Why conversational AI is an effective listening tool

We are excited to bring Transform 2022 back in-person July 19 and virtually July 20 – 28. Join AI and data leaders for insightful talks and exciting networking opportunities. Register today!

Unstructured data is by its very nature difficult to wrangle.  

But unstructured conversational data? It is one of the hardest sources of data to manage, said Amy Brown, founder and CEO of business-to-business (B2B) software-as-a-service (SaaS) startup Authenticx. 

“AI allows organization of this really messy data source,” Brown said. Still, she said, “it takes a commitment and a desire to use that data source.”  

Conversational AI, often referred to  as natural language processing (NLP) or natural language understanding (NLU), helps computers listen to, understand and extract meaning from human language. These tools are rapidly being adopted across enterprises as organizations look to improve customer experiences in an increasingly digital age.  

According to Allied Market Research, the global conversational AI market was $5.78 billion in 2020 and is expected to grow to $32.62 billion by 2030 – representing a compound annual growth rate of 20%.  

“For someone who’s just exploring for the first time conversational intelligence or conversational AI, I would say that it’s like listening at scale,” said Brown. “It’s listening to a bidirectional conversation. AI allows us to pull out at scale big themes that leaders may want to dive deeply into.” 

A healthy look at conversational AI 

Catered to health insurers, health and hospital systems and pharmaceutical manufacturers, Authenticx’s platform uses conversational AI to aggregate and analyze customer interactions at scale. This includes performing automated topic identification and rule-based classification, targeted evaluations, ML autoscoring and data visualization. Used collectively with real-life call montages, this helps to build context around conversations, identify big thematic issues and target strategic planning, Brown said.  

Most interactions with contact centers involve product or service complaints or compliments, but they also indicate when a customer is “stuck in their journey,” Brown said. Authenticx has trademarked a term for this: The Eddy Effect. This is named after the process of eddies – swirls of water – that form around obstructions in a current. Brown calls it “one of the worst brand experiences possible.” 

The goal is to get them unstuck and use the broad range of data beneath to identify and diagnose problems: Was the issue driven by the claims process? Was it driven by the scheduling process? What are the underlying processes that are broken?  

Brown pointed to one Authenticx client that had implemented a new chat function to streamline customer service and reduce incoming calls. But once it was rolled out, call volume didn’t decrease. By applying conversational data, Brown explained, they were able to determine that users got stuck because they simply couldn’t figure out how to navigate the chat.  

“Listening allows you to diagnose and then to follow whether or not the changes you’re implementing are actually having the intended effect,” Brown said. “So, by ongoing listening, you can keep listening and ask, ‘What’s the next thing that we should be listening for?’”  

Conversational AI also helps companies assess the effectiveness of their contact center representatives and audit their regulatory compliance.  

Brown noted that the U.S. healthcare system is the most expensive in the world yet has the worst outcomes in the world – largely because it has built in this “maze of complexity” into its administrative aspects.  

“If we listen well enough,” she said, “we can understand how to streamline, how to create a better experience and how to start to reduce the administrative costs of our system and hopefully start to drive better outcomes.”  

Don’t think of conversational AI as a cost center

Another inherent issue across enterprises is the perception of customer service contact centers as cost centers. They should instead be considered a sales and marketing part of the business, Brown said. The conversations being aggregated are providing insights that could directly inform marketers about how they can message more effectively or improve their website and other platforms. It also gives them affirmations about what they’re doing right via positive feedback.  

Marketers pay millions of dollars every year for customer surveys and interviews, when “conversational data has testimonials literally happening every day,” Brown said. “It’s a marketer’s dream.” 

That’s not to mention that surveys are inherently skewed, she said. They only tap a fractional amount of a customer base and questions are biased because they’re asked within a company’s own paradigm. “It’s the concept of unsolicited feedback that is available at our fingertips, if we only listen to the data that we already own.”  

A trained social worker, Brown started Authenticx in 2018 after spending 20 years in a range of public and private sector healthcare roles, including public relations, policy development, quality improvement and insurance operations.   

“I knew there was tons of value within these customer conversations, it was just really hard to extract and surface,” she said.  

Listening is knowledge

The company has assembled a diverse team of social workers, nurses and customer experience professionals with experience in healthcare. “So, when they’re listening, they’re listening with knowledge,” Brown said.  

An organization with just a “homogenous type of human being” is training its AI to have bias. “We’re working very hard to have models that are as inclusive and unbiased as humanly possible,” Brown said. “AI is only as good as the training data and the training data is trained by human beings.” 

The fast-growing startup expects to raise a series B round by the end of the year and plans to expand its reach eventually to other healthcare areas including retail pharmacy and telehealth, care management and long-term care. But overall, the intent is to stay in healthcare because there is still so much to do in the space, Brown said.  

“The beauty of conversational intelligence is that you can start to listen from the perspective of your customer,” she said. “And so, you’re just going to get a much more authentic version of the truth.” 

Originally appeared on: TheSpuzz