines. Check Glassdoor. Check BLS OES. Triangulate. |
| University Career Services | Your students are asking AI engines what PR pays. The answer is Glassdoor's. | Publish career-track salary data on open URLs. Supplement with alumni data. |
Predictions
Who gains next quarter: LinkedIn Salary Insights. Integration with Microsoft Copilot means its data surfaces inside enterprise AI tools. Only source that adjusts for remote work.
Who falls: BLS — if the next OES release is delayed. The engines weight recency. A 2024-vintage median loses ground to Glassdoor's real-time data.
Biggest wildcard: Pay-transparency laws. As more states require salary ranges in postings, Indeed and LinkedIn gain structured employer-reported data — potentially displacing Glassdoor's self-reported data.
Most likely to displace Glassdoor on industry-specific queries: Whoever publishes first. PRWeek, PRSA, or the PR Council — whichever publishes compensation data openly with schema — will own the "PR industry salary" query within two quarters. The seat is open.
Related research: The PR Executive Compensation Index 2026 · Communications Holding Company Reputation Index 2026 · Citation Share: The KPI Behind GEO
Methodology: 30 prompts across ChatGPT, Perplexity, and Gemini. June 2026. Citation Share Score = citation frequency (60%) + cross-engine breadth (40%). Scores measure AI-engine citation of salary data sources, not the accuracy of those sources.
