State of AI Search: B2B Challengers
A fresh 100-company panel of growth-stage B2B challengers, measured for how often ChatGPT and Perplexity actually recommend them.
Overview
For Q3 we rebuilt the panel from scratch: 100 growth-stage B2B challengers across 7 verticals, none of them carried over from our Q2 study — no incumbents, no household names, the kind of companies that still have room to win AI visibility. We ran 155 unbranded, buyer-intent queries through ChatGPT (GPT-4o) and Perplexity (Sonar Pro) and measured each company's AI visibility rate: the share of its category's queries where it showed up. 66 of the 100 were cited at least once. 34 were completely invisible. The spread was wide — Luxent (NetSuite) topped the field at a 30% visibility rate, while most cited companies landed in the single digits. By count, B2B SaaS had the most companies cited (13 of 16); Microsoft Integrators had the fewest (6 of 14). Perplexity cited far more of these challengers than ChatGPT (62 vs 25), which tracks with its live-retrieval model surfacing smaller players that ChatGPT's trained knowledge doesn't reach.
Key Findings
34 of 100 challengers (34%) were invisible — zero appearances on either engine
Perplexity cited 62 of the 100 companies; ChatGPT cited just 25
Only 21 companies appeared on both ChatGPT and Perplexity
Luxent led all verticals with a 30% AI visibility rate, driven almost entirely by Perplexity (59%)
Microsoft Integrators had the fewest companies cited (6 of 14); B2B SaaS had the most (13 of 16)
A high score on one engine rarely meant a high score on the other: cross-engine consistency was the exception, not the rule
Top Companies by AI Visibility
Per-Engine Breakdown
ChatGPT
GPT-4o
Perplexity
Sonar Pro
Methodology
Scope & Testing Window
100 companies in the B2B Challengers (7 verticals) vertical were tested across 155 buyer-intent queries using 2 AI search engines: ChatGPT (GPT-4o, OpenAI) and Perplexity (Sonar Pro). Fieldwork dates: October 2026.
AI Engines & Configuration
Queries were run programmatically against the OpenAI Chat Completions API (gpt-4o) and the Perplexity API (sonar-pro) at temperature 0.3. Perplexity retrieves and cites live web sources; ChatGPT answered from trained knowledge without live browsing.
Geography & Account Controls
All queries were issued in English from US-based infrastructure. Results may differ by region, language, and account history. Each query was sent as an independent request with no shared conversation history, custom instructions, or memory. Both engines were queried through their standard APIs.
Query Design
Queries were designed to represent buyer-intent searches — the kinds of queries prospects use when evaluating software or service providers. Informational queries (e.g., "what is endpoint security") were excluded. The full query set is available in the downloadable CSV.
Company Selection
100 growth-stage B2B challengers were selected across 7 verticals (B2B SaaS, Fintech, HR Tech, Cybersecurity, MSP Software, Microsoft Integrators, NetSuite Integrators). Every company was new to this edition — none appeared in the Q2 2026 studies — and the panel deliberately excludes category incumbents and household names to focus on companies with room to grow their AI visibility. Each company was verified as a live, independent business matching its vertical before inclusion.
What Counts as a Citation
A company is counted as appearing when it is named as a recommendation or solution in an AI engine's answer to a buyer-intent query, or when its own domain appears in the answer's cited sources. Passing mentions in disclaimers or generic alternative lists are excluded.
Scoring Rules
Each company's AI visibility rate is the share of its vertical's buyer-intent queries where the company appeared, measured per engine and combined across both. A company counts once per query it appears in, regardless of position. The overall rate is not a raw citation count and should not be compared directly to the Q2 citation-count reports.
Repetitions & Statistical Note
Each query was run once per engine for this snapshot. AI responses are non-deterministic — running the same query again may produce different results. Because each query was run once (not repeated), these results represent a single observation per query-engine pair, not a statistical estimate. Confidence intervals and p-values do not apply. Treat the data as a structured snapshot of AI engine behavior during the testing window.
Engine Coverage Note
This study covers two AI engines (ChatGPT and Perplexity). Google Gemini, Anthropic Claude, Microsoft Copilot, and other AI search products were not included. We plan to expand engine coverage in future editions.
Research Lead
Andrew Swiler is the Founder of AnswerManiac and leads the company's primary research on AI search visibility. With a background in B2B marketing strategy and search optimization, he designed the State of AI Search methodology to give companies verifiable data on how AI engines represent their brands.
Sponsor Disclosure
This research was funded and conducted by AnswerManiac. No company featured in the study paid for inclusion or influenced its methodology or findings. AnswerManiac offers AI visibility optimization services, which creates a commercial interest in the topic studied. The raw data is published to allow independent verification.
Limitations
AI engine responses vary by session, account, and region. Results reflect a single dated snapshot in October 2026.
Only two AI engines were tested. Google Gemini, Anthropic Claude, and Microsoft Copilot were not included.
Visibility is measured as a rate — the share of a vertical's queries where a company appeared — not a raw citation count, so these figures are not directly comparable to the Q2 citation-count reports.
Each query was run once per engine. AI answers are non-deterministic, so this is one observation, not a statistical average.
The 100-company panel is a deliberate sample of growth-stage challengers, not an exhaustive census of each vertical.
Cite This Report
Andrew Swiler et al. "State of AI Search: B2B Challengers." AnswerManiac, October 20, 2026. https://www.answermaniac.ai/research/state-of-ai-search-q3-2026-challengers
Listen to Summary
Quick Stats
Open Data
Full dataset available for download. Licensed under CC BY-NC 4.0.
Request JSON or API accessIs your company listed?
Run a free AI Visibility Score to see where you stand across all major AI search engines.
Check Your ScoreWant your vertical researched next?
We build custom State of AI Search reports for specific verticals. Get the data your competitors don't have.