Language planning in higher education has traditionally been approached as a policy-driven, human-centred activity oriented towards access, equity, and multilingual inclusion. In digitally mediated universities (institutions where teaching, learning, assessment, research, administration, and governance are systematically shaped by digital platforms, data infrastructures, and algorithmic systems that intervene between people, practices, and decisions) language planning increasingly unfolds within complex socio-technical systems, including AI-enabled decision-support tools. This paper argues that the most significant site for meaningful human–AI partnerships in institutional language planning is neither the macro level of policy discourse nor the micro level of classroom interaction, but the meso level of institutional practice. Drawing on Van Dyk’s (2026) meta-framework for language planning in higher education, the paper foregrounds the meso level as the organisational arena where policy intentions are translated into structures, practices, and relational processes. Within this framework, the macro level provides normative and regulatory direction, the micro level reflects lived pedagogical and student experiences, and the meso level functions as the mediating space in which human–AI configurations are designed, enacted, and governed. Using insights from a national language resources audit in South African higher education as an illustrative Global South case, the paper demonstrates how AI-supported tools can enhance meso level functions such as resource mapping, capacity-building, monitoring, and institutional sense-making. At the same time, the analysis shows that these processes cannot, and certainly should not, be reduced to technical optimisation. Language planning decisions remain value-laden, context-sensitive, and socially negotiated. The paper therefore proposes a human-centred meso level model in which AI functions as an augmentative partner that renders linguistic complexity visible, while institutional actors retain responsibility for ethical interpretation, trust-building, and inclusive decision-making. In line with EDEN 2026’s focus on human–AI partnerships, the paper contributes a meso level perspective on ethical governance, institutional capacity-building, and inclusive digital futures