Repo surface could name the first buyer more clearly

My main reaction is that the ambition is obvious, but the first buyer disappears. The homepage language points toward integrated marketing intelligence. The repo language points toward causal digital twin, reference implementation, roadmap, and model-choice justification. Each side makes sense on its own. Together, they make it hard to tell whether the first reader is a growth science team, a researcher, an agency operator, or enterprise marketing leadership. I’d narrow the repo surface to one first buyer and one first use case, something like: "A causal simulation stack for growth teams that want to test campaign scenarios before spending real budget." Then the uncertainty bands, counterfactuals, diffusion model, and training details work better as proof for that buyer instead of trying to define the category in the opening.