Published
AI engines do two different things that almost every visibility tracker blurs into one number. They recommend a vendor as a tool, and they cite a vendor's content as a source.
Those two things barely correlate, and the gap between them is the most useful thing in this dataset.
Across 85 category prompts run through ChatGPT, Gemini, Google AI Mode and Perplexity over the 30 days ending June 18, 2026, ZoomInfo was named just 17 times as a recommended vendor, while its editorial blog pipeline.zoominfo.com earned 2,819 citations as a source.
That made it the second most-cited domain in the whole dataset, ahead of every actual vendor. reddit.com, at 4,620 citations, out-cites every sales tool in the set.
The headline lesson: you can win AI citations as a publisher even when AI does not recommend you as a tool, and you can be a recommended tool while your product pages stay nearly invisible as sources. That is why this leaderboard reports two separate cuts, a recommended-as-vendor table and a cited-as-source table, instead of collapsing them into a single brand-visibility score the way most trackers do.
On the vendor cut, Amplemarket is the most-cited dedicated AI-SDR and sales-engagement vendor, with 2,253 citations and a 2.23% share of all citations, followed by Cognism (1,333 / 1.32%), Apollo (1,308 / 1.30%) and Unify (1,300 / 1.29%).
The single most important caveat is also the first thing you should know: Amplemarket both publishes this leaderboard and appears in it, so read it as a measured, reproducible snapshot of who AI engines cite, not as an endorsement or a ranking of product quality.
The full methodology is below, and every number is traceable.
This is not a "best tools" listicle. It does not score features, pricing or reviews. It answers two questions most rankings conflate: when an AI engine answers a sales-software question, which tools does it name, and whose content does it actually pull in as a source?
Those are different and increasingly decisive metrics, because buyers now ask ChatGPT and Gemini before they ask Google, and the names and the sources an engine returns are what shape the shortlist.
Which sales tools does AI recommend as a vendor?
This is the vendor cut: how often each tool is named as a recommended vendor across ChatGPT, Gemini, Google AI Mode and Perplexity, US, 2026-05-19 to 2026-06-18. Citation share is each vendor's percentage of all citations returned for the tracked category prompts.
The denominator includes every cited domain, Reddit, YouTube and publisher sites included, which is why even the leader sits near 2%. Read this table next to the source cut below; a tool can rank well here and barely register as a content source, or the reverse.
ZoomInfo's editorial subdomain pipeline.zoominfo.com is separately one of the single most-cited domains in the entire dataset (2,819 citations, see the source table below). The 17 here is citations attributed to the zoominfo.com product entity in the vendor cut. The full split is explained in the next section.
Source: AirOps, a third-party AI-search analytics platform that tracks which sources AI engines cite. 85 category prompts across 22 topics, US, 30-day window ending 2026-06-18. Vendor figures cross-checked against the live AirOps export on 2026-06-20.
How to read this honestly
Citation share is small for everyone. The leader is at 2.23%, not 22%. AI answers about sales software draw from a very wide pool of community sites, video, and publisher explainers, so no single vendor dominates and rank order matters more than the absolute percentage.
Citation share is not product quality. A vendor can be excellent and under-cited because it has thin content built for AI answers, or heavily cited because it publishes a large content library. This leaderboard measures AI visibility, nothing else.
Amplemarket is in its own dataset. We are first among the AI-SDR and sales-engagement vendors we track, and we are disclosing that openly rather than quietly ranking ourselves first. The methodology below lets anyone reproduce the pull and check us.
Whose content do AI engines actually cite as a source?
The most-cited domains in the same dataset tell a sharper story than the vendor table: AI engines lean overwhelmingly on community and video sources, and on publisher-grade editorial content, not on vendor product pages. The tools that win the vendor cut and the tools that win the source cut are often not the same companies.
Here is the top of the source-domain ranking, enough to anchor the contrast.
The full 29-domain leaderboard, with source-type labels and a "how to get cited" playbook, is its own forthcoming companion piece and will be linked here on publication.
Source: AirOps AI-search analytics, source-domain data, US, 2026-05-19 to 2026-06-18, verified live 2026-06-20.
The most important detail about the split: the top three cited sources are Reddit, ZoomInfo's editorial blog, and YouTube, none of which is a vendor product page.
Among vendor domains, amplemarket.com is the most cited at 2,253. The lesson for any sales-software marketer is blunt: AI engines cite content and community, not pricing pages.
ZoomInfo is the clearest case in the set, its product entity is named just 17 times while its blog is the second most-cited domain overall, and that is exactly why this report keeps the vendor and source cuts in separate tables.
Engine mix: where Amplemarket's citations come from
Where the citations come from matters as much as who gets them. The four engines are not equal contributors to Amplemarket's 2,253 citations.
Source: AirOps AI-search analytics, engine breakdown, US, 2026-05-19 to 2026-06-18. Counts are Amplemarket's citations by engine and sum to 2,253. Each engine also has a citation rate, the share of that engine's answers that cite Amplemarket at all (Google AI Mode 35.5%, Gemini 30.1%, ChatGPT 12.9%, Perplexity 10.5%); those rates describe different denominators and are not additive.
Google AI Mode and Gemini together account for roughly 77% of Amplemarket's citations (1,724 of 2,253), more than 2.5 times the combined weight of ChatGPT and Perplexity. For sales-software vendors optimizing for AI visibility in 2026, this reorders priorities.
