Back

AI Chatbot for Queue Management

What kind of problem is solved with the project

The solution was aimed to provide a convenient tool for the stylists which they could give to their end customers to manage appointments. The automated queue management enables higher level if time utilization and more convenient service for the end customers

Project description

The platform has a scheduling interface for stylists, dedicated phone number for sending SMSes, and administrator management pages

The bot itself has a rather complicated AI functionality which gives an opportunity to select the needed stylist service, agree about the time (which means checking the time is free from other appointments), request to stay in a waiting list, reschedule or cancel the appointment.

The ability to automatically manage the waiting list is the most value added. Stylist is not involved in the process of finding the replacement the client who has canceled the appointment. The system automatically maintains the waiting list and starts searching the available client just once there is a cancellation. That is the source of a significant efficiency boost as this reduces not effective stylist time.

The administration part has plenty of interesting features for the stylist as well, like internal CRM for managing clients, own calendar management with the ability to setup working hours and blackout dates.

Duration: 3 months.

Development team

Front-end(Angular), Back-end (PHP), AI Chatbot developers, QA Engineer.

What we built

The solution includes

Twilio SMS, Facebook Messenger (for testing), web client for Admin panel for stylists

The services and Frameworks we used

AWS Lambda – a compute service from that allows running code without creating, setting up and managing servers. The AWS Lambda helped to place code that provides the addition of stylists’ services name from the database to DialogFlow

AWS API Gateway – a service that allows to create and maintain scalable API service, that can interact with any other Amazon service (Lambda, S3, App running on ec2 instance etc.) in the project.

Node.js – is a server-side, asynchronous, event-driven runtime environment, that allows us to build a server-side application using javascript syntax with the most recent ES standards support.

Botkit – is a Node.js framework, that has implemented some easy and convenient methods for working with different bot platforms and programming flow of the dialog with complicated business logic and various API integrations support.

DialogFlow – the platform is necessary to gather and recognize a client’s intent and produce human-like responses.

PHP API were used to create the backend project part. It allows us to focus not only on the framework code but on the business logic as well.

Development process

Firstly client target was to build the simple bot that allows for making appointments and changing the time or services in case of need. We created the logic and backend using Node.js and Botkit platforms to hold the conversation.
Then we added additional tools to make the bot more beneficial for clients and helpful for users. One of them is push notifications regarding available time frames when it comes free. Using DialogFlow we provided the most proximate to human speech conversation flow.

Bot architecture

Technical challenges

Cleaning content that was gathered by DialogFlow. Using Node.js project part developers deleted unnecessary content;
The edition of existing clients’ data. The DialogFlow isn’t able to provide the solution. Developers moved this project partly to Node.js and created a transition to the necessary part with data.

The result

Technical result: We built easy-to-use chatbots to make appointments quickly without additional actions like calls or app downloading. The bot manages the queue to a specialist without human assistance. It helps to utilize the time and increase revenue.

Business result: We satisfied customers’ demand on the ability to edit users data during the chat with the bot. We created a chatbot that holds human-like conversations 24/7.

Still have questions?

We will give a preliminary assessment within 3 minutes (fill out the form) and are ready to start the project within 4 working days.