{"id":1443,"date":"2022-01-05T20:23:30","date_gmt":"2022-01-05T20:23:30","guid":{"rendered":"https:\/\/salarydistribution.com\/machine-learning\/2022\/01\/05\/how-to-approach-conversation-design-with-amazon-lex-building-and-testing-part-3\/"},"modified":"2022-01-05T20:23:30","modified_gmt":"2022-01-05T20:23:30","slug":"how-to-approach-conversation-design-with-amazon-lex-building-and-testing-part-3","status":"publish","type":"post","link":"https:\/\/salarydistribution.com\/machine-learning\/2022\/01\/05\/how-to-approach-conversation-design-with-amazon-lex-building-and-testing-part-3\/","title":{"rendered":"How to approach conversation design with Amazon Lex: Building and testing (Part 3)"},"content":{"rendered":"
\n

\"\"<\/a><\/p>\n

In parts one<\/a> and two<\/a> of our guide to conversation design with Amazon Lex<\/a>, we discussed how to gather requirements for your conversational AI application and draft conversational flows. In this post, we help you bring all the pieces together. You\u2019ll learn how draft an interaction model to deliver natural conversational experiences, and how to test and tune your application.<\/p>\n

Draft an interaction model<\/h2>\n

In the second post<\/a> of this series, you identified some use cases that you wanted to automate and wrote sample interactions between a user and your application. In this post, we use these use cases to build an Amazon Lex framework, called an interaction model, but first, let\u2019s review some important definitions.<\/p>\n

An interaction model<\/em> defines the tasks that your application can help a user accomplish in a structure that a bot can understand. In your interaction model, you define the five major components for Amazon Lex: intents, sample utterances, slot names, slot values, and slot synonyms. You can make an interaction model (also called a bot definition<\/em>) in whatever format works best for you and your team, such as in a spreadsheet or directly on the Amazon Lex console.<\/p>\n

We start by defining intents. An intent<\/em> is a user\u2019s goal\u2014what is the user trying to accomplish when they interact with your bot? Continuing with the banking bot example, a user goal might be something like blocking a card or opening a bank account. For example, in the following dialog, the user\u2019s intent is to set up a recurring payment.<\/p>\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n
Turn<\/span><\/td>\nParticipant<\/span><\/td>\nDialog<\/span><\/td>\nNotes<\/span><\/td>\n<\/tr>\n
1<\/td>\nAmazon Lex<\/td>\nThanks for contacting Your Favorite Loan Company. What can I help you with today?<\/td>\n<\/tr>\n
1<\/td>\nCustomer<\/td>\nI wanna set up a recurring payment.<\/td>\n<\/tr>\n
2<\/td>\nAmazon Lex<\/td>\nYour current balance is [$200.00]. That includes an overdue balance of [$100.00]. A payment of [$200.00] is due on October 15th. To set up a recurring payment, you\u2019ll need the routing number and the account number. Do you have that information ready?<\/td>\nCan customers change their debit date? Or only the account?<\/td>\n<\/tr>\n
2<\/td>\nCustomer<\/td>\nNo<\/td>\n<\/tr>\n
3<\/td>\nAmazon Lex<\/td>\nOkay. I can wait. Let me know when you have the information.<\/td>\n<\/tr>\n
3<\/td>\nCustomer<\/td>\n\n

\u2026<\/p>\n

I\u2019m ready.<\/p>\n<\/td>\n<\/tr>\n

4<\/td>\nAmazon Lex<\/td>\nTo start, what\u2019s the 9-digit routing number?<\/td>\n<\/tr>\n
4<\/td>\nCustomer<\/td>\n999-999-999<\/td>\n<\/tr>\n
5<\/td>\nAmazon Lex<\/td>\nAnd what\u2019s the account number?<\/td>\n<\/tr>\n
5<\/td>\nCustomer<\/td>\n1234 567 890<\/td>\n<\/tr>\n
\u2026<\/td>\n\u2026<\/td>\n\u2026<\/td>\n<\/tr>\n<\/tbody>\n<\/table>\n

Each intent should come directly from a use case that you previously established in the earlier steps. We go into more detail about how to define intents later in this post.<\/p>\n

Continuing on, a prompt label<\/em> is a value defined by the application\u2019s designer that maps to something the bot says.<\/p>\n

A sample utterance<\/em> is something the user says to the bot that is defined in the interaction model to help the bot classify customer intent. For example, if you\u2019re creating an intent for opening a bank account, you\u2019d likely want to include utterances like \u201copen an account,\u201d \u201chelp with opening an account,\u201d or \u201cHow can I open a bank account?\u201d The idea behind sample utterances is that by defining a class of utterances with similar semantic content, the bot can use these to make an educated guess about what the user\u2019s goal is. Even if you don\u2019t define every possible utterance (and you shouldn\u2019t), the bot can guess what the user is trying to do.<\/p>\n

A slot<\/em> is a piece of information that the user provides in order to accomplish their goal. For example, if a customer wants to open a bank account, we need to know the type of account. We can use a slot to collect those account types, and name it something that builders will understand, like AccountType<\/code>. Slots can be either required or optional, depending on the use case. For example, you might need a required slot like BirthDate<\/code> to authenticate your user, but collect an optional slot type of AccountType<\/code> to disambiguate between the different accounts a user might have. Slot values<\/em> are the pieces of information that you want the bot to recognize as a slot, like \u201cchecking\u201d or \u201csavings.\u201d Synonyms<\/em> are alternate ways of saying a slot value, like \u201cISA\u201d or \u201cdeposit account.\u201d<\/p>\n

Finally, a slot corresponding utterance<\/em> is an utterance that contains a slot value, but doesn\u2019t contain an intent, such as \u201cto my savings account\u201d or \u201cit\u2019s for my savings account.\u201d In these utterances, you can\u2019t tell what the user is trying to accomplish without the context of the rest of the conversation, but they do contain valuable slot information that you need the bot to capture.<\/p>\n

The bot also has some available actions, ElicitIntent<\/code> and ElicitSlot<\/code>, which mean that the bot is either trying to capture the user\u2019s intent or the bot is looking to gather some slot information.<\/p>\n

Now that you\u2019ve defined the values for all these components and put all those pieces together, you\u2019ve created the first draft of an interaction model. Here\u2019s an example, complete with the bot\u2019s available actions.<\/p>\n

\"\"<\/a><\/p>\n

Turn user stories into intents and slots<\/h2>\n

From your user stories, you\u2019ve identified the use cases that you want your application to be able to help your users fulfill, such as blocking a bank card due to fraud or opening a new credit card. Make a complete list of all use cases that you developed. Now, it\u2019s time to work backwards from the use cases to create user intents.<\/p>\n

Start by getting a group of people together from all different teams of your organization\u2014business analysts, technical pros, and leadership team members should all be present. Ask each person to create a list of possible things that they might reasonably say to a human agent or to an AI application for help with their use case. For example, if your use case is to open an account, you might list things like \u201cI\u2019d like to open a bank account,\u201d \u201cCan you help me open a new account?\u201d or \u201cOpening a savings account.\u201d Be flexible with what you write. Have each person write 10\u201320 utterances per use case. Keep in mind the variety available in human language:<\/p>\n