Two forces behind every adoption curve
The Bass diffusion model, developed in 1969 to describe how new products get adopted, splits adoption into two forces: an innovation coefficient (p) capturing adoption driven by external influence — advertising, media coverage, direct outreach — and an imitation coefficient (q) capturing adoption driven by social influence, where existing adopters recruit new ones by example.
Why enterprise AI skews imitation-heavy
Enterprise buyers are typically risk-averse and heavily influenced by what peers and competitors are already doing — "everyone in our industry already uses this" carries more weight than an advertisement ever could. That pushes q well above p for most enterprise AI categories: early adoption is slow, driven by a small number of genuinely innovative buyers, and then compounds sharply once a critical mass of visible adopters exists.
Reading the shape of the curve
Higher q relative to p compresses the curve into a sharper hockey-stick, with a later but steeper takeoff once imitation kicks in. Higher p relative to q spreads adoption more evenly across time, since external influence doesn't depend on how many others have already adopted. Knowing roughly where a given AI category sits on this spectrum is directly useful: it's the difference between betting a market will reward being early versus betting it will reward being ready to move fast once mass-imitation begins.
A model, not a guarantee
The Bass model is a widely used forecasting tool, not a crystal ball — it assumes a fixed addressable market and stable coefficients, both of which real markets can violate as new segments open up or competitive dynamics shift. Its real value is as a structured way to reason about adoption timing, not a precise prediction of any specific company's outcome.
🧪 Try it yourself: the Future of AI Business Lab simulation lets you experiment with everything described above directly in your browser.