Periodic Reporting for period 1 - OmniBot (OmniBot: Artificial Intelligence and Natural Language Processing Bot platform)
Periodo di rendicontazione: 2018-05-01 al 2018-10-31
Understanding natural language has been hugely challenging for Machine Learning (ML) algorithms. While certain ML technologies work well for specific subtasks (e.g. voice-recognition, machine-translation), discerning the accurate meaning from conversational text has continued to be elusive even for the largest technology companies. Chatbot developers have reacted to the challenges often by restricting the users’ degrees of freedom by offering them menus or highly specific dialog flows. Facebook even enabled Chatbots without any possibility to “chat”, i.e. just buttons to click on. While such “Chatbots” may still have some handy uses, they do not provide the intended conversational experience.
The OmniBot solution advances state of the art Artificial Intelligence technologies and Natural Language understanding (NLU) and facilitates building conversational interfaces into any chatbot application that uses voice and text. Our advanced chatbots and voice enabled virtual assistants are easily trained with just a few training examples and an easy to use GUI based dialog flow Bot-Builder. Also, Bot skills or tasks can be quickly trained and shared for re-use and customization. Every Bot comes with human supervised machine live-learning capabilities, ensuring that the Bot learns and improves its contextual understanding very quickly. This Human-in-the-Loop machine learning functionality represents an important state-of-the-art innovation.
OmniBot’s innovation addresses European business challenges, as it creates a new technology concept that will help European businesses resolve difficulties relating to digitization, customer satisfaction, competitiveness and growth. The feasibility study as part of the Phase 1 project helped to validate the viability of OmniBot technology and the market acceptance and future market potential of our solution. Phase 2 project promises a rapidly achieving production capacity and commercialization of the OmniBot solution in our target markets.
The OmniBot solution advances state of the art Artificial Intelligence technologies and Natural Language understanding (NLU) and facilitates building conversational interfaces into any chatbot application that uses voice and text. Our advanced chatbots and voice enabled virtual assistants are easily trained with just a few training examples and an easy to use GUI based dialog flow Bot-Builder. Also, Bot skills or tasks can be quickly trained and shared for re-use and customization. Every Bot comes with human supervised machine live-learning capabilities, ensuring that the Bot learns and improves its contextual understanding very quickly. This Human-in-the-Loop machine learning functionality represents an important state-of-the-art innovation.
OmniBot’s innovation addresses European business challenges, as it creates a new technology concept that will help European businesses resolve difficulties relating to digitization, customer satisfaction, competitiveness and growth. The feasibility study as part of the Phase 1 project helped to validate the viability of OmniBot technology and the market acceptance and future market potential of our solution. Phase 2 project promises a rapidly achieving production capacity and commercialization of the OmniBot solution in our target markets.
During the Phase 1 project, in 2018, we have organized customer pilots with Volkswagen, Netcall Telecom, Bigtincan and IBB in order to gain a complete understanding of the OmniBot solution’s technical acceptance, including: demand, willingness to pay, pricing validation, and expected conversion rates. All pilots have been individual projects and each of them had very unique functional requirements and objectives. These four projects were chosen as pilots because of their very different set-ups, needs and feature requirements, which help to validate the commercial flexibility and feasibility of OmniBot’s technology. Feedback to and reception of our technology and platform (users’ feedback) has been sincerely positive throughout all three projects. Testing within each individual pilot project team has finished and the solutions are at various stages of production deployment and integration and will be evaluated during next stage broader deployment environments. Willingness to pay, price validation, and expected conversion rates were demonstrated in each particular case. Also, the collected pilot data and user experience reports provided us with sufficient data to create a development plan towards TRL9 and also generated relevant technical and user-related content which has served to help us and our pilot partners build showcase model solutions (use-case) which are currently being used to approach new customer opportunities. The Phase 1 project also helped to validate the business concept by demonstrating the market need for our product as well as evaluate the prospects for European growth and international expansion, consistent with the company’s commercialisation plans. OmniBot has also undertaken several complementary actions in order to maximize its market success, by defining and implementing the communication needs (product positioning and marketing message) to be used during the commercialization phase, thereby creating the basis for successful product launch.
OmniBot provides best value for money solutions through 6 highly differentiated value propositions, including: 1. Integrated voice options: OmniBot is the only European based conversational AI vendor that owns its own speech technology. Additionally, OmniBot supports the integration of any 3rd party speech technology allowing full deployment flexibility. 2. Training data is optional: NLP models are quickly trained with a few manually defined training samples. Bots built without pre-existing training data immediately have good accuracy which is then easily improved through the Human-in-the-Loop tool. 3. Advanced Interactive Bots: OmniBot is among the most customizable platforms in the industry (as validated by Deutsche Telekom & others) and includes APIs, SDK and programing tools optimized for advanced customization and automatization. This allows even the simplest of Bot dialogs to drive the most complex client interactions and service access. 4. Modularity: Bots are built with custom or pre-defined Bot Modules (similar in concept to Amazon Skills). Each Bot typically has multiple modules which can also be reused and customized for use in additional Bot use cases. 5. Ecosystem platform: For larger enterprises and OEMs, OmniBot provides an attractive low cost of ownership alternative to the current market players. Companies can build, integrate and manage a complete Bot framework one time and then leverage the same framework to support any user/client channel or company touch points, including: IVR or telephony, Web browser or native mobile apps, SMS, IM, Alexa, Google Home, Slack, etc. 6. Enterprise class features: OmniBot is designed for large collaborative teams and has many features to help manage and control projects and project teams, while also providing tools for business and technical people to plan, design, test and manage Bots across enterprise wide applications.