An "elevator pitch" is a brief, persuasive speech that an experience seller can use to attain the attention of a prospective client. Unfortunately, when selling complex enterprise products and solutions, there is no one pitch that works for all customers. To craft a good pitch, a seller must study a large amount of documentation, including product descriptions, client references, and use cases. Leveraging experience developed over the years, sellers then determine which marketing message will work best with a client. The goal of our research is to automatically create knowledge snippets from a large set of enterprise documents that can be used in elevator pitches. We refer to these snippets of text as points of view (POVs). Our method is based on natural language understanding (NLU), clustering and ranking techniques where the most relevant and informative content are selected as POVs for a given product. In addition, our approach is tailored to create POVs for a given aspect of the product, like the business challenges or the benefits of deploying the product.