The objective of this article is to apply a comparative analysis of machine learning techniques to predict project sales prices for a consulting company in the South of Brazil. The company is involved in various fields such as strategy, production, quality, and innovation. Due to this diverse range of projects, the company managers face challenges in accurately determining the sales value of new projects, as they deal with different types of predictor variables such as consultant type, project type, and number of hours. Hence, there is a need to utilize a method that can predict sales values through multivariate analysis and yield results close to the company's expectations. To achieve this goal, the article conducted a literature review on two research topics: Production Planning and Control (PPC) and machine learning techniques. Subsequently, the current sales prospecting process of the company was mapped out. Data were collected, analyzed, and prepared, followed by testing and selection of the best model. Finally, the proposed improvement was discussed with the organization. The results revealed that the application of Gradient Boosting Machine (GBM) technique achieved the lowest error rate among the tested machine learning techniques. The error rate was approximately 22%, which is deemed acceptable within the analyzed segment. Consequently, this study successfully met stakeholders' expectations by demonstrating the potential of utilizing computational algorithms for demand forecasting and project pricing.The objective of this article is to apply a comparative analysis of machine learning techniques to predict project sales prices for a consulting company in the South of Brazil. The company is involved in various fields such as strategy, production, quality, and innovation. Due to this diverse range of projects, the company managers face challenges in accurately determining the sales value of new projects, as they deal with different types of predictor variables such as consultant type, project type, and number of hours. Hence, there is a need to utilize a method that can predict sales values through multivariate analysis and yield results close to the company's expectations. To achieve this goal, the article conducted a literature review on two research topics: Production Planning and Control (PPC) and machine learning techniques. Subsequently, the current sales prospecting process of the company was mapped out. Data were collected, analyzed, and prepared, followed by testing and selection of the best model. Finally, the proposed improvement was discussed with the organization. The results revealed that the application of Gradient Boosting Machine (GBM) technique achieved the lowest error rate among the tested machine learning techniques. The error rate was approximately 22%, which is deemed acceptable within the analyzed segment. Consequently, this study successfully met stakeholders' expectations by demonstrating the potential of utilizing computational algorithms for demand forecasting and project pricing.