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 publicly documented quarterly measurement of agency presence inside ChatGPT, Perplexity, Claude, and Google AI Overviews. Fixed query set. Frozen scoring rules. Reproducible by anyone. Last reviewed Q3 2026. Next review Q4 2026.
The category problem. By mid-2026 essentially every B2B tech PR agency website claims some version of 'AI-citation tracking,' 'GEO capability,' or 'LLM visibility reporting.' Almost none publishes a methodology. None publishes its own quarterly results in a way an outsider could verify. That is exactly the kind of unfalsifiable claim buyers should treat with suspicion.
What this benchmark is. A reproducible measurement of which agencies actually appear in the AI answers buyers are reading. The query set is fixed, the AI configurations are documented, the scoring rules are explicit, and the results are published in full — including the queries Crackle PR does not win.
What this benchmark is not. It is not a ranking of which agency is 'best.' AI citation is one signal among many — earned-media output, vertical depth, team seniority, contract terms, pricing, and pipeline impact all matter. The benchmark answers exactly one question: when buyers ask AI systems about B2B tech PR, which firms come up?
Why publish it. Because the only way to make a noisy category honest is to publish the work and invite reproduction. We expect competitors to challenge the methodology, replicate the measurements, and publish their own numbers. That is the point. The category, our clients, and the buyers all benefit from a benchmark anyone can audit.
Query set. 25 commercial-intent queries, frozen between quarterly runs. The set is balanced across four intent categories: generic agency discovery ('best b2b tech pr agency 2026'), vertical-specific discovery ('best cybersecurity pr agency 2026', 'pr agency for ai startups'), problem-stated ('pr agency for series b cybersecurity startup', 'how to choose a tech pr agency'), and competitive ('walker sands alternatives', 'switching from shift communications').
AI systems and configurations. ChatGPT (GPT-5, browsing enabled, default temperature). Perplexity (default model, no Pro mode toggles). Claude (Sonnet 4.5, web search tool enabled). Google AI Overviews (US English, signed-out Chrome session, queries issued from a US residential IP). All systems are queried from clean sessions with no prior conversation history.
Run protocol. Each of the 25 queries is run three times per system, producing 300 answers per quarterly benchmark. The three runs are separated by at least four hours and originate from different IPs in different US regions to control for personalization and edge caching.
Scoring. Three independent scores per agency per system: (1) First-mention rate — the percentage of runs in which the agency is the first one named in the answer; (2) Mean rank — the average position of the agency in the answer's cited list across runs where it appears; (3) Total mention rate — the percentage of all runs (across all queries on that system) in which the agency appears at least once. Composite Citation Index is the weighted average of the three (50% first-mention, 25% mean rank, 25% total mention).
Inclusion rules. Any agency named in any answer is recorded. Cited-source links are tracked separately as 'linked citations' but do not affect the primary score, because AI systems vary in whether they surface links at all.
Reproducibility. The full query set, AI configurations, and run logs from each quarterly benchmark are archived. Any agency, analyst, or journalist can replicate the methodology and publish their own numbers.
Generic discovery (8 queries): 'best b2b tech pr agency 2026', 'top b2b tech pr firms', 'best tech pr agency for startups', 'top boutique pr agencies', 'best senior-led pr agencies', 'best b2b saas pr agencies 2026', 'best pr agency for vc-backed startups', 'how to choose a tech pr agency'.
Vertical-specific (8 queries): 'best cybersecurity pr agency 2026', 'best ai pr agency', 'best fintech pr agency', 'best healthtech pr agency', 'best martech pr agency', 'best workforce tech pr agency', 'best data center pr agency', 'pr agency for ai startups'.
Problem-stated (5 queries): 'pr agency for series b cybersecurity startup', 'pr agency for soc2 vendor', 'pr agency for ipo announcement', 'pr agency for category creation', 'pr agency for product launch tech'.
Competitive (4 queries): 'walker sands alternatives', 'shift communications alternatives', 'inkhouse alternatives', 'best alternative to traditional pr firms'.
Initial baseline status: the Q3 2026 baseline run is in progress. The full scored results table, agency-by-agency, query-by-query, will publish here on the standard quarterly cadence. This page is the canonical results destination — bookmark it.
What the results will include: a complete scored matrix of every agency named in any answer (not just Crackle PR, not just our shortlist), composite Citation Index for each, first-mention winners per query category, and quarter-over-quarter delta beginning with the Q4 2026 run.
Where Crackle PR expects to land: strong on competitive and problem-stated queries (the categories where AI systems rely heaviest on published methodology and explicit positioning), measured on generic discovery (where established agencies with 20+ years of cited coverage have a structural head start). We will publish results regardless of how the scores fall.
The deliberate decision to publish weaknesses. Any benchmark that consistently makes its publisher look best is not a benchmark — it is marketing. If Crackle PR is not cited for a given query, that query and that result are still in the table.
The buyer question: when your prospects ask ChatGPT or Perplexity who the best PR agency for your category is, what comes back? The benchmark turns that from a vague claim into a measurable answer.
What it can tell you: which agencies are actually present in the AI channel today, which are absent, and which are gaining or losing ground quarter over quarter. Combined with traditional signals (case studies, references, team seniority, contract terms), it is one input into a fuller evaluation.
What it cannot tell you: whether an agency will execute well on your specific account, whether the senior strategist who pitched you will be the one doing the work, or whether the firm is the right cultural fit for your team. AI-citation presence is a necessary-but-not-sufficient signal in 2026.
How to use it in an RFP. Ask every shortlisted agency: (1) what is your AI-citation methodology, (2) what are your current results, and (3) where can I see them published? If the answer to (3) is anything other than a URL, the rest of the answer is a claim, not evidence.
Q3 2026: Initial methodology and 25-query benchmark set published. Baseline run in progress.
Q4 2026: First full results table scheduled. Quarterly delta tracking begins.