The New Battleground: How AI Search is Rewriting B2B Software Comparisons and Leaving Brands in the Dark

By Global Tech Desk
Published: September 2026

Introduction: When Your Competitor Writes Your Product Review

If you ask Google’s AI Overview to evaluate the reporting capabilities of applicant tracking system (ATS) giant Greenhouse against its rising rival, Ashby, the AI engine will deliver a surprisingly nuanced assessment. It might point out, for instance, that Greenhouse’s native reporting suite occasionally requires integration with a separate business intelligence (BI) tool to unlock advanced analytics.

For a prospective buyer, this seems like an objective, helpful insight. But for Greenhouse’s marketing team, there is a catch: the primary source cited by the AI for this claim is not a neutral industry analyst, nor is it a piece of documentation published by Greenhouse. Instead, the citation leads directly to Ashby’s own proprietary comparison web page.

This scenario highlights a rapidly escalating operational challenge for B2B enterprises across the globe. As artificial intelligence fundamentally reshapes how business buyers discover, evaluate, and purchase software, brands are discovering an unsettling truth: when an AI-powered search tool or conversational agent evaluates your product against a competitor, it builds its narrative from whatever evidence is accessible on the open web. And more often than not, that evidence does not belong to you.

Main Facts: The "Versus Layer" and the Crisis of First-Party Control

To understand the mechanics of this shift, a recent comprehensive study conducted by research teams at Foundation and AirOps dove deep into the software buying journey. Focusing on applicant tracking systems—a highly competitive category where prospective enterprise buyers routinely cross-shop multiple vendors—the researchers ran 20 distinct ATS buying prompts across six major AI platforms twice.

The platforms tested included ChatGPT, Perplexity, Gemini, Claude, Google AI Mode, and Google AI Overview. Across these tests, the team gathered a massive dataset of 240 distinct AI-generated answers, systematically auditing 2,695 recorded source presences to determine exactly which web pages powered the responses.

The findings expose a stark vulnerability in modern digital marketing strategies. Across six direct head-to-head product matchups (involving Greenhouse, Ashby, Lever, and Workable), the actual software vendors accounted for a mere 27% of the visible source presence in AI citations. The remaining 73% of the evidence utilized by AI models came from sources entirely outside the two vendors being compared. Furthermore, not a single one of the six direct matchups was led by a first-party comparison page from either of the competing companies.

Instead, AI engines rely heavily on what researchers have dubbed the "Versus Layer"—an ecosystem of independent comparison portals, third-party review sites, community forums, and, crucially, competitor-owned web properties. Within the ATS category, independent and commercial publishers dominated this layer, supplying 42.7% of the citations. Other software vendors contributed 20%, traditional review sites accounted for 7.4%, and community sources made up 2.9%.

For modern Chief Marketing Officers, this creates an existential risk: if a business fails to curate and dominate the narrative within the Versus Layer, competitors and third-party publishers will effectively write the definitive product specifications for them.

Chronology and Methodology: How the AI Discovery Landscape Shifted

The evolution of search from traditional keyword indexing to generative AI synthesis has been swift, catching many legacy digital marketing departments flat-footed.

For over two decades, B2B software marketing relied on a predictable playbook: optimize product landing pages for search engines, capture high-intent keyword traffic, and drive prospective buyers toward gated whitepapers or product demos. If a buyer wanted a comparison, they would navigate to dedicated review aggregators like G2 or Capterra, or search for independent blogs.

However, the mass adoption of conversational AI engines radically compressed this journey. By 2026, user behavior had shifted decisively toward multi-platform AI discovery. According to recent market data from Comscore’s Q2 2026 AI Intelligence Report, user traffic and prompt volumes have fractured significantly across platforms. While ChatGPT historically dominated the generative search landscape, its share of total AI prompt volume fell from 70% to 50% over the course of the year. Simultaneously, Google’s Gemini surged from 17% to 30%, and Anthropic’s Claude climbed from 2% to 11%.

Recognizing that these distinct platforms do not index or prioritize the web in the same way, the Foundation and AirOps research collaboration set out in mid-2026 to map the exact anatomy of B2B AI discovery. By executing hundreds of structured prompts simulating real-world enterprise software evaluations, the teams tracked how algorithms select, weigh, and display external sources.

The resulting data proved that traditional search engine optimization (SEO)—measured purely by domain authority, backlink profiles, and monthly traffic—fails to guarantee visibility in generative AI answers.

Supporting Data: Four Core Patterns of AI Source Selection

The empirical teardown of the 2,695 recorded source presences revealed four distinct behavioral patterns governing how AI engines select pages to construct product comparisons:

1. Domain Authority is No Longer a Shield

In traditional Google search, a web page published by a multi-million-dollar enterprise with high domain rating (DR) almost always outranks a smaller competitor. Generative AI, however, operates differently.

