StoneTemple analysts conducted a test of popular voice assistants, to determine which of them better understands the needs of people and responds to commands. In the test participated Siri, Google Assistant, Microsoft Cortana, and Amazon Alexa. The effectiveness of each assistant based on the quality of answers to over 5,000 questions.
The questions were related to news, sports scores, weather, Internet searches and other things. Asking questions and comparing the answers of the assistants, the researchers took into account and gave the program a verbal explanation or requested a response from the database, such as the Knowledge Graph draws information from third party sources, including “Wikipedia”. Also taking into account the error, when aides did not understand or misinterpreted the request.
The smart assistant was Google Assistant, which handled the response to 68.1% of the questions. 90.6% of responses functions were correct. Followed by a small gap is Cortana: she gave the answer of 56.5%, 81.9% of the cases correctly.
From Alexa the percentage of correct answers higher — 87%, however, Amazon was able to answer all of 20.7%. Not too well proved and Siri: she said 21.7% of the questions, and correctly only by 62.2%.
It is fair to say that in the tests, Siri worked on iOS 10. The release of iOS 11, which was announced at WWDC 2017, it will become much smarter and will get more functions. Siri will begin to apply the algorithms, deep learning, and her voice would sound softer and more natural. Perhaps a new test will show its best side.