Research & Development
AI assistants
Our AI virtual assistants can engage in a dialogue with the customer, categorise the problem and then recommend a solution. If necessary, they can forward the customer to a dedicated customer service agent, providing the agent with the content of the conversation so far in text form, along with an indication of the customer’s mood category. The biggest advantage of our systems is that they can eliminate customer service waiting time almost completely.
OUR AI ASSISTANT SOLUTIONS
AI assistants
Our AI assistants are best described as a two-component artificial intelligence solution. The two main components of our AI assistants are:
(1) An RNN (recurrent neural network) algorithm – typically LSTM or GRU – which is responsible for providing context-based dynamic communication capability, i.e. the function of asking and answering.
(2) An Attention Mechanism-based, Transformer algorithm (a Natural Language Processing (NLP) algorithm), typically an enhanced version of BERT – the NLP algorithm developed by Google -, which is responsible for processing and interpreting spoken and/or written text.
Our AI assistants are able to operate in both voice (phone) and text (online chatbot) formats, and can receive and handle incoming requests. Our AI assistant solutions are able to communicate autonomously with the customer, understanding their questions and answers, and are able to maintain the conversation in context. Our systems use a speech-to-text algorithm to convert the content of the conversation into text, which is then stored in both text and voice formats, in compliance with the relevant legislative requirements (e.g. GDPR). If the customer’s problem requires the involvement of a personal customer service agent, the AI assistant will provide the agent with a text transcript of the conversation so far, so that they can continue the problem-solving process as efficiently as possible.
One of the key technical benefits of the design of our AI assistant systems is that after initial training, they can be used in a live work environment. This is because our AI assistants are self-learning systems, which means that they evolve and refine their own functionality independently with each customer interaction. The initial training of our systems is targeted at the client’s area of expertise and the problems they face, ensuring that the self-learning system will evolve in the right context. Another important advantage of our AI assistants is that they are multilingual, thanks to our BERT and mBERT (Multilingual BERT) text processing algorithms, making them ideal for international companies or local companies with many foreign language clients.
Independent and context-dependent, dynamic communication
Spoken and text-based communication skills
Self-learning systems, with continuous improvement
Systems capable of multilingual operation
Excellent in a customer service environment
OPPORTUNITIES FOR FURTHER DEVELOPMENT
Sentiment models
One of our most important current research areas for the further development of our AI assistants is the modeling of the emotional state of the customer during conversations and its implementation in the dynamic design of communication.
Using properly trained and enhanced BERT and mBERT-based models, not only the processing and interpretation of heard and/or written texts can be achieved, but also the modeling of the emotional content of individual texts and dialogues can be developed. The virtual assistant is thus able to dynamically assess the current sentiment of the conversation for each customer and react accordingly, and, in case it has to hand over the customer to a personal customer service agent for any reason, it can inform the colleague of the customer’s emotional state, in addition to the content of the conversation so far.
Our researchers hypothesize that emotional modeling can not only improve efficiency when handing over to a customer service agent but also when communicating with the AI assistant itself, as knowing the sentiment of the conversation can lead to a more accurate understanding of the information, more successful customer handling, and more effective communication, which can ultimately increase customer satisfaction.
Dynamic modelling of the sentiment of a conversation
More accurate understanding, more effective communication
Better customer management, higher customer satisfaction
Research and Development with Trilobita
Phases of our R&D projects
R&D Project planning
In the R&D project planning phase, we help our clients find the most optimal use of resources. We prepare the financial and technical design of the project and prepare the proposal for the selected funding scheme.
Applied research
In the applied research phase, we prepare the necessary research plans. We carry out and document the series of experiments based on our research methodology.
Evaluation of research results
We evaluate the results of the series of experiments using various data analysis methods and prepare the research summary document.
System Planning
Based on the research results, we design the systems for our customers. We use our own system design methodology and tools for the planning.
Development and testing
Our development methodology combines elements of classic waterfall and agile methodologies, flexibly adapting to the needs of the given client and project. The efficiency of our development and testing work is further enhanced by a number of our already tested, ready-to-use system modules.
Support
After the completion of our R&D projects, we always provide follow-up and support services to our customers for the solutions we have delivered. Our goal is to establish successful, long-term partnership with our clients.
We believe that every hour spent on design pays off many times over in the implementation and roll-out of our systems. Our ergonomically designed user interfaces provide our customers with a new user experience and ease of use.