Methodology

What this scorecard measures

AI looks for authority the same way your buyers do - in third-party credibility. This scorecard shows where that credibility exists for you and your competitors, and which of it the AI engines are actually drawing on.

Google AI Overview and AI Mode results are sampled from live searches; VAIV has no access to Google's internal data.

How answers are weighted

Answers are weighted by what B2B buying teams actually decide on: where the question sits in the buying journey (awareness, consideration, decision) and which purchase criterion it tests — risk and compliance, proof of outcomes, fit, integration, support, commercial terms and vendor viability. A decision-stage security or proof question counts roughly a third more than an open browsing question.

Missing evidence is ranked by what assistants rely on to surface a vendor: independent third-party sources, corroborated across several outlets, editorially vetted and clearly tied to your brand as an entity. Press and trade coverage, top 10 lists and analyst references outrank forums and self-published pages, which is why they sit at the top of your next moves.

The research behind it

• GEO: Generative Engine Optimization — Aggarwal et al., Princeton / IIT Delhi — ACM SIGKDD (KDD '24).Across a 10,000-query benchmark, adding citations, quotations from credible sources and statistics raised a source's visibility in generative answers by up to ~40%, while keyword-stuffing tactics carried over from classic SEO did not.
• AI Answer Engine Citation Behavior: an empirical analysis of the GEO-16 framework — Kumar & Palkhouski, Wrodium Research (2025).On B2B-targeted prompts, pages that scored highly on structure, clear entity identification and verifiable evidence were cited far more often than pages that read as promotional copy.
• Generative Engine Optimization: How to Dominate AI Search — Wang, Chen, Chen & Koudas, University of Toronto (2025).Generative engines synthesise answers from a small set of retrieved sources, so being present and corroborated across several independent third-party pages matters more than the rank of any single page.
• What Drives Citations in Production Large Language Models? — Discovered Labs — observational study of ~2M citations across ChatGPT, Claude, Google AI and Gemini.Citation frequency tracks independent, editorially-published pages with clear attribution and recency far more strongly than it tracks vendor-owned marketing pages.
• The B2B buying journey — Gartner.A typical complex B2B purchase involves 6–10 decision makers who spend most of their time on independent research and validation rather than with vendors — which is why decision-stage, risk and proof questions are weighted highest here.

These studies provide directional evidence about how assistants behave. We sample prompts; we don't census every possible query.