Artificial Intelligence is at the heart of contact center digital transformation as it has far-reaching consequences. In fact, there are many areas of contact center that are getting impacted by AI:

• Multi-accent multi-language Voice Recognition – English is spoken in different accents in the US, UK, Australia, Canada, India and in several European countries, such as, France, Germany, Spain, Italy to name a few. It is virtually impossible to train the speech recognition model with all possible types of accent, and hence a pure machine learning model may not always work. Artificial intelligence becomes relevant in the context of the model that “learns” continuously and gets better with its understanding of different types of accent. The situation is far more complex when multiple languages are taken into consideration.

• Natural Language Processing (NLP), Natural Language Understanding (NLU) and Natural Language Generation (NLG) – While the term NLP is used broadly to refer to the technology needed for a machine to interact with a human, there are many sub-problems under the broad umbrella that need to be handled differently, and it goes without saying that AI plays a critical role in each of these areas. First, there are several (unstructured) ways of saying the same thing in English. Dissecting such unstructured sentences to extract entities and converting them into a structured sentence is the realm of NLP and is powered by AI. Second, the same meaning can be conveyed in English by multiple sentences that are syntactically different. Interpreting a sentence and extracting “intent” from the sentence is the realm of NLU and is also powered by AI. Finally, once the key words are retrieved, creating a human-understandable sentence is the realm of NLG and this again requires the help of AI. Finally, when the sentence is read out to the customer by the virtual agent, Text-to-Speech technology is used and to make it sound like a human rather than a machine, once again AI techniques are a must. The challenges become even more complex when we need to analyze sentences in different languages and extract intent from them. It is practically impossible to first encode and then use the grammar in every possible language to do syntactic analysis of sentences spoken in a language of customer’s choice, let alone the semantic analysis. AI offers some clues as to how this complex multi-language problem can be solved in a practical way.

• Sentiment Analysis – when an interaction happens between a customer and the call center agent, it is important to understand if the customer is happy or angry, satisfied or frustrated and that can be inferred from the choice of words and the way they are delivered. AI can be used to make that assessment and deliver value both tactically as well as strategically. On the tactical front, AI can advise the agent to use appropriate words and sentences during the conversation to suit the sentiment of the customer. Strategically, AI can be used to correlate various factors to the displayed sentiment of the customers, identify patterns and create a playbook for how to respond to customers under what circumstances. AI algorithms can then advise the human agents to follow through the guidelines in the playbook, thereby improving the customer satisfaction score in a continuous manner.

• Knowledge graph generation based on Enterprise data – there are many sources of data for an enterprise based on which a call center agent is expected to respond to a customer’s query. These data sources for a client could be troubleshooting guides, FAQs, technical specification, engineering manuals, historical call records etc. Unless the data from these sources can be ingested and organized based on business taxonomy and industrial ontology to create a knowledge graph, it becomes almost impossible to extract the most relevant and contextual answer to the customer’s query in real time. AI is the technology that can be used for automatically creating this knowledge graph.

• Agent Assist – while creating the knowledge graph is important, equally important, or more important it is to do semantic search to retrieve the most relevant answer to the customer’s query. Semantic search is a data searching technique in which a search query aims to not only find keywords, but to determine the intent and contextual meaning of the words used by a customer. AI can be leveraged to not only extract the customer’s “intent”, but also to use the “intent” combined with a mix of factors, such as location or the user’s past history, to retrieve the most relevant answer. Once one can provide exactly what the customers are looking for, customers feel that they are listened to, and that increases customer satisfaction and hence the opportunity for better conversion rate.When virtual agents are used, and a virtual agent cannot resolve the customer’s query, she seamlessly transfers the call to a live human agent. When the call is transferred, not only is the context of the call transferred but also recommendations based on customer’s purchase history, browsing history and other related information are passed to the agent for up-sell and cross-sell purposes. These recommendations are powered by AI/ML algorithms.

• Automated Workflow – there are several steps associated with the execution of a business process in the contact center. For example, a contact center agent, typically, needs to look at multiple screens to find the relevant data and aggregate them doing copy-and-paste from these screens before sending a response back to the customer. The agent needs to complete several after-call works, such as, sending reminders, order confirmations to the customer via email and/or SMS and do updates on multiple systems. Many, if not all, of these steps can be automated using AI-based Robotic Process Automation (RPA).

• Digital Self-Service – customers get frustrated when call center agents spend a lot of time on the phone but cannot help them solve their problem. Sometimes, this happens because they cannot look up relevant content which could be a combination of text, diagrams, tables, technical drawings, and videos, and consolidate the information in a comprehensible manner. However, using AI-based algorithms, all relevant information spread out in different systems in different formats, can be stitched into a cohesive story, from which the virtual agent can extract step-by-step guidance and enable self-service for the customer, helping solve her problem in a seamless manner.

• Personalized Service – in order to provide personalized service to a customer, it is mandatory to understand the customer’s need precisely, have a good understanding of her preferences, be aware of the previous conversations and disputes, if any, and blend them all in an intelligent manner. AI is the core technology that enables this.

Thus, Artificial Intelligence (or AI) is dramatically transforming each and every function associated with the contact center, starting with the front-end interaction with the customer to the back-end automation, from understanding the customer’s voice to extracting the intent to performing a semantic search in a multi-dimensional knowledge graph to retrieving the most relevant information and saying it back to customer in a human-like voice. Even creation of the knowledge graph from heterogeneous customer data repositories, and automation of routine tasks for the agent to providing assistance to the human agent with all relevant information about the customer and guiding him/her to respond to the customer in a way that is consistent with the customer’s sentiment are all driven by AI. With all of these enhanced features packed into contact center operations, AI powers what is called a Cognitive Contact Center, bringing in an unprecedented amount of efficiency while also enhancing customer satisfaction.