CampaignUS has published an in-depth analysis of how companies discuss artificial intelligence and marketing on earnings calls. The piece examines the language corporate executives choose when they brief investors on AI-driven marketing programs.

The reporting zeroes in on earnings-call transcripts — the quarterly recordings where CEOs and CFOs explain results and strategy to analysts. These calls have become a primary venue where AI marketing claims reach the market, and CampaignUS reviewed how those claims are framed.

The full analysis is available through CampaignUS.

What we know

The article's scope covers corporate communications on AI and marketing in investor settings. No specific figures, named companies, or direct quotes were carried in the syndicated headline feed that reached this desk. Mart Signal was unable to verify the underlying data points before publication.

We will update this item once the complete text and its figures are accessible.

Why earnings calls matter here

Earnings calls remain the highest-scrutiny venue for corporate claims. Statements there carry legal exposure under securities rules, which is one reason analysts treat AI marketing language in transcripts as a signal — of genuine capability, or of a company riding the hype cycle.

That distinction is exactly what trade watchers are trying to make. When a marketer says AI "transforms" campaigns, the follow-up questions on those calls tend to be blunt: what shipped, what did it cost, and did revenue move.

Mart Signal's standing guidance to readers is unchanged: discount verbs, demand numbers. Any vendor or operator claiming AI-driven marketing gains should be able to name the deployment, the spend, and the measured lift. If an earnings call does not, treat the claim as positioning, not performance.

Editor's note

This item is based on a headline-level syndication of the CampaignUS report. The source feed did not include the article body, named executives, quoted figures, or cited companies. Per Mart Signal editorial policy on unverified syndicated material, we publish what is confirmed — the existence and subject of the analysis — and flag the rest as pending.