In recent years, there has been huge progress in the field of Deep Learning for Natural Language Understanding. Beyond the entertainment industry, however, there are important aspects to consider for professional, AI-based voice assistants. These include the requirement that decisions are explainable and reproducible, as well as a frequent lack of valid training data. Allinga Engine therefore follows an approach based on state-of-the-art Deep Learning technologies to achieve high-performance speech understanding while at the same time reliably reproducing business processes.
Allinga Engine facilitates flexible dialog with deterministic, reliable behavior on clearly defined paths, and masters dynamic dialog elements. The latter allows it to, for example, enter unordered information into forms and automatically request missing information.
With Allinga Engine, users can use natural speech to intuitively request information from external knowledge sources and databases. Allinga Engine utilizes information in existing knowledge representations to achieve this, for example, texts such as manuals or knowledge graphs. Allinga Engine also supports the creation of new knowledge graphs by combining heterogeneous company data.
Allinga Engine can carry out individual commands and connect external services. In this way, machines can be easily controlled or data entered into ERP systems using voice commands.