Analyzing unstructured data from textual information, such as consumer reviews of different services and products, is becoming increasingly important. Text sentiment analysis (SA) is often applied to transform such information into quantitative form. SA allows us to measure emotional tone reflecting alternatives evaluation. This study proposes a multistage approach to decision-making using Z-numbers based on verbal information. Initially, information on decision criteria contained in reviews is transformed into numerical form by applying SA. Next, Z-numbers are used to capture this numerical information. Finally, a decision-making method compares alternatives with Z-number-based criteria values. An example of decision-making for hotel comparison is considered.