Balto Improves Real-Time Guidance With Intent-Based Voice Processing
Balto has released its next generation of intent-based conversational dialogue models, raising the bar in conversational guidance capability, and reaffirming the company’s reputation as the premier real-time guidance platform for contact centres. Using state-of-the-art deep learning techniques, Balto’s conversational dialogue modelling has been the gold standard for AI-powered real-time guidance during calls. With the latest […]
Topics
Balto has released its next generation of intent-based conversational dialogue models, raising the bar in conversational guidance capability, and reaffirming the company’s reputation as the premier real-time guidance platform for contact centres.
Using state-of-the-art deep learning techniques, Balto’s conversational dialogue modelling has been the gold standard for AI-powered real-time guidance during calls. With the latest generation of intent-based language processing, Balto’s proprietary AI library and Smart Checklist are now even smarter, recognising the true intention of the speaker — be it an agent or customer — and allowing for more natural conversation to take place while maintaining script integrity.
Mike Goldstein, VP of Engineering at Balto, elaborates on the upgrade, “Before we were decoding in a very syntactically based way, looking for pattern-matches. That works really well in a lot of use cases. If you have a very strict standard, that’s ok. But that can be restrictive to agents. Our next generation of intent-based natural language processing allows agents to be more natural in their conversations and still hit script marks.”
The intent-based language processing enhancements augment the company’s existing real-time guidance platform by relying less on specific keywords. Going beyond the AI’s ability to hear “trigger” words that would then prompt a subtle nudge to agents during calls, the new model identifies and decodes synonyms, rephrasing, paraphrasing, substitutions, and common idiomatic expressions. A good example of this intent-based decoding is when a customer says “that will hurt my pocketbook.” The AI is able to recognise that phrase as a sales objection in real-time, and then prompt the agent with ways to handle the objection.
This next generation of intent-based language processing has already been rolled out to Balto customers, and is the first wave of several for 2022. Customers should notice immediate flexibility in their agent conversations. For more rigid call scripts that necessitate strict compliance marks by agents, the processing enhancements can be disabled.
