AI machine learning from your website

Artificial Intelligence (AI) and Machine Learning (ML) are advancing the field of chatbots very rapidly.

At 3d2, we incorporate various AI/ML models to improve the effectiveness of the support the bots are giving. Although it is called machine learning, the learning is happening at various levels on the platform.  

Machine Learning (ML)

The ML area most frequently used is the natural language recognition model. The ability to recognise phrases and the context in which they are used are critical in analysing what the user is asking. This learning and improvement over time are at the base of the platform.

Author’s / Caretaker’s responses

The second area of learning is from the author’s or caretaker’s responses. When a user emails the company’s support email, the bot is learning from the responses that the caretaker is supplying to the user. Only the responses that were not learned yet, are passed to the caretaker of the bot. This reduces the amount of time spent by the caretaker as the bot is learning to answer more questions. 

User’s level

The third area of machine learning is on the user’s level. The more users that interact with the bot, the wider is the variety of samples that the bot uses to learn how to respond. This level of ML means that over time the support bots effectiveness increases without anybody’s effort. Automated learning is built-in.


The main goal is to get the autonomous support bot to take the lion’s share of the user queries and give time back to the caretaker. The second advantage is that the users are helped a lot quicker since they don’t have to wait for answers from the caretakers in the form of email replies for most of their queries.

Fourth area of learning

The wider experience, which is rich in content and context, is teaching the user that it’s quicker to ask the bot than to email the support desk. So finally it trains the users too as the fourth area of learning.