Publicação
Application of generative Artificial Intelligence in the automation of commercial procedures in the road freight transport sector
| Resumo: | This project aims to increase the operational efficiency of the sales department at Rangel Transitários S.A., a company specialising in road transport services. To achieve this, the natural language processing capabilities of Generative Artificial Intelligence models were leveraged to automate the repetitive tasks performed by sales staff when handling transport quotation requests sent by customers via e-mail. The project began with a literature review focused on understanding the range of models available on the market and their respective implementation methodologies. This was followed by a detailed study of the department’s operations, which revealed an average e-mail response time exceeding one hour and a significant amount of time spent on data reading, extraction and validation. These factors limited the attention given to higher value-added tasks - such as customer follow-up and prospecting - and contributed to low quotation acceptance rates. The project was developed iteratively, following Scrum Agile principles, and was structured into two main phases: the development of a Minimum Viable Product - designed in accordance with core Software Development domains - and a subsequent phase focused on refining the solution in a production environment. During this second phase, access to the application was gradually extended to additional users (a total of five), with the research project concluding upon the full rollout of the solution to all salespeople in the department. The final results were highly promising. Following a fine-tuning process, the data extraction model achieved accuracy levels exceeding 85%, repetitive processing tasks were reduced to just 7 seconds and improvements are anticipated across several departmental performance indicators, including response time and quotation acceptance rates. These findings were supported by user feedback, which indicated high levels of receptiveness to the automated solution and a clear recognition of the value brought by integrating Artificial Intelligence into departmental activities. |
|---|---|
| Autores principais: | Maré, João Ricardo |
| Assunto: | Generative Artificial Intelligence Commercial Process Automation Operational Efficiency Road Transport Software Development Inteligência Artificial Generativa Automatização de Processos Comerciais Eficiência Operacional Transporte Rodoviário Desenvolvimento de Software |
| Ano: | 2025 |
| País: | Portugal |
| Tipo de documento: | dissertação de mestrado |
| Tipo de acesso: | acesso embargado |
| Instituição associada: | Universidade do Minho |
| Idioma: | inglês |
| Origem: | RepositóriUM - Universidade do Minho |
| Resumo: | This project aims to increase the operational efficiency of the sales department at Rangel Transitários S.A., a company specialising in road transport services. To achieve this, the natural language processing capabilities of Generative Artificial Intelligence models were leveraged to automate the repetitive tasks performed by sales staff when handling transport quotation requests sent by customers via e-mail. The project began with a literature review focused on understanding the range of models available on the market and their respective implementation methodologies. This was followed by a detailed study of the department’s operations, which revealed an average e-mail response time exceeding one hour and a significant amount of time spent on data reading, extraction and validation. These factors limited the attention given to higher value-added tasks - such as customer follow-up and prospecting - and contributed to low quotation acceptance rates. The project was developed iteratively, following Scrum Agile principles, and was structured into two main phases: the development of a Minimum Viable Product - designed in accordance with core Software Development domains - and a subsequent phase focused on refining the solution in a production environment. During this second phase, access to the application was gradually extended to additional users (a total of five), with the research project concluding upon the full rollout of the solution to all salespeople in the department. The final results were highly promising. Following a fine-tuning process, the data extraction model achieved accuracy levels exceeding 85%, repetitive processing tasks were reduced to just 7 seconds and improvements are anticipated across several departmental performance indicators, including response time and quotation acceptance rates. These findings were supported by user feedback, which indicated high levels of receptiveness to the automated solution and a clear recognition of the value brought by integrating Artificial Intelligence into departmental activities. |
|---|