Abstract-Background: Businesses are increasingly dependent on IT services, and providers need to deliver fast, with high quality and low cost. An incident is an event that can lead to loss or disruption of services. Incident management reinstates normal service operation as quickly as possible and mitigates negative impact to business, ensuring agreed levels of service quality. So, reduce resolution time is usually the most important goal for incidents. Aims We aim to obtain knowledge about process and identify adequate metrics for Incident Management to help reduction of resolution time. Our research questions are: (i) Which Incident Management sub-process is causing more impact to resolution time? (ii) Which metrics can be used to measure this sub-process? (iii) What actions can be taken to improve Incident Management process in order to reduce impact of this subprocess in resolution time? Method: We present a case study in a global large company that considers reduction of incidents resolution time as a goal. Results: By applying BPM, BPI and Process Mining we were able to discover the underlined process and a bottleneck for resolution time. Moreover, we proposed metrics to improve process and service quality by applying GQM+Strategies.

Empresas e sistemas de TI estão enfrentando um aumento na complexidade do ambiente caracterizada pela colaboração, mudança e variedade de clientes, fornecedores e produtos. Aplicar a técnica de group storytelling pode contribuir para a Gestão do Conhecimento da organização. A contagem de histórias traz benefícios desde a captura até fixação da informação, passando pela comunicação e entendimento dos conceitos. Empresas americanas (3M e Apple), japonesas (Sony e Toshiba) e européias (ClubMed e Océ) já utilizam esta abordagem na prática. Por outro lado, a Engenharia de Ontologias pode contribuir na melhoria da qualidade da informação e oferecer uma solução para lidar com a gestão do conhecimento de forma sistemática. No entanto, a especificação e gestão de ontologias realizadas de forma manual podem ser caras, tediosas, enviesadas e propensas a erro. O aprendizado automático de ontologias é uma abordagem que extrai ontologias a partir de dados, tanto estruturados como não estruturados (textos). Este trabalho apresenta, em fase exploratória, uma proposta capaz de especificar automaticamente os elementos que compõem uma ontologia, a partir do conhecimento tácito dos envolvidos no domínio. Um estudo exploratório foi capaz de obter os conceitos de uma ontologia, de forma automática, a partir de histórias contadas numa ferramenta de group storytelling sobre os processos de negócio de uma das secretarias de uma Universidade de Ensino Superior Federal.

Abstract. Business and IT systems are facing increasingly complex environments characterized by collaboration, change and variety of customers, suppliers and products. Appling group storytelling technique can contribute to the organization knowledge management. It brings benefits from capture to disseminating information, through communication and understanding of the concepts. American companies (3M and Apple), Japanese (Sony and Toshiba) and European (ClubMed and Océ) already use this approach in practice. On the other hand, the Ontology Engineering can contribute towards improving the quality of information and offer a solution to address knowledge management systematically. However, the specification and manually made of ontology management can be expensive, tedious, biased and pruned to error. Aiming to contribute with the management and quality of information, we explore the automatic learning of ontologies, which is an approach that extracts ontologies from data, both structured and unstructured (text). This work presents a proposal to extract an ontology from the tacit knowledge of those involved in the field. An exploratory study was able to get an ontology automatically from stories told by a group from a university department.