With SPEXIAN we consistently implement technologies from Artificial Intelligence (AI), Machine Learning and Deep Learning. Regardless of the industry our customers come from, we have the right solutions for everyone. Much more effective is the analysis of the converted and uploaded media on our platform. Here is an extract.
SPEXIAN is in the development stage and many customers give us concrete application examples for their own use. CLX tries to supply these customers with the right product.
Even though MIFID II is not an urgent problem at Swiss banks, solutions are sought after by EU banks. MIFID II prescribes, that telephone and face-to-face conversations of any client action (for example, share purchase) must be traceable. E-banking applications can transcribe the conversations into text, thus allowing a customer to proof-read and confirm them via tool.
From a compliance perspective, it is more interesting to report a misconduct by an agent or a customer (Dashboard Compliance Manager). These and other tools or analytics can be configured and displayed by the customer as widgets.
When requested, emails, SMS or red/green warning messages on compliance violations can be sent to responsible people.
Today, recordings of the call centers cannot be fully analyzed to extract business or performance related information. However, with SPEXIAN it is now possible to analyze the mass of audio data and extract important statistical information that help clients improve their business.
The next wave of innovation in a CLX Banking APP. The customer can interact with the app through oral speech. This in return allows the customer to perform certain actions directly on the bank account by giving the app voice commands (display the data from the bank account, make money transfer between accounts, etc.). These commands are transcribed by SPEXIAN, where it understands what a customer said and, as a result, executes the commands in the app.
Clx applied for a joint research project (InnoSuisse grant) to further improve the speech recognition and natural language understanding technologies for the German language. Title: SM2: extracting Semantic Meaning from Spoken Material (practical implementation expected in 2019).
Currently English, French and High German are available (Swiss German can be available on request).
The most frequently asked question was: Can other customers benefit from various speech recognition improvements?
Answer: Evgeny Bogdanov, Chief Developer at CLX:
«Yes, on our SAAS solution all improvements of Acoustic and Language models are merged together. Our on-premise customers receive a quarterly update of the licensed language model. This is a component of the Artificial Intelligence Engine to continuously improve itself by learning from more and more available data.” »