Scheduler chatbot

This project involved implementing a chatbot on an existing platform. The project in question enables the creation of event calendars, and the association of speakers and activities. The chatbot had to be made in Typescript and enable the user to create/modify/delete events using text or voice queries.

To develop this chatbot, I first had to familiarize myself with the data structure used in the project. Modifying the database wasn't so easy due to the complex data structure which, instead of storing events one by one, stores them in groups obeying the same rules.

Once I was familiar with the data structure, I listed all the actions the chatbot should be able to perform and divided them into categories.

I then used two GPTs queries: one to determine the purpose of the query (category), and the second, chosen on the basis of the first, to format the user's query into a usable format for performing the desired modification.

I then processed the query response, formatted it in a format compatible with that of the database, and made the desired modifications. If the modification is not possible, I send feedback to the user via the chatbot to explain the problem, for example, if a piece of information is missing.

I created a robust chatbot capable of handling complex commands. As you can see, here, I created a specific activity to a precise location each tuesday and thursday of a month and it successfully created the events:

The chatbot can also handle complex deletion and modification requests.