TECHNOLOGY

Nexidia’s award-winning, patented Phonetic Search Engine (PSE) technology enables audio search by using phonemes - the smallest unit of human speech. The Nexidia technology supports all audio qualities and audio variances such as a speaker’s age, gender, accent, language and dialect. Searches using phoneme pattern matching are executed on all words and phrases including:

• blended words
• proper names
• slang
• code words
• nonstandard grammar patterns
• ad-hoc use of different languages

Using the PSE, the audio is indexed or made searchable at over 83 times real time on a single CPU. This audio is searched for key terms and phrases related to the business drivers of the organization. The results of these searches are written to a database and exposed to the users in a flexible, sophisticated reporting view. Here the user can ‘cut & dice’ the results in order to get multi-dimensional views of the intelligence contained in their media files.

The Nexidia server ingests audio files to create a .PAT file (phonetic audio track), thus marking the audio files so that they are rendered phonetically searchable. Search terms and queries are created and applied against these .PAT files. The results of these searches are then compiled and generated into dashboards, charts, trending and analysis reports. Using Nexidia Replay, these audio files can be played back and further searched, so as to enable the enterprise to drill further into the intelligence contained in the actual voice of the customer.

Multiple Language Support

Today, over 50% of the world’s voice communication is recorded with more than 70% of most customer interaction data captured in voice transactions. Additionally, over 600 radio and television feeds are manually monitored in 40 languages from around the world. Manually translating, transcribing, and searching these communications can not only take up to 5 times real time to complete, but can also be expensive finding the appropriate manpower.

The Nexidia technology is designed to easily support multiple languages due to its phonetic core. The technology supports both:

• high level audio - radio and television at broadcast levels of quality
• low level audio - landline and cellular telephony, across a range of levels of quality and noise

Note: The technology can effectively search regardless of the speaker’s gender, age, dialect, accent, or speaking style.

By using a unique patented, phonetic-based approach, the Nexidia technology requires less than 400 sound bytes or phonemes to process all languages and dialects in the entire world. Today, the Nexidia solutions offer support for 33 languages. Additional language plug-ins can be added continuously without the addition of cumbersome resource-demanding dictionaries.