TinyMe: Open-source chatbot for personalized guidance

Like it or not, chatbots, in any form, are here to stay. They have become extremely popular in the last few years and can be found in almost every industry and website.

One of the reasons why chatbots are so widespread is that they save time and effort, which is achieved via automation. They provide a lot of possibilities like answering customer questions, collecting information from users or customers, automating repeating tasks, etc. Meet TinyMe, a conversational flow bot, a type of rule-based chatbot that engages customers in a conversation like no other.

Chatbot types

All these chatbots have different conversational complexity and, in general, can be organized into three types:

Rule-based bots

These bots use a set of predetermined rules and algorithms to understand and respond to user inputs. They are limited in understanding context and providing more complex responses.

Generative bots

These bots use machine learning algorithms, such as deep learning, to generate responses based on patterns in the data they have been trained. They are capable of handling more complex and open-ended conversations.

Hybrid bots

As the name suggests, these bots combine features from different types of chatbots, using a mix of rule-based and machine learning-based approaches to understand and respond to user inputs.

Here’s a great infographic depicting the history of chatbots and how they developed and improved over the years.

How TinyMe came to be

Around two and a half years ago, at the end of 2020, an investment company called VisualVest GmbH was looking for a way to integrate a chatbot into its own platform. This is when we entered the picture, and the ProductDock and VisualVest collaboration began.

They had a list of requirements for the chatbot, which included. The chatbot to be constantly updated by content creators with new, simple, and intuitive conversations. It can have targeted conversations depending on the webpage’s context (on which the chatbot is integrated). The most importantly, to have a chatbot with a great user experience

These are solid requirements, but VisualVest, being a financial consultant, needed to cover one more aspect—data privacy and GDPR. 

Most companies rely on third-party products to build their custom chatbots. Some do it by modeling it on a chatbot platform (PaaS, SaaS) and then integrating it into a website or an application—you can recognize this type of chatbot dialog by having something like “powered by XYZ” written below.

It’s important to mention that the data the customers provide while interacting with a bot will not be within the website domain but will sit on the PaaS platform, which is an important indicator of how limiting these bots could become, considering the data privacy policies and GDPR compliances.

Since the original project uses many open-source projects within its core implementation, it made sense to contribute to the broader community as well. A separate twin brother was built, with the learnings and difficulties from the original project adequately addressed. The name of the twin brother is TinyMe.

What is TinyMe?

TinyMe is a conversational flow bot, a type of rule-based chatbot that engages customers in a conversation like no other. It is an improved version of the bot we built for VisualVest. The guiding principle behind its creation is ensuring content creators and users find it enjoyable. It takes the form of a chatbot-like window that appears on the side of your landing page and is designed to attract your attention.

Additionally, it includes various features such as widgets, reactions, and interactions. You don’t even need to type what you want – it will guide you through!

What TinyMe does

TinyMe relies on tiny things to provide a richer experience. Plugins are included in the chatbot, enabling you to start building flows from day one. These include text with markdown support, powerful interactions such as sliders, buttons, single/multi-selects, and more.

The powerful dashboard combines three parts in one user interface: 

Powerful analytics that is easily connected to third-party data analysis tools via its extensive REST API.

An editor to create your flows.

A frontend for testing your available flows.

How TinyMe works

Under the hood, it uses a bundled NodeRED to create flows in a flow “programming” language. You don’t need to be a programmer to get started – the advanced plugin system contains nodes that you can drag and drop, fill in the configuration, and deploy.

The plugin system also extends a React application, which is where customers will interact with the platform. Plugins, which can be searched and installed on the live installation, allow you to customize the experience.

Here are a few screenshots.

Next steps

Stay tuned for the upcoming Part 2, where we will dive deeper into architecture and technical decisions.