IBM Watson’s Tone Analyzer Service Aims to Disrupt Customer Engagement

Share this post:

A new Beta feature launched today within the IBM Watson Tone Analyzer Service. Now you can analyze your customer support conversations with the new Tone Analyzer for Customer Engagement Endpoint. Escalate customer conversations when they turn sour, or find opportunities to improve customer service scripts, dialog strategies, and customer journeys.


When you analyze your data with the new Tone Analyzer for Customer Engagement Endpoint, you’re analyzing it with an alternate Tone Analyzer model specifically trained on customer support conversations on Twitter. This model detects a new set of tones designed specifically for customer engagement use cases. The tones included are frustrated, sad, satisfied, excited, polite, impolite and sympathetic. Learn more about how the model was developed here.

With this addition, we now have two endpoints available:

1. Tone Analyzer General Purpose Endpoint

Use the Tone Analyzer General Purpose Endpoint to monitor social media and other web data. Analyze short-form text like tweets or reviews, or longer documents like articles and blog posts.

–       Tones detected within the General Purpose Endpoint include joy, fear, sadness, anger, disgust, analytical, confident, tentative, openness, conscientiousness, extraversion, agreeableness, and emotional range.

2. Tone Analyzer Customer Engagement Endpoint (BETA)

Use the Tone Analyzer Customer Engagement Endpoint to monitor customer service & support conversations.

–       Tones detected with this endpoint include frustrated, sad, satisfied, excited, polite, impolite and sympathetic


Check it out today, and let us know what you think!

More Watson Stories

Watson Speech to Text and Text to Speech have released Lite Plans!

On Nov 1st, 2017, both Watson Speech to Text and Watson Text to Speech released Lite Plans, as part of IBM Cloud Platform’s initiative to roll out Lite Plans across public cloud services.

Continue reading

Enhance a chatbot conversation with Context Variables and System Entities

Imagine, you are in a conversation with a chatbot and you feel that the human angle is completely missing because the bot starts it's dialog with a usual "Hi" or "Hello". You may want to personify the conversation by adding the name of the person (who's logged in) to the boring "Hi" or "Hello". Ever thought of this? It's not just personification, How about wishing appropriately based on the time of the day someone invokes your chat application? Also, how about passing values back and forth during a conversation between the nodes or from application to a node?

Continue reading

Translate Watson Conversation Chatbots

Creating and maintaining chatbots in multiple languages can be costly, error-prone, and not easily scalable. Each change in a bot needs to be manually replicated across each language the bot supports. The manual nature of maintaining multilingual chatbots can have a real impact on continuous delivery as the language-specific changes must be performed by language experts and this takes time. Moreover, we need to pay for every single change – intents, entities, and bot output. This can get expensive very quickly — so, what's the best way to translate chatbots into different languages without impeding continuous delivery and disrupting DevOps?

Continue reading