Helping sales managers to understand their business environment Germ´ an S ´ anchez-Hern´ andez GREC–ESADE Ramon Llull University german.sanchez@esade.edu N´ uria Agell GREC–ESADE Ramon Llull University nuria.agell@esade.edu Juan Carlos Aguado GREC–UPC BarcelonaTech juan.carlos.aguado@upc.edu M` onica Casabay ´ o GREC–ESADE Ramon Llull University monica.casabayo@esade.edu Abstract Within a business framework, one of the challenges of marketing and sales departments is to ensure that sales managers of the company have updated infor- mation about their customers’ interests. In order to help commercial representatives in their day-to-day work, it is important to offer them technologies able to identify opportunities for each of their assigned points of sale. Such technology should be based on proactive and context aware systems, i.e. these systems should automatically warn sales managers when an op- portunity in a close point of sale is detected. In addition, these systems should be personalised (to the profile of the commercial representative) and should show adaptive features, i.e. providing an adequate response to changes in the environment (such as the adaptation of thresholds according to some economic data). Finally, these technologies should be naturally integrated and embedded into the devices used by sales managers to interact with the information system of the firm, i.e. employing both digital and physical sensors (by continuously taking into account updates of the sales data for ad- justing the definition of opportunity, or by detecting the location of the sales manager for the system to provide information about the nearest shops). Having this in mind, in this paper a system allow- ing the commercial representatives to obtain an au- tomatic description of the different groups of points of sale is introduced. This will permit the sales manager to rely on a quick and updated overview of the environment in which s/he is placed. Specif- ically, the project refers to the design of a natural language generation (NLG) system to qualitatively describe the most important characteristics of each class, cluster or segment of points of sale previ- ously defined. An adaptation of a generic archi- tecture for data-to-text systems consisting of four stages is proposed. It includes the detection of the most relevant patterns of the data and the definition of a grammar that generates the natural language description of the considered clusters. A case study addressing a challenge in a marketing environment is included. The study takes place in a business-to- business (B2B) environment, in which a firm dis- tributes its products via other firms. $,7$  