This thesis was developed in collaboration with COMPANY-002, commercially known by the same coded identifier, as part of the SYSTEM-TH-001 digital transformation project. At the start of the project, the author was the company’s CIO and system architect. Informal discussions led to a company–university collaboration through an industrial doctorate, combining theoretical knowledge, academic research and business experience to build a solution tailored to the company. This collaboration made it possible to deploy research in a real business setting and strengthen the relationship between applied research and operational needs.
1.1 Context
COMPANY-002 The company. Founded at 2008 That's a niche company at the custom travel segment. Some important information from the company are at 2020 a turnover of 3 M = C and year 2023 of 5 M = C / year, with a single public place situated at Arroyomolinos (Madrid) and an opportunity capture website. On his 16 years of existence, COMPANY-002 have been transformed from a travel agency startup to one of the reference providers for this segment in Spain, competing with some and most important players
and achieve great achievements along the way as was achieved by achieving Spanish tourism quality standards " ORG-TH-010 "at 2017.
1.2 Identification of problem
A few months before the pandemic, COMPANY-002 decided to start the creation of a new system as it optimized its relationship with customers and allowed for more efficient travel agents, this renewal through scientific innovation or an interest in artificial intelligence technologies have been enabling them To achieve their business objectives. That decision was built on the idea that corporate management with a light of business intelligence processes, basically analytic reports and more complex every day by increasing data volume as well as being really far from automation. The company was seeking best adaptation to change within a market with an uncertain future that will likely converge into the online model, so as To compete on equal terms and conditions for high-level vendors. The main goal of COMPANY-002 have been focused on offering best experience to their customers and with a project as SYSTEM-TH-001 getting them every day is simply easier. Today, the automation of our business as far up to what level we wish is at hand by using artificial intelligence.
1.3 Objectives
SYSTEM-TH-001
The aim of the project is to put in place a new Intelligent Central System for Company operations by introducing innovation with its introduction into this system. Using these modules, the system is capable of adapting to changes with different scenarios and using available data.
That thesis has been developed as a strategic part of the project. SYSTEM-TH-001 and research results have been implemented in different subsystems that make up SYSTEM-TH-001 to solve business problems that resolve tools such as:
Enterprise Resource Planning (ERP): SYSTEM-TH-001 It includes a corporate management subsystem that allows the company to manage and optimize all its business processes with only one technological platform. It has, among other things and its functionality with respect to the management of travel archives or production
re financial reporting to management and business intelligence reports for best strategic decision-making.
Customer Relationship Management (CRM): SYSTEM-TH-001 It implements an analytical subsystem of customer relations management that allows the company to optimize interactions with current and potential customers. The CRM develops opportunity and client management functions, guided by an analytic scouting model based on the company's historic architecture.
eCommerce: The future subsystem for actual implementation of the business electronic, currently continuing development research phase and aims at selling tourist services through internet with functionality or to sell all kinds about travel from simpler up To more complex as those combined. from web services, thus obtaining prices dynamically and in real time.
The project focuses on applying innovation to the company's various software pieces by introducing an intelligent feature, enabling its system as a decision support and knowledge-based information from historical data. The introduction of these intelligent functionality aims to optimize adaptation and change in the business so that his system remains but with an optimal performance for new scenarios. At the start of this project, it was expected that its introduction would lead to a quick increase in sales and an improvement with respect To quality after-service recommendations as basic pillars for today's commercial strategy.
A thorough research has been undertaken with the use of different techniques for artificial intelligence and information about his or her historic data, as a means to offer knowledge that will allow an improved efficiency.
Objectives
Three main objectives are put forward: the first goal is to define an intelligent architecture for a new system SYSTEM-TH-001 Based on (Multi Agent System), the second is construction of analytics that have been built up using an automatic learning model with intelligence generated by knowledge extracted from company's historic data. The third goal was not first defined as it appeared during his thesis development and aims at defining how to start a project path The construction of a virtual travel agent, for this purpose have been investigated (Natural Language Processing) and text mining technologies using the Support Vector Machine.
The challenges to achieve these objectives have been put into practice at the same time as different research lines are complementary. To restrict the scope of research, for its first aim it was decided to focus on finding a solution with respect and how work loads have been distributed at system level. The aim was to investigate the use of PLN technologies and be prepared for a future stage in which virtual agents are an application solution.
During the development of activities to achieve first goal, an intelligent architecture for system was defined SYSTEM-TH-001 with a Multiagent System (MAS), this approach was chosen for them to manage and distribute work loads effectively in an enabling environment such as these.
The second goal is achieved by applying a solution to the business problem from calculation of propensity and purchase, through an identification with potential customers that are key for commercial activity. To this end, we have focused on the recognition of customers' characteristics and an analytic model for scouting with historical data that can be updated over time. That allows to optimize the available time of work for travel agents, so that best financial performance can be achieved from commercial operations. Each commercial agent will have a binary classification of every business opportunity and this has an important impact on the decision-making by his or her travel agents, enabling them to get better focused with their conversion rate increased. To achieve this goal, an automatic learning model (ML) has been put into place with the aim of defining purchase propensity using analysis and treatment as a basis for variables that define different client typologies. The model has been developed with an overview of a final production environment that will be followed by incremental update as it should adapt to new daily data obtained from the system. To sum up, the goal is to develop a score model that will be useful and efficient with high evolution capacity enabling an improved sales process.
The third goal that had been established during the main research of his thesis was to define how a virtual travel agent could be put into practice and have used technologies for Natural Language Processing (PLN) or text mining. To achieve the goal, we have created prototypes of virtual travel agents for evaluating best implementation in next versions. SYSTEM-TH-001. A text's gender has been classified with the use of SVM and this can have an opportunity to re-orient virtual travel agent talks as a result.
a significant improvement in the available knowledge of our virtual travel agent.
Innovation
The whole project has been developed with strict use of free software. The system's implementation SYSTEM-TH-001 That's how this means a guarantee of flexibility and cost reduction while keeping the company independent from other private software providers. That's why all the benefits of using free software have been used and has allowed a company to enjoy high degree with digital sovereignty. The system put in place is robust and adapted to the needs of our business, with a 3 computer expert team including his own. With a well-defined strategy, SYSTEM-TH-001 It could become a clear example of innovation and efficiency, worthy to be studied in an advanced world dominated by proprietary software.
1.4 Structure of the document
The structure of this document is as follows: the Introduction addresses its context, challenges and objectives with respect to innovation. The Section 2, entitled: IA's state of art and includes key concepts such as automatic learning with knowledge management systems. The chapter 3, Management of tourist services and development SYSTEM-TH-001 and explain his architecture, the modules that make up him or her. The Section 4 To measure the propensity to purchase, it analyzes interested analytical processes including using natural language processing tools and virtual agents. The Section 5, SVM adaptation. It's about optimization and experiments performed for textual classification based on gender The Section 6, Conclusions and information about the results obtained. Also, references and appendices are included that quote the sources or resources used throughout this document with additional indexes of figure (s).