Chatbots work for VP Immigration all day answering potential clients' questions and book a paid consultation.
Full Service Chatbot
Web and Mobile
Always available customer service
VPimmigration law firm provides paid consultation to the clients, and clients can only book the appointments by calling them. As a person has to take appointments, VPimmigration can only take appointments during business hours, leading to losing clients. VPimmigration does not want to lose any potential clients, so they want to provide appointment booking through their website. They want to be creative and provide appointment booking on their website with an interactive chatbot. They also want to take intake sheets and provide immigration information through a chatbot. The amount of data they want to provide is enormous and context-based. The chatbot should also take payment from the clients before confirming an appointment, create an outlook calendar invite, and send a confirmation email with an intake sheet link.
VPimmigration commissioned NvisionKC to develop this complex chatbot. NVisionKC decided to leverage the existing frameworks and services to develop this chatbot. NVisionKC determined that all the chatbot requirements can be developed with the Microsoft bot framework and cognitive services. The backend of the chatbot is developed with Microsoft Bot Framework, and the Microsoft webchat library is used for UI. Developing this sophisticated chatbot requires custom-tailored solutions. NVisionKC ingeniously solved problems that were raised while developing the chatbot. The chatbot developed by NVisionKC has the following advantages or features.
Easier Data Entry
The chatbot must answer questions related to the immigration services, and some questions should lead to an appointment booking suggestion. The amount of data that the chatbot can provide to the clients is enormous. We found that Microsoft QNAmaker is suitable to store and query huge data. QNAmaker also easily integrates with the bot framework. QNAmaker can only store questions and answers and allows multi-turn questions and answers, which means an answer to a question can have further questions. So we requested VPimmigration to convert the data into question and answer format for chatbot conversation. Multi-turn questions and answers are so deep that VPimmigration cannot use a regular editor like MS word to enter the data. NVisionKC then came up with a solution to use mind map software to enter the data and export it to a tab-delimited text file. The file is then converted to QNAmaker JSON format and pushed to QNAmaker using REST API.
Elegant UI
In some cases, instead of just sending a text to the chatbot, we want to send tables, action buttons, take input as form, etc. Fortunately, the Microsoft Bot framework supports adaptive cards, allowing us to create and send custom cards to the chatbot, which are highly customizable. Adaptive cards are used to display the list of available appointment times, welcome message, and payment form.
Confirmation Email with Any Email Address
The bot sends a confirmation email to the customer after booking an appointment. The confirmation email also contains a link to the intake sheet. It lets users submit the intake sheet at any time. The bot sends emails to the clients using SendGrid API. SendGrid API allows the software to send emails with a custom registered email address.
More than a Chatbot
VPimmigration wants to take client intake form through the chatbot, but the number of fields inside the client intake is numerous, and it is not feasible to take through the chatbot. We created a webpage with the intake form and provided the link to the clients after completing the appointment booking. Luckily, the Bot framework is an extension of the asp.net core; it can serve web pages. We exploited the asp.net core feature to develop the intake form functionality. In this way, we don’t have to develop multiple applications. The bot itself can handle the intake sheet feature.
Secure File Uploads
Clients can also upload the files in the intake sheets, and this raises some security concerns. NVisionKC provided a secure file upload feature in the intake sheets. The bot stores the uploaded files in the Azure blob storage. We enabled windows defender for the blob storage and configured it to report any malware uploaded to the blob storage; in this way, we prevented clients from uploading a malware and VPimmigration from downloading a malware
Logs and Analytics
During an event of failure or compromise, logs allow us to determine the root cause of the problem. We deployed the bot in Azure and enabled app insights. The bot is configured to log every event to the app insights. The app insights gives information about the bot uptime and metrics. Even though app insights provide metrics, it doesn’t provide any analytics about the user actions. To get the user action analytics, we made a custom workbook. When the workbook is opened, it displays the stats of the user actions.
Low Development Time and Reduced Deployment Cost
Higher development times incur a lot of money. We created the chatbot in a couple of months by leveraging the existing technologies and third-party open-source libraries. The UI of the web chatbot is deployed in the Azure blob storage instead of using a web server, reducing the deployment cost. The chatbot itself can take intake sheets from the users without requiring another program. It eliminated the need to deploy another program which further reduced the deployment cost.
Modular
The design of the chatbot is modular. It is easy to add or remove a feature. Each skill of the chatbot is designed as a module.