Shieldan end-to-end workplace intelligence platform for compliance teams, has expanded its offering with behavioral analysis capabilities, allowing an automatic understanding of dialects, context and jargon across multiple languages.
The Israel-headquartered platform can now surveil and decipher context, dialects, semantics and jargon in Spanish and Chinese, among many other business-driving languages including the Nordic languages and all English dialects.
Through its application of artificial intelligence (AI), semantic analysis and mixed-language natural language processing (NLP) technology, Shield is able to alert, in a full linguistic context, for market abuse, risk and compliance concerns across all of its communication channels. , including voice.
Banks and businesses alike rely heavily on electronic communication channels, like Zoom voice/video or WhatsApp messages, to communicate across borders.
Despite this heavy dependency, existing market solutions have been cited as incapable of properly alerting for various risks when different dialects of the same language are used. This can be attributed to the low accuracy of the models in understanding these regional differences and accents.
The addition of behavioral analysis capabilities to Shield is forecast to improve the accuracy of transcription and market abuse detection through its multi-language models, whilst reducing operational inefficiencies of false positives.
“We’re witnessing a transformational shift in the way we communicate, and banks and businesses no longer operate in communication silos,” explains. Shlomit Labin, VP of data science at Shield. “They’re in dire need of an all-encompassing solution that they can scale to their needs, which is what Shield offers.”
Where legacy solutions fail to understand the subtlety of language, Shield’s ability to apply all of its language models to each sentence provides banks with a solution they can introduce into their existing processes amid ever-changing global compliance regulations.
“For too many banks, simple communication between branches in different countries can be a compliance nightmare, especially if those branches are speaking the same language but have different dialects, continues Labin.
“When the Mexico City branch of a bank is speaking with the Bogotá office, the resulting transcript, crucial to security and compliance, is often rampant with false positives because of the nuanced differences in dialect.
“With Shield, it doesn’t matter if the conversation is in different dialects or jumps between multiple languages, we’re able to accurately alert on it.”