Successfully reported this slideshow.
We use your LinkedIn profile and activity data to personalize ads and to show you more relevant ads. You can change your ad preferences anytime.
Upcoming SlideShare
What to Upload to SlideShare
What to Upload to SlideShare
Loading in …3
×
1 of 10

Extracting value from text and audio to inform business strategy

0

Share

Download to read offline

Tailwind Traders recent internal employee survey showed their employees are frustrated with lengthy processes for simple actions, such as booking vacation and other company benefits. They want to reduce the friction of reviewing and booking vacation so it’s a simple, easy and pleasant process for their employees. In this session you will see how Tailwind Traders applied Conversational AI best practices to simplify the vacation process for their employees. Using the Bot Framework Composer tooling you can quickly build conversation flows, incorporate intelligence services such as Q&A maker and LUIS, test and deploy your virtual assistant to the cloud and embed it where your customers and employees spend their time.

Related Books

Free with a 30 day trial from Scribd

See all

Related Audiobooks

Free with a 30 day trial from Scribd

See all

Extracting value from text and audio to inform business strategy

  1. 1. MICROSOF T CONFIDEN TIA L – INTERNA L ONLY Tailwind Traders
  2. 2. MICROSOF T CONFIDEN TIA L – INTERNA L ONLY • Python • Notebooks • VSCode • Cognitive Services • Speech-to-text • Text Analytics • LUIS • Blob storage • PowerBI
  3. 3. MICROSOF T CONFIDEN TIA L – INTERNA L ONLY Store Prepare and train Model and serve Azure Blob Storage CSV files (structured) PowerBI Azure Cognitive Services Azure Blob Storage Audio files (unstructured)
  4. 4. MICROSOF T CONFIDEN TIA L – INTERNA L ONLY Azure Blob Storage CSV Files (unstructured) Analytics Engine Azure Blob Storage Services Speech-to-text Text Analytics Content Output Table Custom Notebook Audio files (unstructured) LUIS PowerBI Azure Logic Apps
  5. 5. MICROSOF T CONFIDEN TIA L – INTERNA L ONLY Support Ticket ID Date Created Date Completed Cust ID Audio file path Transcribed Text Theme Key Phrases Sentiment 0 8/20/19 8/22/19 132459 None “The delivery took so long that I no longer need this device. How can I return it?" Speed ['delivery', 'device'] neutral 1 8/23/19 8/24/19 183619 http://<blob storage> “I have seen similar products for slightly lower prices. How can I return this product?" Price [similar products', 'lower prices'] neutral
  6. 6. MICROSOF T CONFIDEN TIA L – INTERNA L ONLY • SupportTicketID • CustomerID • DateCreated • DateCompleted • Escalated • Theme • Sentiment Category • Key Phrases •
  7. 7. MICROSOF T CONFIDEN TIA L – INTERNA L ONLY Responsible AI considerations • Account for accents and other languages into input • Storing private recordings in Azure • ChatBot
  8. 8. MICROSOF T CONFIDEN TIA L – INTERNA L ONLY Github: https://github.com/microsoft-us-ocp-ai/developers-guide-to-ai Cognitive Services: https://azure.microsoft.com/en-us/services/cognitive-services/ PowerBI: https://docs.microsoft.com/en-us/power-bi/create-reports/sample-tutorial-connect-to-the- samples Blob: https://docs.microsoft.com/en-us/azure/storage/blobs/storage-quickstart-blobs-python
  9. 9. © 2018 Microsoft Corporation. All rights reserved. Microsoft, Windows, and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other countries. The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft must respond to changing market conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information provided after the date of this presentation. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.

Editor's Notes

  • PowerPoint 
    LUIS
    Logic Apps
    Power BI 
  • 1 –2 min: Show the website for Tailwind Traders and its features 
    (Add Quantitative Data – TT has 1000 tickets per day? It takes one person forever to look at the tickets one by one. This is why we need AI to solve these tickets more efficiently. Understand the intent of the customers and act upon it ) , cost savings? Productivity and branding will increase. Resources will be reduced to scan tickets,...

    Problem Statement : Increased amount of support tickets regarding usage of these offerings. 
    Goal: Store, Analyze, and Extract insights from their text and audio data to make better product backlog decisions and reduce their support tickets
    Learning Obj:
     - Solving the goal of the company with metadata of each ticket using Azure Cognitive Services Text Analytics and Speech to Text 
    - Aggregate these findings to inform their product backlog and implement improvements 

  • LUIS - ML algorithm, supervised learning : mention train and test
  • Rows = issue x entity
    Columns = characteristics
  • If time permits
  • ×