THE AI EARNINGS CALL INDEX — Which Public Companies' Earnings Calls Does AI Actually Cite?

Scoring the S&P 100 on Earnings Call AI Retrieval

2026 Virgo PR Original Research | July 2026

This study is part of the Virgo PR research franchise on how AI engines shape capital markets. See also: The Chatbox Is the New Broker | The AI Short-Seller Index | The Disclosure→AI Comprehension Gap | The Analyst Disappearance

Executive Summary

Which Part of the Earnings Call Does AI Actually Cite?

This research scores the S&P 100 on whether ChatGPT, Claude, Gemini, and Perplexity cite the earnings call transcript, the 10-Q, the press release, or the analyst note when answering investment questions. We tested 400+ prompts across four engines to build a 100-point Virgo Earnings Call Citation Score.

Key Finding: The earnings call transcript is the single most-cited source — at 61% primary citation rate. CEO prepared remarks account for 74% of all transcript citations. The first 90 seconds of the CEO's opening determine whether the earnings call shapes the AI answer or gets buried.

The Retrieval Hierarchy

When asked "What did [Company] report this quarter?", AI returns sources in this order:

• Earnings Call Transcript — 61% primary citation

• Press Release / 8-K — 54%

• Company Investor Relations Website — 47%

• Business News Coverage — 41%

• 10-Q / Annual Filing — 38%

• Sell-Side Analyst Reports — 22%

The earnings call dominates because it is the richest source of current, quotable, entity-dense language. CEO prepared remarks are conversational, specific, and structured in a way engines can parse. The 10-Q is technically authoritative but linguistically dense — engines cite it for financial data but paraphrase the earnings call for narrative.

Three Surprise Findings

1. Shorter CEO Openings Outperform Longer Ones.

CEOs who open with a tight 60–90 second summary — revenue, guidance, one strategic priority — generate higher citation rates than CEOs who deliver 8–10 minute prepared remarks. Engines extract the signal from concentrated language. Diluted language produces diluted citations.

2. Founder-CEOs Are Cited at 2.3x the Rate of Professional CEOs.

Founder-CEO language is more distinctive, more opinionated, and more entity-rich. Professional CEO language trends toward regulatory caution and consensus framing. Engines cite distinctive language; they summarize consensus.

3. Long Q&A Sections Reduce Citation Quality.

When the Q&A runs 45+ minutes, engines begin cross-contaminating analyst questions with CEO answers. The result: AI answers that attribute analyst speculation to company guidance. Shorter, tighter Q&A produces cleaner AI retrieval.

Sector Rankings

Technology: 74/100 — Best-performing sector. Tech earnings calls are entity-rich, product-specific, and heavily covered. Apple, Microsoft, Nvidia lead.

Banks / Financial Services: 68/100 — Strong on financial data accuracy. Weaker on narrative — bank earnings calls tend toward regulatory language that engines flatten.

Healthcare / Pharma: 63/100 — Pipeline updates and clinical data citations perform well. But regulatory hedging language ("we believe," "subject to approval") confuses engines.

Consumer / Retail: 58/100 — Same-store sales and guidance language cites cleanly. Brand-level commentary gets diluted across competitors.

Industrials: 52/100 — Lowest-scoring sector. Industrial earnings calls are technically dense and linguistically uniform across competitors. Engines struggle to differentiate.

The AI Earnings Playbook

Rule 1: Front-load the first 90 seconds. Revenue. Guidance. One strategic headline. This is the retrieval anchor.

Rule 2: Use named entities. "We signed a $40M contract with Lockheed" cites. "We secured a significant defense contract" doesn't.

Rule 3: Keep Q&A under 35 minutes. After that, cross-contamination risk rises sharply.

Rule 4: Publish the transcript within 24 hours. Freshness matters. Engines that index the transcript first set the narrative.

Rule 5: Monitor AI answers within 48 hours of earnings. If the engine got it wrong, the correction window is narrow.

Conclusion

The earnings call is the single most powerful AI retrieval surface a public company controls. The companies that treat it as a communications asset — structured, entity-rich, front-loaded — own the AI answer. The companies that treat it as a compliance obligation lose the narrative to whoever files first.

IR is becoming AI Relations. The earnings call is where the answer gets written.

---

The Virgo PR Research Library

• The Chatbox Is the New Broker — The flagship microcap AI retail investor study. Virgo AI Visibility Score. Top 25 Most AI-Visible Microcaps.

• The AI Short-Seller Index — How AI engines retrieve and amplify short-seller claims versus company responses.

• The Disclosure→AI Comprehension Gap — How SEC filings get misread by AI engines.

• The Analyst Disappearance — Sell-side research is being displaced from inside out.

About Virgo PR: Virgo PR is the boutique AI Communications research firm. We produce original studies on how AI engines shape capital markets, investor relations, and corporate reputation.