Propensity modelling is an AI-driven predictive technique that calculates the likelihood of a specific user taking a future action based on their historical behaviour and data patterns. In a modern sales stack, propensity modelling is used to identify which leads are "high-intent" and most likely to convert, allowing businesses to route their best resources—whether human sales reps or aggressive retargeting ads—to the most promising prospects.
This algorithmic approach to sales moves away from broad, generic outreach and into hyper-personalization. For example, if your model identifies a lead as having a high propensity to buy, you might trigger a Retrieval-Augmented Generation (RAG) agent to provide real-time, technically accurate answers to their questions. Conversely, low-propensity leads can be placed into automated, low-touch nurture sequences. This technical efficiency reduces your overall CAC by ensuring that manual labour and expensive ad spend are only applied where the probability of return is highest. It is a central component of the "Post-Purchase Flywheel," as it also helps predict churn risk and identifies the best timing for an automated upsell.