Insights · Analysis
Analysis of how ChatGPT, Claude, Gemini, Perplexity, Copilot, Grok, DeepSeek, Meta AI, Google AI Overviews, and AI Mode are rewriting discovery, and what to do about it.
177 insights · search & filter
A long tail of 43 unknown software comparison sites took 14.4% of ChatGPT's citations - 3.5x what G2, Capterra and TrustRadius managed combined.
Clio holds the strongest mean rank of any brand in the study at 1.20. But 45.9% of legal citations go to third parties, one of which listed a funeral home.
Healthcare EHR drew the heaviest retrieval in the study at 60.5 results per answer, and the lowest vendor citation share at 48.7%.
Of 1,249 dated results ChatGPT retrieved, 90.4% were published within twelve months and only 5.0% were older than two years. Freshness is near-absolute.
Accounting had the highest vendor citation share of any category tested. One budget prompt sent all seven citations to a single vendor's own site.
Ecommerce platforms took 83.8% of citations and project management 87.5%. Shopify and Asana hold two of the strongest mean ranks in the study.
CRM was the most stable category tested. HubSpot led 67% of prompts and hubspot.com was the single most-cited domain in the entire 60-prompt study.
Six phrasings of the same buying question returned brand sets overlapping by a mean Jaccard of 0.338. In HR, all six prompts had a different winner.
Across 367 citations in 60 software buying answers, vendor-owned pages took 65.9%. Review aggregators took 4.1%. Reddit and YouTube took zero.
Primary research · August 2026
An audit across 10 categories and 60 buying questions. I recorded what ChatGPT reads, throws away and links to when a buyer asks it which software to buy, and what that decides.
Free download
35 pages · PDF · 536 KB. Enter your details and it downloads straight away.
