A bridge between breakthrough science and the industry by transforming the latest developments into value-generating solutions: MindTitan

The number of artificial intelligence (AI) projects is going to increase by double in leading organizations, Gartner reports. However, for businesspeople, it can be unclear where to start those projects as well as what to consider before taking them on. Overall, MindTitan platform supports 157 languages, including all the European languages, as well as Arabic and Turkish. For higher performance, MindTitan fine-tune the final language model based on your data, as this enables AI to achieve 90%+ classification accuracy. The chatbot has built-in integration API capabilities, meaning you can run any API queries to your background or CRM systems as part of the chatbot Solution Flows. In cases where APIs are non-existent, the chatbot can run RPA scripts to achieve the same effect. The RPA scripts can be deployed both on the cloud and on-premises. Depending on the complexity of your system and the number of solutions that you would like to automate it can take from 4 weeks to 4 months. The time is mostly dependent on the number of custom integrations that are required. Simple questions can be automated in several days, whereas longer solutions require weeks. MindTitan’s onboarding and customer success team ensures smooth and fast onboarding, guiding you on each step.

Titan Chatbot is the first truly intelligent chatbot that can automate not only chats but also calls. Titan Chatbot platform enables the automation of highly complex conversations and customer support use cases. This is achieved in tandem by the machine learning capabilities and the Solution Flow Management (SFM) tool.  SFM follows the business process modelling notation (BPMN) to allow for describing virtually any kind of process. It is adjusted for the TitanCS chatbot to enable conversational techniques, Application Programming Interface (API) integrations and custom Python code for the most complex logic. Titan Chatbot is built using the latest natural language processing models and made by customer service experts for customer service professionals. Cooperation with the biggest Telecommunication companies like Elisa and Veon helped to shape the business process. The TitanCS Chatbot is GDPR compliant and is physically located in the Google or Amazon cloud closest to the client infrastructure. In the case of European deployments, the chatbot is deployed at an EU-based cloud provider, again closest to client infrastructure. If required, the bot can also be deployed on-premise for the client. The link between the bot instance (has a static IP) and client backend systems is protected by VPN or firewall rules, depending on the client’s preferences.

Make the onboarding of your agents faster and the quality of responses better

On average, customers wait 2.4 minutes on the line before they are connected to a customer service representative. Only after reaching the agent, customers start explaining their issue, which might lead to a further call transfer to another department. Button IVR is also misleading as in 30% of cases customers press the wrong button, simply because the traditional IVR system was designed for button phones and it’s not so comfortable for smart phones, meaning there are quite a lot of people who get confused about how to use it. This leads to a long journey from one agent to another and all over again. In the end, it all increases the abandoned call rate, frustrates customers, and actually costs companies around 2% of potential extra sales. Conversational IVR uses an enhanced company-specific speech recognition model and artificial intelligence that enables clients to describe issues with their own words and get forwarded to the right agent instantly. Conversational IVR is an AI solution that automates your call center and enhances your representatives to serve inbound contacts more effectively. Instead of listening to a bunch of options in an IVR and pushing buttons when calling customer service, Conversational IVR simply asks the person “What can I help you with?”. The customer describes the problem in their own words and the AI forwards the call to the appropriate specialist.

Allow your customers to describe their issue in their own words

People tend to get lost in traditional IVR systems where they need to press buttons to go through predefined selections. For some people it is difficult to understand the logic and they may lose their patience or make a simple mistake. This all leads to wrong routings and an increased abandonment rate. AI will minimize these problems by understanding to which department the call should be transferred. This component interfaces with call center software. It behaves as an ordinary software phone and communicates with the call center software over the SIP protocol. As this is the component that handles the call sessions, it also coordinates everything that happens during the call, such as routing the sound to the ASR component, sending the text to the classifier and playing prompts and responses, forwarding calls, requesting information from the client’s systems, and performing other actions as required. The AI component consists of speech-to-text and intent classifier. The speech-to-text system is a real-time speech recognition system which converts sound to text. The classifier deals with intent classification – using the voice and text it helps to understand why the customer called so that we would know what needs to be done with the specific call. The identified intent is returned to the coordinated component which handles the call according to the defined business logic.

Kristjan Jansons, Co-founder and CEO

“We believe that AI will play a huge role in an ever more competitive future of how products are made and services are being provided as data enables to provide insights and optimizations which would be difficult for humans to grasp.”