State of AI Search Reports
We ran real queries through ChatGPT (GPT-4o) and Perplexity (Sonar Pro) across 5 B2B verticals to find out which companies get recommended — and which ones don't exist in AI search. These are the results.
All Reports
Each includes full data, methodology, PDF download, CSV dataset, and audio summary
State of AI Search: B2B SaaS
The report that started it all. We analyzed 100 B2B SaaS companies across HR Tech, Fintech, and general SaaS to see which ones ChatGPT and Perplexity actually recommend. 47 got cited. 53 were invisible.
State of AI Search: Cybersecurity
Cybersecurity had the broadest coverage of any vertical studied. Palo Alto Networks leads with 39 citations, but the real story is the long tail of vendors fighting for positioning below the top 5.
State of AI Search: MSP Software
Managed Service Provider software is a surprisingly tight race. NinjaOne leads (43 citations), but Atera, ConnectWise, and Syncro are all within striking distance at 36 each.
State of AI Search: Microsoft Integrators
The consulting and integration partner landscape for Microsoft is dominated by Avanade (54 citations). Accenture sits at #2 with 42, and Hitachi Solutions rounds out the top 3 at 38.
State of AI Search: NetSuite Integrators
NetSuite's integration partner ecosystem shows RSM with a clear lead (42 citations). Folio3 and Myers-Holum tie for second at 29 each. Only 12 companies appear on both AI engines.
Our Methodology
Every report follows the same process. Full methodology details, including query lists, engine configurations, and scoring rules, are available on each individual report page.
Define buyer-intent queries
We map the exact questions your prospects ask when evaluating software — the queries that drive pipeline, not vanity traffic. Query counts range from 12 to 42 per vertical.
Run queries through AI engines
Every query is tested on ChatGPT (GPT-4o, web browsing disabled) and Perplexity (Sonar Pro). Fresh browser sessions with no prior conversation history. All queries run from US-based IP addresses.
Record and score citations
A citation is counted when a company is named as a recommendation, listed in a comparison, or presented as a solution. Each company receives one citation point per appearance. No position weighting is applied. We track total citations, per-engine citations, and cross-engine presence.
Analyze and publish with open data
Results are compiled into per-engine breakdowns with full company rankings, methodology documentation, and limitations disclosures. Each report includes the complete dataset as a downloadable CSV under a CC BY-NC 4.0 license.
Engine coverage note
These reports cover two AI engines: ChatGPT (GPT-4o) and Perplexity (Sonar Pro). Other major engines — including Google Gemini, Microsoft Copilot, and Claude — are not included. Companies that perform well on the two tested engines may perform differently on others. Future editions may expand engine coverage.
A note on reproducibility
AI engine responses are non-deterministic — running the same query twice can produce different recommendations. Each query was run once per engine per testing session, so these results represent a single observation, not a statistical average. Our reports capture a snapshot of AI engine behavior during a specific testing window. We document engine versions, query lists, testing dates, and browser configurations so others can attempt replication. Individual results may vary.
Sponsor disclosure
This research is self-funded by AnswerManiac. No company paid for inclusion, higher placement, or favorable treatment. AnswerManiac offers AI visibility services; this research informs those services but was not commissioned by any client.
Open Data
All datasets are available for download under a CC BY-NC 4.0 license. Use the data for your own analysis, cite as: Swiler, A. et al. “State of AI Search,” AnswerManiac, 2026.
Want your vertical researched next?
We build custom State of AI Search reports for specific verticals. Get the data your competitors don't have.