The emergence of spatially resolved omics technologies has transformed our ability to study tissues in their native context, capturing not just what genes cells express, but where they sit and who their neighbours are. This spatial dimension has enabled questions about how cells communicate across the tissue, and how the local microenvironment shapes inter- and intracellular signalling. This talk presents a progression of work that addresses these questions at increasing modelling complexity: from the inference of cell-cell communication in single-cell and spatially resolved data, through frameworks that link intercellular signals to downstream pathway activity, to generative models that enable counterfactual reasoning about cellular states under altered neighbourhood composition.