Artificial intelligence (AI) is becoming a prominent topic in corporate communication, yet investors face substantial uncertainty about its economic implications. We investigate whether and to what extent financial markets respond to firms’ managerial communication about AI during quarterly earnings calls. Drawing on signaling theory, we conceptualize AI talk as a rhetorical signal conveying forward-looking information about firms’ strategic orientation toward AI under conditions of high uncertainty. Using S&P 1500 earnings call transcripts from 2010 to 2024, we develop an AI dictionary to measure the incidence and intensity of AI-related communication and document its evolution, thematic content, and market impact. Across multiple empirical approaches, we find evidence that AI talk is associated with higher cumulative abnormal returns, controlling for firm characteristics, earnings surprises, industry effects, and linguistic features. Portfolio strategies formed on AI talk yield positive and significant risk-adjusted excess returns over the subsequent quarter. We further show that market responses to AI talk are moderated by firms’ R&D intensity and operational efficiency, with stronger reactions when these attributes provide clearer cues about firms’ innovation and efficiency profiles. Notably, firms with high R&D intensity and high operational efficiency elicit the strongest positive reactions, whereas firms with low R&D intensity or operational efficiency exhibit smaller but positive responses. In contrast, firms with moderate R&D intensity or operational efficiency generate muted reactions. Overall, these findings show that AI talk functions as a consequential rhetorical signal with economically meaningful stock price effects.