As an illustrative example, Lever published a comprehensive comparison page evaluating not only its direct rivals but also cross-shopping scenarios between third parties, such as Greenhouse and Ashby. Despite having a URL Rating of zero, just two referring domains, and a paltry eight monthly visits according to Ahrefs metrics, Lever’s comparison page was cited by AI engines in 27 separate answers. In contrast, Ashby’s own first-party page on the same subject—bolstered by a URL Rating of 19, six referring domains, and roughly 20 monthly visits—was only cited 18 times. The AI cared less about backlink prestige and more about the presence of direct, structured answers to the prompt.

2. The Funnel Progression Dictates Source Types

The nature of the sources cited shifts dramatically depending on where the prospective buyer is in their research cycle:

• Neutral Category Prompts: When users asked broad, top-of-funnel questions about ATS options without naming specific brands, broad industry listicles accounted for 51.1% of surface-weighted source presence, while dedicated comparison pages captured just 12.6%.
• Brand-Inclusive Prompts: Once a buyer introduced a specific vendor into the query, the utility of dedicated comparison pages jumped to 33.8%, while listicles dropped to 21.4%.
• Head-to-Head Prompts: For direct comparisons between two specific products, dedicated comparison pages surged to 42.1% of all citations.

3. Structural Clarity Wins Over Marketing Fluff

AI algorithms prioritize efficiency in information retrieval. Pages that surfaced buyer fit, distinct decision criteria, operational trade-offs, and clear recommendations near the top of the document were consistently favored over pages buried under marketing jargon.

For instance, an independent comparison of Workable versus Greenhouse published by Venture Harbour directly outlined target buyer profiles, key differentiators, and functional trade-offs within its opening paragraphs. This page surfaced in 13 AI answers. Conversely, Workable’s own official page for the exact same matchup—which dedicated its introductory sections to high-level product benefits and demo conversion messaging—appeared in only two answers.

4. The Competitor-Sourced Product Definition

When a software vendor fails to publish clear, accessible comparative data, AI engines fill the vacuum with whatever authoritative-looking text is available. This brings the narrative full circle: because Greenhouse lacked a robust, owned comparison page addressing Ashby, generative AI models defaulted to pulling structural evidence about Greenhouse’s limitations directly from Ashby’s promotional literature.

Platform Divergence: The ChatGPT Anomaly

While platforms like Gemini, Perplexity, and Claude rely heavily on the distributed Versus Layer of listicles, independent blogs, and third-party comparisons, enterprise technology leaders must note that ChatGPT operates under an entirely different architectural logic.

According to the 2026 discovery data, ChatGPT—which remains the most widely utilized AI conversational agent in the United States across multiple demographics—diverges sharply from its competitors. When tested with mid-market recruiting queries regarding software matchups like Lever versus Workable, ChatGPT largely bypassed third-party listicles and competitor blogs.

Instead, 77.3% of its visible source citations pointed directly to the vendors’ own foundational digital footprints: 63.8% of citations stemmed strictly from first-party product and pricing pages, while the remainder pulled from official help and support documentation. Dedicated third-party comparisons and listicles accounted for a negligible 3.3% of ChatGPT’s source presence.

This critical behavioral split means B2B marketing teams cannot apply a monolithic strategy. Winning visibility across Gemini and Perplexity requires mastering the Versus Layer, whereas securing influence on ChatGPT demands meticulous optimization, structuring, and updating of core product, pricing, and technical documentation pages.

Implications and Strategic Recommendations for B2B Marketers

The transformation of discovery from human-navigated search engines to AI-synthesized answers demands an immediate overhaul of modern enterprise marketing playbooks. Passive SEO, keyword stuffing, and relying on G2 badges are no longer sufficient defenses against competitors willing to game the Versus Layer.

To reclaim control over brand narratives in the age of generative search, industry analysts and marketing strategists recommend a three-pronged framework:

1. Audit the Versus Layer Immediately: Digital marketing teams must systematically test how major AI models evaluate their products against key competitors. Identifying which third-party sites, publisher blogs, or competitor pages currently supply the "evidence" for your product is the first step toward remediation.
2. Restructure First-Party Comparison Pages: Brands must abandon long-winded, top-of-page marketing fluff on comparison pages. To satisfy AI retrieval patterns, web copy must immediately present clear buyer fit profiles, objective trade-offs, pricing differentials, and direct decision criteria above the fold.
3. Diversify Optimization Tactics by Platform: Because platforms like ChatGPT index heavily toward deep product and support documentation, while engines like Gemini and Perplexity draw extensively from editorial and comparison ecosystems, marketing budgets must balance first-party technical clarity with aggressive third-party positioning.

As the lines between objective software evaluation and algorithmic synthesis continue to blur, one fundamental rule of modern commerce remains immutable: If your enterprise fails to supply the definitive evidence about your own product, a competitor, publisher, or algorithm will gladly write the story for you.