What Industries Will Adopt Voice-Related AI Applications First?

Raghav Bharadwaj

Raghav is serves as Analyst at Emerj, covering AI trends across major industry updates, and conducting qualitative and quantitative research. He previously worked for Frost & Sullivan and Infiniti Research.

What Industries Will Adopt Voice-Related AI Applications First?

Episode summary: In this week’s episode we focus on AI application in the customer service business function, – specifically in the context of call centers. We speak with Ali Azarbayejani, CTO of Cogito based in the Boston area, which works on coaching and providing feedback for call center agents in real time.

We aim to focus on what our readers and business executives can do today with AI in the context of call center applications, and how they can go about seeing measurable impacts over a predetermined period of time.

We speak with Ali about what is possible with analyzing voice in real-time today and what kind of ROI can businesses expect for this application. Lastly we touch-base on what factors will make AI inevitable for some companies in the next two to three years.

Subscribe to our AI in Industry Podcast with your favorite podcast service:

Guest: Ali Azarbayejani, CTO Cogito Corporation

Expertise: Machine learning, computational geometry

Brief recognition: From 1984 to 1996, Ali earned four degrees from the Massachusetts Institute of Technology in aerospace engineering, electrical engineering and computer science including a Ph.D. in computational geometry. Ali was previously with Mitsubishi Electric Research Labs for a span of 6 years are the principal technical staff. He went on to found his own consulting firm before joining Cogito as CTO.

Big Idea

In recent years we have seen a lot of vendors operating in the text-based AI space, yet another emerging related application is using AI in voice analysis and feedback. According to Ali Azarbayejani, AI can today listen to conversations in real time and process indicators on three fronts and feed them back to agents at call centers:

  • Short-term indicators like speaking rate, time spent speaking or not speaking for all participants in the call etc. are extracted directly from the audio recording of the call in real time. For example AI can prompt action if an agent is speaking too fast or is leaving long pauses in between speech.
  • Long-term patterns of speech and identification of undesirable conversational patterns with domain specific contextual awareness. For example AI can rate the quality of long term speech patterns for an agent in a call center and prompt recommendations.
  • Conversational dynamics and patterns of interaction from verbal cues. For example identifying cues in the audio to attach features to tone like angry or dominant or check the empathy or rapport.

Ali adds that the largest application today for AI in voice is in the customer service business function, although he is also witnessing a steady increase in adoption for sales applications. For example in a typical customer service call center application, AI can use information obtained by analyzing all indicators and provide real-time feedback in the form of notifications in a small window on the screen. The platform can prompt notifications when it notices that certain conversation patterns are not aligning to the set objective.

Apart from individual level phone conversations, the AI platforms of today can also help supervisors monitor all calls and take control when they deem necessary. In terms of configuration and training Ali said that in a typical new call center, the AI vendor would take around two or three weeks to understand the norms and the different kinds of conversations that can be expected in that specific use-case.

When asked about where he sees the technology becoming ubiquitous in the future, Ali’s response was aligned to the following themes:

  • Large member centric organizations that interact with customers through call centers are going be a big market in the near future for this AI application. In smaller call centers of non-member centric, transactional call centers, this AI application may not be the best fit.
  • Industries where long-term customer relationships matter a lot more, like banking, finance or insurance, or high-value brands in other industries looking for a lot of repeat customers are the ideal first movers who will benefit from this technology.

Interview Highlights with Ali Azarbayejani from Cogito

The main questions Ali answered on this topic are listed below. Listeners can use the embedded podcast player (at the top of this post) to jump ahead to sections they might be interested in:

  • (3:00) What is possible today in phone-conversational AI technology and analyzing voice in real-time today?
  • (6:15) Why is the customer service business function so well suited to this AI-application?
  • (14:50) How are these systems trained? What goes into the configuration of these systems?
  • (20:50) What sector is likely to need this tech the most in the coming 5 years and where will it become ubiquitous first?

Subscribe to our AI in Industry Podcast with your favorite podcast service:

Header image credit: Adobe Stock

Stay Ahead of the AI Curve

Discover the critical AI trends and applications that separate winners from losers in the future of business.

Sign up for the 'AI Advantage' newsletter: