Algorithm Suggestion for a Ticket Routing System Using Machine Learning

Background information

I am a beginner in ML, let me start there. I'm trying to implement an intelligent system capable of routing a ticket (in a ticket system) to the appropriate location depending on a few parameters, 5-10.

How should the algorithm work?

For example, the ticket "1234" is sent to "accounting" by a human, normally by reading the description and title of the ticket (search for keywords). The ML algorithm should learn where tickets go based on where similar tickets (based on the same keywords) have been used before.

What I've tried

I have implemented a very simple NN in JavaScript using the sigmoid function to predict the discrete type of outputs yes / no. If I remove the sigmoid function, I could predict where the ticket should go, based on the keywords converted into parameters, perhaps using a linear regression.

The problem

I do not know how to turn keywords into numeric parameters (or vectors?) That can be integrated into a simple linear regression implementation. With my limited knowledge, I'm not sure if linear regression is the way to go, but it certainly sounds like that. Linear regression is also a simpler algorithm that I can implement myself.

Is linear regression the solution? How to transform a paragraph of arbitrary length (ticket description) into meaningful keywords to use by the algorithm?

Ideally, it would be JavaScript, but I do not see any tags for JavaScript. Odd.