Imagine finding an error while developing a new product and then just clicking on your software assistant that then displays an appropriate debugging technique. Perhaps the tool also identifies the possible location of the fault. Perhaps the tool makes an automated correction to your code. That is the anticipated future of knowledge-based, artificial intelligence (AI)-supported software assistants.

This paper presents an extensive review of existing research systems that are identified as software assistants in the area of software design, construction, and maintenance. The authors’ methodology for conducting a mapping study is thorough. The decisions are well documented.

The included short history of software assistants is interesting, and the review of previous studies appears thorough. The authors identify 112 primary studies and clearly classify the capabilities of existing research systems. Included charts display the classifications well and show the current state of research. The reference list is excellent.

The discussion section looks at the many challenges and limitations of the current state of research on software assistants. These include a decrease in the number of new research articles, lack of research involving new results in machine learning (ML) and human-computer interaction (HCI), emphasis on coding issues over software design issues, and a predominance of work in Java and Eclipse. Future research directions are discussed.

This paper should interest at least two groups: software developers who are interested in trying the current tools, and researchers interested in developing the next generation of software assistants.