How to plan a chatbot for your organization
Chatbots have been around for quite a while, but lately, most of the enterprises have started adopting chatbots as a new channel of communication and to reach customers with better efficiency. Before you start building a chatbot for your enterprise, you need to plan. In this article, I have mentioned the steps which can help you have the right mindset while developing a bot for your organization. Purpose of chatbot Before you begin building a chatbot, you should know the reason for building it. In order to know the reason, you will have to find answers to these questions: What problem will the bot solve? What is the reason for building the bot? What kind of requests or queries will the chatbot handle? For example, you might need a helpdesk bot to reduce the queries received by your helpdesk team or a sales bot to provide the team with required data or to set up meetings. There are various fields in which chatbot can be integrated. Decide the purpose carefully and if your chatbot has to solve more than one purpose split them and build conversation for each purpose individually. Target Audience Deciding the audience is the second step while building a chatbot. It is essential to understand the way people of a geographical area converse so that the bot can use the words and phrases used by the target audience. Also, one must understand the age group of the target audience and whether the chat must be friendly or formal. All these factors play a crucial role in building interesting conversations to engage the target audience. Building engaging conversations To be able to construct engaging conversation the purpose of developing the chatbot has to be clear. For example, if the bot is for real estate sales, then the bot should be able to build interesting conversations around this topic. Here are some tips for creating a query data set that can engage the users in interesting discussions. The conversations should not sound like FAQs or queries and answers written to the point. The communications should be seamless, and the user should not realize the presence of a machine behind the bot. Create category-based use-cases. Below is an example of a use-case. Bot: Hello, how can I help you? User: I am looking for a property Bot: oh! Can I know the purpose of the purchase? For residential For business User: Residential Bot: Are you looking for a House Flat Villa User: Flat Bot: Ok, for how many members are you looking for? User: 4 Bot: which location would you prefer? (Options can be displayed) User: Mumbai Bot: Here are some 2-3 BHK options. Have a look. The conversation can go long based on user inputs and queries. All the use cases of the bot have to be identified and then split into queries. Once the queries are identified note down the answers/responses for these queries. Some queries can be answered by giving multiple options to users. Deciding the right technology for your chatbot The technology of the chatbot can play a significant role in managing the cost and timeline. Currently, there are many technologies available like Microsoft Bot Framework with LUIS, Dialog Flow with Node JS or MBF, IBM Watson Conversation, Wit.AI to name a few. Based on the technologies your organization has currently quipped the technology for chatbot can be decided. It doesn’t end there… Although chatbot technology is extensively used, building a chatbot is a complex task as numerous aspects have to be considered in order to create interesting conversations. The process of building a bot does not end once it is deployed, but the bot has to be observed and retrained for the queries that are not present in the datasheet. Our team of experts at Advaiya can help you in chatbot implementation which can be customized and integrated with existing services in your organization.
How to convert FAQ into a Bot using Azure Bot Service
In my last article “Build FAQ Knowledge Base with QnA Maker”, I have explained steps to create knowledge base from existing FAQ pages using QnA Maker service. In this blog as a next step, I will explain how we can convert this FAQ knowledge base into a chatbot using Azure Bot Service. To use Bot Service, if you have at least an Azure Account. If not, you can start one for free. To create a Bot, follow the below steps: Log into the Azure Portal Click on New > AI + Cognitive Services > Web App Bot A new blade will open with information about the Web App Bot. In the Bot Serviceblade, provide the requested information to create a Bot Service: Bot Name: This will be our Bot name and used to respond to Chat with the user Subscription: Microsoft Azure Subscription, you can very well try this out with the free trial subscription also. Resource Group: This is a logical grouping of all Azure resources. Bot service is part of Azure resources it must be associated with either existing resource group or create a new resource group. Location: Select the data center region you would like to use Pricing tier: This defaults to S1 Standard, change this to FO (10K Premium Messages) Or else you will pay for the service App Name:It should be unique application name for the bot which consists of alphanumeric and underscores characters only up to 35 characters. The App name cannot be changed once the bot is created. Scroll down and fill remaining details: Bot template: As we are using FAQ knowledge base which is in question and answer form, select Question and Answer App service plan/Location: You will need to create a new one or use an existing. Azure Storage:This is a data storage account. You can create either a new or use an existing one. Application Insights:This service is not free, so be careful about turning it on. Microsoft App ID and password: Select Create New option if you need to enter a Microsoft App ID and password manually. Otherwise, a new Microsoft App ID and password will be automatically created for you in the bot creation process. See the below image: Click Createto create an Azure Bot. It will validate provided choices, and on completion, it will begin the deployment. On completion of deployment, we would have Azure Bot service. Now to consume the knowledgebase we have created using QnA Maker service, we need to connect both services. To connect the Azure Bot service, select your bot, it’ll be the one which has a TYPEof Web App Bot From the left menu, click on APP SERVICE SETTINGS > Application Settings. Scroll down to App settings There are two blank settings called QnAKnowledgebaseId and QnASubscriptionKey. Provide your valid QnAKnowledgebaseId and QnASubscriptionKey from your QnA Maker service. Click Save To test this newly created Bot, click on Test in Web Chat on the left side and ask queries based on the knowledge base. By default, bot will be configured and running for Web Chat To check all the channels for this bot, click on Channelson the left menu. There are more channel options available like Teams, Skype, Cortana, etc. To test a chatbot in Web Chat, click on Test in Web Chat option under BOT MANAGEMENT section in the left panel Now enter a message and bot should respond leveraging the knowledge base created using QnA Maker service. Note: If the Web Chat control is not fully loaded after a few minutes, try refreshing the page. In this blog, I have explained how you can create a basic Web App Bot by using Bot Service and tested the bot’s functionality by using the built-in Web Chat control within Azure. I hope the entire blog series helped you understand and explore the capabilities of Azure Bot Service and QnA Maker service. You can view the entire blog series here: Everything about Microsoft QnA Maker Build FAQ knowledge base with Qna Maker How to convert FAQ into a Bot using Azure Bot Service