The secret value of recorded audio data for contact centres

Language technologies
Tilde speech technologies

Most companies providing customer or internal user support by phone make recordings of these conversations for quality assurance, but there’s so much more you can do with this data by introducing automated speech recognition (ASR) technology into your workflows. Let’s talk about these benefits and opportunities, which have long only been available for the world’s dominant languages. At Tilde, we are working hard to also make such technology-driven business gains available for companies operating using Baltic languages.

Gaining new value from existing voice data

Contact centres have a huge opportunity to derive new meaning and insights from voice data. Almost all (95%) C-level executives regard voice data as “valuable” or “very valuable” to their organisation and over three quarters (76%) believe that a ‘Voice First’ strategy will be in place within less than five years*, according to a survey conducted by Sapio Research.[1]

Understanding the customer voice and sentiment helps to deliver key metrics on agent performance. By tracking success, contact centre managers can motivate and empower agents to deliver better customer experiences and create a data-driven workforce giving staff more time for analysis and consultation instead of monotonous technical operations. The combination of content analysis tools and speech-to-text cost-efficient technology will help to quickly assess the ever-increasing flow of non-written information, making it possible to see the big picture and make data-based decisions.

Searching audio and video archives have never been easier

There is no need to physically watch or listen to dozens of hours of records every day as ASR technology makes it possible to easily find a specific place in the audio or event video recording, and can even show the timeline with time titles, indicating which sections contain the required keywords. Traditional search engines generally find only what is included in the file title or the description, whereas implementing speech technology considerably simplifies the use of archived call recordings.

Dealing with Latvian, Lithuanian or Estonian – there are no longer any limitations!

Thanks to the efforts of the language technology research community, including our experts at Tilde, AI empowered speech technologies are now also available for all three Baltic languages. We have succeeded in overcoming the biggest challenge for our languages – morphological complexity and an insufficient amount of training data. New generation neural network technology can quickly convert sound to text and apply grammatical rules, i.e. using punctuation marks. And best of all - by combining Tilde’s deep neural network ASR technology with the data your organisation already possesses (both audio recordings and special business keywords), it is possible to reach up to 90% accuracy in speech recognition! Empowered by AI based tools, we strongly believe that there is no challenge too big or language too complicated for us!


Martins Suna
AI Business Developer Connect via Linkedin

Interested in utilizing speech technology into your workflow?
We’re here to help!

Request demo 

[1]Survey conducted by Sapio Research for Red Box asking 588 IT Directors or C-level executives responsible for IT across UK, US and Singapore.