A retro robot with a vintage black phone - conversational AI in the call center concept
PHOTO: Adobe

Natural language processing (NLP) and conversational AI are often used together with machine learning, natural language understanding (NLU) to create sophisticated applications that enable machines to communicate with human beings. This article will look at how NLP and conversational AI are being used to improve and enhance the Call Center.

NLP is a technological process that facilitates the ability to convert text or speech into encoded, structured information. By using NLP and NLU, machines are able to understand human speech and can respond appropriately, which, in turn, enables humans to interact with them using conversational, natural speech patterns.

Predictive Forecasting Isn’t Just for Weather

Predictive algorithmic forecasting is a method of AI-based estimation in which statistical algorithms are provided with historical data in order to predict what is likely to happen in the future. The more data that goes into the algorithmic model, the more the model is able to learn about the scenario, and over time, the predictions course correct automatically and become more and more accurate.

Workforce management software that utilizes AI is able to analyze huge amounts of historical volume data and recommend the best forecasting algorithm that will result in more accurate forecasts. For instance, Amazon’s AutoML functionality creates a predictor which trains the optimal model for a brand’s datasets. Additionally, Amazon Forecast includes six built-in algorithms which can be selected, but ultimately, using Forecast to train the model for the most optimum algorithm produces the best results.

A 2019 paper by ResearchGate on predicting call center performance with machine learning indicated that one of the most commonly used and powerful machine learning algorithms for predictive forecasting is Gradient Boosted Decision Trees (GBDT). Gradient boosting works through the creation of weak prediction models sequentially in which each model attempts to predict the errors left over from the previous model. GBDT, more specifically, is an iterative algorithm that works by training a new regression tree for every iteration, which minimizes the residual that has been made by the previous iteration. The predictions that come from each new iteration are then the sum of the predictions made by the previous one, along with the prediction of the residual that was made by the newly trained regression tree (from the new iteration). Although it sounds (and is) complicated, it is this methodology that has been used to win the majority of the recent predictive analytics competitions.

Related Article: 4 of the Top Call Center Challenges for the Coming Year

Customers and Agents Work Better Together

Puneet Mehta, founder and CEO of Netomi, an omnichannel AI-based customer service solution provider, shared that conversational AI agents are being used within call centers to reduce costs, enhance agent productivity and improve the customer experience. “By sitting alongside human agents within platforms like Zendesk, Salesforce, Gladly, Freshworks, etc., virtual agents act as the first line of defense when a ticket comes in,” explained Mehta. “Netomi’s AI agents take the best course of action based on the specific ticket. Highly repeatable tickets like order status or refunds can be automatically resolved without human intervention.” 

NLP and AI are able to provide detailed information to agents who handle more complex queries, and even to suggest the most appropriate agent for each scenario. “For more complex questions, the AI gathers information from the customer and back-end systems and drafts a response for an agent to quickly review and send,” said Mehta. “For the most complex queries, AI agents summarize and route tickets to the right agent based on experience, bandwidth, sentiment, or other specific business rules.”

By using natural language understanding (NLU), conversational AI bots are able to gain a better understanding of each customer’s interactions and goals, which means that customers are taken care of more quickly and efficiently. “NLU-powered AI agents are making a significant impact on support teams. Netomi’s NLU automatically resolved 87% of chat tickets for WestJet, deflecting tens of thousands of calls during the period of increased volume at the onset of COVID-19 travel restrictions,” said Mehta.

Related Article: How Customer Data Platforms Can Benefit the Call Center

NLP & NLU Enable Customers to Solve Problems in Their Own Words

The use of AI-based Interactive voice response (IVR) systems, NLP, and NLU enable customers to solve problems using their own words. Today’s IVR systems are vastly different from the clunky, “if you want to know our hours of operation, press 1” systems of yesterday. Jared Stern, founder and CEO of Uplift Legal Funding, shared his thoughts on the IVR systems that are being used in the call center today. 

"NLP has revolutionized IVR systems and has made routing very effective. Conversational IVR enhances customer experience as it is easier than the traditional methods. NLP can also be used for data analysis,” said Stern. “Based on customer interaction, content that will push them to an advanced stage in the sales funnel can be identified. Call centers can use NLP for speech-to-text applications. Generic data like name and address can be collected quickly. Agent data processing can be reduced, and security can be increased."

NLU converts unstructured text and speech into structured data which allows the AI to more precisely understand intent and context. NLU is able to achieve this through the combination of three different technologies: Syntactic analysis applies rules that are specific to sentence structure, i.e. syntax, to determine part of the meaning of what's being said; semantic analysis looks at the relationship between words in order to understand meaning; pragmatic analysis determines the context of sentences to more fully understand intent.

“Natural language understanding enables customers to speak naturally, as they would with a human, and semantics look at the context of what a person is saying. For instance, ‘Buy me an apple’ means something different from a mobile phone store, a grocery store and a trading platform. Combining NLU with semantics looks at the content of a conversation within the right context to think and act as a human agent would,” suggested Mehta.

Related Article: Call Centers vs. Contact Centers: Understanding the Key Differences

Conversational Intelligence Facilitates Smarter AI

Raj Gupta, chief engineering officer at Cogito, an AI coaching system provider, thinks that with customer and employee expectations so high, and call center complexity increasing exponentially, emerging technologies such as conversational intelligence and NLP have become vitally important. “Conversational intelligence combines forms of artificial intelligence (AI), including machine learning (ML) and NLP technology,” said Gupta. “It is used to create and train algorithms to deduce intent and emotional sentiment from customer speech or text. This analysis can then provide customer support to human agents to improve interactions and customer experiences to quickly and efficiently resolve customer needs and issues, improve satisfaction, and even simplify coaching and onboarding agents.”

Conversational intelligence is typically focused on human-to-human and human-to-machine speech, which makes it perfect for customer support channels, call centers, and chatbots. “Yet, the actual value of conversational intelligence and NLP comes when it reveals the sentiment and intent behind customer interactions to help augment a human agent versus a chatbot, as consumers overwhelmingly prefer to interact with people today,” suggested Gupta. “Chatbots in call centers are limited to using these tools for highly repetitive tasks in well-defined, closed interactions. Augmented intelligence has far more possibilities by focusing on human-aware technologies for machine collaboration with human control.”

Final Thoughts

AI technologies such as natural language programming, along with natural language understanding, machine learning, and natural language generation, allows machines and their associated applications to have conversations with humans in a manner that is natural — either through text or speech. By using algorithmic forecasting and conversational intelligence, AI technologies enable customers and agents to work together more effectively and efficiently, improving and enhancing the call center experience for both customers and employees.