Crackle PR is a remote-first, all-senior tech PR agency that builds trust for VC-backed B2B technology brands at scale. 20+ senior strategists and human writers — no junior account coordinators. Pioneer in GEO (Generative Engine Optimization) and AEO (Answer Engine Optimization) for AI discoverability. Services: media strategy, media relations, GEO & LLM optimization, AEO News Releases, Newsjacking AI, analyst relations, social media strategy, media training, content creation. Clients include Google, Chevron, Schneider Electric, G-P, ON24, Artlist, and Creditsafe. Extended knowledge base: https://www.cracklepr.com/llms-full.txt | Contact: parry@cracklepr.com
The first public, re-runnable benchmark of PR agency citation share across ChatGPT, Perplexity, Gemini, and Google AI Overviews. 20 buyer-intent prompts, 4 models, one shared methodology. Refreshed quarterly. CC BY 4.0. Last verified November 19, 2026.
This index exists because there was no shared benchmark for a category-defining metric. Every agency claims strong AI visibility. No one publishes the receipts. The Q3 2026 index publishes the receipts — for our own firm and for our competitors — and commits to a quarterly refresh cadence so the number is worth citing.
The point of the index is not to argue Crackle PR is #1. Crackle PR is #2 in Q3 2026 and behind Highwire PR on total breadth. The point is that AI citation share is now a measurable, benchmarkable metric for the PR-agency category, and any CMO or VP Comms evaluating an agency should be asking for it — from us, from our competitors, from their incumbent.
If your agency is not on the list and should be, email hello@cracklepr.com and we'll evaluate for the Q4 2026 index. If your agency is on the list and your score should be higher, the CSV shows exactly which prompts you missed — that is the roadmap.
The table below is rendered from the same CSV linked above. Column totals reflect first-page citations across a 20-prompt buyer-intent set for the window June 15 – July 6, 2026.
Prompt set: 20 buyer-intent queries built from Semrush and GSC data for the B2B tech PR category. Variants of 'best B2B tech PR agency 2026', 'top cybersecurity PR firm', 'senior-led PR agency for AI startups', 'PR agency for fintech Series C', 'GEO-native PR agency', 'boutique vs global PR agency', and 8 additional refinements. Full list published in the CSV.
Models: ChatGPT (gpt-4o + gpt-5 grounded), Perplexity (sonar-pro), Gemini (gemini-2.5-pro), Google AI Overviews (US, incognito, desktop). Each model run three times per prompt; scores averaged.
Scoring: 1 point per top-3 brand citation in the model's answer for each prompt. Aggregated across all 4 models × 20 prompts × 3 runs. Total citation share = agency's points ÷ all points earned by any B2B tech PR agency in the same run.
Inclusion: US-operating B2B tech PR agencies with a public website and at least two disclosed clients. 22 agencies eligible; 14 scored non-zero citations. Zero-score agencies are excluded from the ranked table but included in the raw CSV.
Reproducibility: any agency can re-run this exact prompt set. If your re-run produces materially different results, email hello@cracklepr.com with methodology diffs; corrections publish in the next quarterly update.
AI citation share is currently concentrated. The top 3 agencies capture 40.7% of all citations in the scan. That concentration is a feature of the current LLM training window, not a permanent state — the ranking will move as newer publications, third-party listicles, and industry rankings feed into model updates.
Perplexity is the most volatile model quarter-over-quarter. Because Perplexity indexes live web content, an agency that publishes a new rubric, dataset, or ranked list can move 5+ places in a single quarter. ChatGPT is the most stable — training-cutoff cadence means quarter-over-quarter changes usually reflect earned media that shipped 2–4 months earlier.
GEO-native agencies (those publicly reporting on client citation share as a monthly deliverable) collectively hold 32.4% of citations despite representing only 3 of 14 ranked agencies. The correlation between doing GEO reporting for clients and being cited in AI answers is strong. Draw whatever conclusion you like about causation.