Skip to content

Latest commit

 

History

History
26 lines (15 loc) · 3.09 KB

File metadata and controls

26 lines (15 loc) · 3.09 KB

Final Case

Objectives

This culminating activity will allow you and your group to get hands-on experience designing and architecting an intelligent agent from end-to-end. The output of this activity is for attendees to work together and build on what they've learned to design and architect an intelligent agent.

Case Study

A hotel chain CEO recently announced in a trade magazine that she is looking to enhance the customer experience by augmenting their in-person concierge service by embedding bot capabilities into their member services mobile app. The hotel guest reservations are held within an Azure SQL Database which contains the dates and the hotel location of the guest that is presented to the application. The bot is intended to handle the common request that guest will make such as setting up wake up calls, ordering services such as a pool cabana or spa session, or handling cab reservations with their in-house cab company. Therefore, the bot should be able to offer to book these services, providing a choice on dates and durations. The intention is to free up the real concierge’s time to focus on informing guest about the local area using local knowledge. It is critical that access to the bot is secured. The hotel chain is a Microsoft Shop. Their network is setup using AD, the databases using SQL Server and cloud applications developed in .Net in Azure. The CFO and CIO has made budget available for this project to be delivered within 6 months.

Exercise

Using what you've learned throughout the airlift, develop a potential bot design and architecture. Think about the bot design and what problems you would like the bot to solve. Next, come up with a LUIS schema design that will help you address the problems. In addition, come up with an architecture and what other services/enhancements you could integrate.

If you choose to use Cognitive Search in your solution, develop a potential Cognitive Search enrichment pipeline plan that contains how you will gather the documents, what cognitive skills you may use (predefined and custom), the skills sequence, how you'll integrate it into the larger solution, and any other details you think are important to call out.

Time-permitting, your group is encouraged to create a simple POC or some mock conversations (this resource may help) to present.

Your group should be ready to present. Feel free to use any of the materials you've used thus far including but not limited to: