Google has introduced Dialogflow Enterprise Edition (beta), a platform for building voice and text conversational applications. The Enterprise Edition expands on all the benefits of the Standard Edition of Dialogflow providing higher default quotas for greater scalability and flexibility. Google has also integrated speech recognition with Dialogflow so that developers can use the service to build voice-based applications. Voice control and speech recognition is available within a single API call (Dialogflow Enterprise API).
Dialogflow was formerly API.AI, a startup that provided APIs and other tools developers could use to add natural language processing capabilities to applications. In September 2016, Google acquired API.AI, and last month API.AI changed its name to Dialogflow to better match the company’s product offerings.
Dialogflow Enterprise Edition uses Google’s machine learning capabilities to extract pre-built entities such as numbers, time, and date. Using natural language understanding, the service recognizes the intent of the user. Dialogflow builds unique models that learn what data should be extracted and what actions to trigger. The models are based on information provided to the system such as what a user might say.
Agents can be trained to identify custom entity types, and the platform includes more than 30 prebuilt agents that can be used as a template. Examples of prebuilt agents include a car agent for interacting with a vehicle’s systems, date agent for looking up holidays and calculating the date, and currency convertor agent for getting exchange rates for major currencies. Prebuilt agents are available in more than eleven languages including English, Chinese, Spanish, French, German, and Italian.
For more information about the differences between the Enterprise and Standard versions of Dialogflow, view the Google Cloud Platform documentation.