Social media is a powerful force in business, one that cannot be ignored. It is a challenge to provide social media content to current customers and potential customers to consume. However, the greater challenge is the kind of customer relation management that social media demands. A large organization might be the recipient of thousands of posts and tweets per day and on multiple platforms. Actually keeping track of all the various platforms, for each product or service of a complex business, can be a daunting task. That’s why we’ve created the Cognitive Social CRM pattern.
We provide code and documentation to demonstrate how to use IBM Watson Conversation, Natural Language Understanding (NLU), Tone Analyzer, and Cloudant DB to monitor Twitter feeds. The tweets are processed so that the intent can be understood, the sentiment and emotional tone are predicted, and keywords are extracted. This information is presented in various charts, graphs, and is persisted in a Cloudant database for future use.
We use the example of an airline, with Watson Conversation service providing the intents that an airline tweet might contain, such as cabin, delay, service, or seating. Watson NLU extracts keywords from the tweets. These keywords are mapped to give insight into customer concerns. The Tone Analyzer service can report if a tweet is positive or negative, as well as more subtle emotions like anger, fear, joy, sadness, and disgust. Finally, we’ll archive the tweets in a Cloudant database to enable long-term analysis of trends and inflections. A developer can quickly deploy and test this pattern, and we believe you will find it easy to adapt to your individual business domain and needs. To get started, check out our Analyze Twitter handles and hashtags for sentiment and content developer pattern. Enjoy!