Business processes (BPs) have become extremely complex with thousands of tasks to be completed. These processes should be modelled to manage their complexity, but deviations from these models can occur during execution. To account for the potential outcomes, models are often over-specified, but in fact, it is impossible to anticipate every scenario. Therefore, the issue of how to model the response to these problems arises. The approach put forward in this study entails the creation of a primary model with the aid of domain experts. If a deviation occurs during execution, the system searches for compatible patterns in a collaborative library of workflows, rather than relying on dedicated solutions for specific problems. Nevertheless, the main challenges lie in selecting the most appropriate set of candidates from a vast number of patterns, as well as identifying the optimal injection points on the primary model to correct the deviation. One main objective is to rapidly eliminate incompatible patterns by employing simple mathematical techniques to narrow down the pool of candidate solutions. This serves to minimize the number of patterns that should be considered where more sophisticated methods can then be utilized to rank and select the best solution from this smaller set.