CEO and Founder of Vivid Australasia and Director of CogntiveCX Vivid Australasia is the exclusive sales and marketing agent for Dapresy a global provider of data visualization and data integration software. Before his fulltime work at Vivid and CognitiveCX, Heath spent 15 years working as one of the Directors at Sprout Research. As the CEO Heath is responsible for running all facets of the business. Heath has a proven executive management track record and over 20 years experience driving sales and growth across multiple industries. He is practical and pragmatic entrepreneur with a strong focus on the ‘So What’.
What is the goal of our presentation? To educate market researchers and insight professionals on Artificial Intelligence (AI). Practical examples and case studies will be woven into this education piece. The audience will work away with a clear and concise understanding of what Artificial Intelligence is, how it works and its application for market researchers and more importantly how market researchers can use Artificial Intelligence to generate and deliver actionable customer experience insights. CogntiveCX is a Brisbane based; we have been using Watson Knowledge Studio, Watson Natural Language Understanding and Watson Discovery Services to uncover insights in our customers’ unstructured data for over 24 months. In this session, we show how we use IBM Watson to analyse existing omnichannel datasets (from social media to call centre logs) and quickly extract contextualised insights as they apply to different stakeholders in real time enabling improved decision making. CogntiveCX is a global reference for IBM Watson. At the request of IBM, we will be attending the IBM ThinkTank in Las Vegas to present how AI can be used to transform Customer Experience. The conference runs over three days with 1500 sessions and 50,000 people attending. New learnings on AI and its application for market researchers will be presented as part of the session. In 2017 we, in conjunction with one of our clients won the cognitive BigData and Innovation award. The award was granted on the application of cognitive to the customer experience industry. This practical case study will be used throughout the to help the transition between learning and applying. Artificial Intelligence enables analysis of the organic and unfiltered voice of the customer. Every second a conversation about a company’s products, brand, service or processes is happening. These conversations are often defined as a moment. In the United States alone it is said that 30 billion moments happen every day. Big Data discussions are quickly being replaced with discussions around “Dark” or unstructured data. In 2017 more data was created than the previous 5000 years. It is estimated that 2.5 quintillion bytes of new data are generated every day. Dark data is often defined as an organisations information asset. It is information of extreme size, complexity and is unstructured. Dark data is data that organisations are currently collecting. It is collected, processed and stored but do not leverage. Dark data is varied and exists within the individual silos of an organisation. The small amount of dark data that is currently leveraged is done in isolation. A single view of this data provides organisations with a single view of their customer. With AI we can connect departments and provide Customer Experience Managers with one single view of the customer. So why is this important for market researchers and why as market researchers do we need to learn about AI? Customers are already telling us about their experiences. They email, the call, they write to us, they complete feedback surveys, and they speak out on social media and in the comments on news sites or forums. Stories are being told without a single question being asked. The answer to improving customer experience is in the data that is already available. The volume, speed and complexity of this data have made it difficult to access, analyse and draw insight from. Artificial Intelligence is currently used by organisations to process large amounts of data and or develop new ways to interact with the customer, ie chatbots. However, Artificial Intelligence through its use of NLIP and Deep Learning can extract meaning from free text, helping companies to Listen, Understand and React to moments. Defining then redefining the customer experience is possible if we understand the intent, sentiment and emotion associated with these moments. Artificial Intelligence can help researchers define, redefine and improve the customer experience. But first, we, as researchers need to understand Artificial Intelligence and how we can use it. In just a couple of years, AI has jumped from Sci movie plots to the mainstream application. Artificial intelligence (AI) makes it possible for machines to learn from experience, adjust to new inputs and perform human-like tasks. Most AI examples that you hear about today – from chess-playing computers to self-driving cars – rely heavily on deep learning and natural language processing. Using these technologies, with today’s computing power AI can ingest, process, understand and allow analysis on large amounts of data. Sample size has been a challenge for researchers and insight managers and is often one of the c-suites reasons for not adopting new insights. With AI sample sizes of 1200 are replaced with a sample size of millions. In one study conducted in Australia, we used AI to process and understand over 20 million data points from every consumer touch point. Traditional methods of mining text data are done through manually coding each data point/comment into a broad theme/topic. This traditional method is labour intensive, time prohibitive cost prohibitive has greater margins for error, and is rigid (particularly with big data, or when more than one data source is being used). This is also the reason this data is often put to one side and not used. Artificial intelligence uses cognitive natural language processing (NLP). Watson has a collection of natural language processing API’s (called Alchemy Language), that help you understand sentiment, keywords, entities, high-level concepts and more. And can process billions of data in a short amount of time. The sentiment output from Watson is far more granular than any other sentiment analyser tool, which can only tell you whether a piece of data is ‘positive’ or ‘negative’. It is also faster and more accurate than traditional coding methods. Also, Watson’s Tone Analyzer uses linguistic analysis to detect three types of tones from the text: emotion, social tendencies, and language style. Emotions identified include things like anger, fear, joy, sadness, and disgust. Emotions are a key driver of good or poor experiences. Understanding the emotion associated with the experience will enable organisations to bridge the experience gap between what they are currently delivering and what the customers expect. Each data source is ingested into Watson and provided with a context of the data, e.g. survey results, containing verbatim commentary on reasons for good and poor service with X, as a minimum. As each data source is inputted to Watson, a knowledge domain is built on, that contains information specific to an organisations business. Watson is, therefore ‘trained’ to understand each organisation’s data, e.g. names of entities, such as services, products and locations, as well as relations between entities and characteristics of those entities from text. AI uses Natural Language Processing (NLP). NLP applies knowledge about the structure of language to extract the names of entities, such as companies, services and locations, as well as relations between entities and characteristics of those entities from text. Natural Language Processing relies on two core components: (1) A large knowledge domain that contains and understands the different entities that may be involved in any piece of text. This might be business units, dates, people, products etc. (2) A strong sentiment/emotion analyser. Having the data structured is not enough if you can’t easily extract the sentiment of the author. Traditionally these processes have been time-consuming and relying heavily on manual input. However, IBM Watson has revolutionised this process by providing a solution that is both highly accurate and incredibly efficient. The two modules used in this process are Watson Knowledge Studio and Watson Discovery. Using Watson Knowledge Studio, you can easily train a custom machine learning model to detect entities in your text. Entities that are specific not only to your industry but to your organisation. Whether these are business units or staff members or services performed, IBM Watson Knowledge Studio can categorise and then predict when these entities are involved in customer response forms. Once you have a comprehensive knowledge domain, you can begin extracting insights with Watson Discovery. Traditional sentiment models will read the sentiment of the entire text. This can lead to false positives as shown above. The power of Watson Discovery is in its ability to distinguish different themes within the text and giving individual sentiment readings for each one. It will also analyse the emotion displayed within the text while retaining the context of the individual textual themes.