As large language models (LLMs) become increasingly integrated into workplaces, understanding how individuals interact with them is critical for organisational performance (Fiegler-Rudol et al., 2025). Emerging research suggests that anthropomorphising behaviours, such as politeness, can influence LLM responses (Yin et al., 2024). While personality traits and communication styles are known to shape human–human interactions (Everson, 2025; Sims, 2016), their role in human–LLM communication remains unclear. For instance, traits such as extraversion, agreeableness and neuroticism are associated with linguistic behaviours including politeness, word count, and pronoun use (Koutsoumpis et al., 2022), but whether these patterns hold in LLM interactions in workplace contexts is unknown. This study investigated whether personality traits and communication styles relate to anthropomorphising linguistic features (politeness, pronoun use) and word count across human–human and human–LLM workplace interactions, and whether these linguistic features correlated across contexts. It was hypothesised that (1) linguistic behaviours would positively correlate across interaction types, and (2) personality and communication styles would correlate with linguistic features in both contexts. An English-speaking sample of working adults (N = 335; 53.7% male; Mage = 40.6, SD = 12.8) were recruited using Prolific. They completed the Big Five Inventory-2 (Sotto & John, 2017) and Communication Style Inventory-Brief (Diotaiuti et al., 2020) Scales, as well as human-email and LLM-looking workplace scenario vignettes. The first hypothesis was supported as linguistic behaviours positively correlated between human-human and human-LLM interactions. However, the second hypothesis was not supported, as agreeableness, extraversion, neuroticism, and expressiveness did not correlate with linguistic features. These findings suggest that while communicate patterns extend from human-human to human-LLM contexts, they are not strongly driven by stable individual differences such as personality or communication style. LLM interactions may instead be shaped more by contextual or task-related factors than by dispositional tendencies, suggesting that established personality–language relationships cannot be assumed to generalise to human–LLM communication. Implications for research and practice are proposed.