"What sales tools does ChatGPT recommend" is the question buyers ask, but ChatGPT is not where most citations are won; the Google surfaces are. Optimizing only for ChatGPT leaves roughly three-quarters of the citation opportunity on the table.
Methodology (reproducible by design)
This leaderboard is deliberately built so a skeptic can rebuild it.
Here is everything.
The data comes from AirOps, a third-party platform that measures which sources AI engines cite in their answers (the same kind of AI-search analytics marketers use to track AI visibility). The prompt set is 85 category prompts spanning 22 topics (for example AI SDR tools, cold-email deliverability, B2B data providers, sales engagement, buying-intent signals).
These are category questions ("which AI-SDR tools..."), not brand questions ("is Amplemarket good"), so the pool is competitive by construction. The engines are ChatGPT, Gemini, Google AI Mode and Perplexity. The geography is the United States. The window is 30 days, 2026-05-19 to 2026-06-18.
On what each metric means: a citation is a source link an engine returns in its answer. Citation count is the raw total.
- Citation share is that vendor's percentage of all citations across the prompt set, where the denominator is every cited domain, including community and publisher sites.
- Citation rate is the share of an engine's answers that cite the source at all, so it uses a different denominator and is not additive across engines.
On vendor versus domain: we report both. The vendor cut attributes citations to a product entity; the source cut counts the raw domain. They can diverge sharply, as ZoomInfo shows, which is why we publish both rather than collapsing them.
For verification, the domain figures were re-pulled live from AirOps on 2026-06-20 and matched the snapshot used here. Treat the whole thing as a point-in-time snapshot; AI citation rankings shift quickly with model and search updates.
Disclosure: Amplemarket built and published this leaderboard, and is one of the ranked vendors, finishing first among the dedicated AI-SDR and sales-engagement tools tracked.
The methodology is published precisely so the result does not have to be taken on trust: the prompt set, engines, geography, window and the analytics tool used are all here, and the metric (an engine returns your link, or it does not) is mechanical, not editorial. This is a measured citation-share snapshot, not an endorsement and not a ranking of product quality.
What this measures. These lower citation counts reflect AI visibility for these specific category prompts in this window, not a judgment of ZoomInfo, Salesloft, Outreach or Gong as products.
For feature, pricing and review comparisons, see the roundups linked below.
Why citation share is becoming the metric that decides shortlists
For a decade the question was "where do we rank on Google?"
In 2026 a growing share of sales-software research starts inside an AI engine, and the engine returns a handful of cited sources. If your domain is not among them, you are not in the consideration set, and the buyer may never see your site at all. Citation share is the new shelf space.
Two structural facts in this data make the point.
First, the pool is content, not product: Reddit, a ZoomInfo blog and YouTube outrank every vendor product entity, and vendors that win citations win them with editorial content and community presence.
Second, the engines are concentrating: roughly 77% of Amplemarket's citations come from Google's AI surfaces, so an AEO strategy that ignores Gemini and Google AI Mode is optimizing for the minority of citations.
Amplemarket's own standing here is a factual data point, not a sales pitch. Among the dedicated AI-SDR and sales-engagement vendors tracked, it currently holds the highest citation share (2.23%), and amplemarket.com is the most-cited vendor domain in the set. Whether that lead holds is an open question we will re-measure; this is a snapshot, and AI rankings move fast.
How this connects to the rest of our 2026 index work
This citation-share leaderboard measures AI visibility. Two companion indices measure different things, and we keep them separate on purpose.
For a capability score across 231 features, where Amplemarket scores 219 against ZoomInfo 107, Salesloft 92, Gong 91, Outreach 80, Cognism 75 and Nooks 26, see the feature-bracket analysis. Those scores come from the separate 231-feature capability assessment, not the AirOps citation data on this page.
A companion Outbound Outcome Index, measuring open rate, bounce, AI reply uplift and consolidation from published case studies and customer testimonials, is forthcoming and will be linked here on publication. It is a separate dataset: outcomes, not AI visibility.
For feature and pricing comparisons of the individual tools named above, see our roundups: best AI SDR tools, best AI sales agents, best sales intelligence platforms, and best multichannel outreach tools. For how AI-native search is changing prospecting itself, see find leads the way you think with AI search.
We want to be precise about what is and is not new here, because category-scoped AI-visibility leaderboards already exist publicly.
Several broader B2B-SaaS brand-visibility trackers already cover the wider market, and category-level AI-visibility leaderboards exist. So being scoped to the sales category is not, by itself, what makes this different; others do that. What none of them publishes, as far as we can find, is the recommended-as-vendor versus cited-as-source split.
What is genuinely distinct here is in three respects.
First, it separates "recommended as a vendor" from "cited as a source" and publishes both; every tracker we found reports a single brand-visibility number, and the actionable lesson that you can win citations as a publisher even when AI does not recommend you as a tool is mathematically invisible in any single-number ranking.
Second, the publisher is a ranked vendor and says so; Amplemarket appears in its own leaderboard, finishing first among the tools tracked, and discloses that in the opening, the methodology and the FAQ.
Third, it is fully reproducible; the exact prompt set (85 category prompts across 22 topics), the four engines, the US geography, the 30-day window and the analytics tool used (AirOps) are all published, which are typically the inputs a black-box tracker keeps proprietary.
A leaderboard that separates recommendation from citation, that you can audit, run by a participant who admits it, beats a single-number black box you cannot.
