Hello, I would like to use AI at our

At a glance

The community member is interested in using AI to propose responses to customer support tickets at their helpdesk. They have over 10,000 resolved tickets that they want to use to expand a GPT model. The community member asks if this is feasible, and whether Llamaindex would be suitable. They also ask what data they should store, such as the ticket title, customer content, and comments from both support staff and customers. The community member wonders if it makes sense to note whether a comment is from a customer or support staff, and what format the data should be stored in, such as JSON or text files.

The comments suggest that this is feasible, but the structure of the tickets is important. One community member recommends trying a small knowledge base first to see the output quality. Another community member says storing the correct response is most important, as the customer data could "pollute the retrieval and results". The comments also mention that unstructured help docs can be used with vector search, but metadata or another method may be needed to associate the tickets and comments correctly.

RRomanito

Hello, I would like to use AI at our helpdesk. A customer submits a ticket and the AI would propose a response to their request. My idea is to expand GPT with data from our resolved tickets (there are more than 10,000 of them).

Is this even feasible?
If yes, is Llamaindex suitable for this? If not, what should I look at?
If yes, what data should I store? A ticket has a title, content written by the customer, and then there are comments from both support staff and the customer (when something needs to be resolved further). It makes sense to me to store all these things. Does it make sense to note whether the comment is from the customer or from support staff?
In what format should I store the data? I have access to the helpdesk database, so it's no problem to read and export it in anything. It should probably be clear to which ticket each comment belongs. So it makes sense to me to store it in some structured format. Is that correct? Is a json structure more suitable than exporting it to txt?

Thanks for suggestions 🙂

2 comments

TTeemu

I'd say it's feasible, obviously depends quite a bit on how the tickets are structured. Maybe you could try to make a small simple knowledge base with some examples and then seeing what the output quality is like

Definitely feasible, have used it for similar use-cases

I think storing the correct response is most important, the customer data there might pollute the retrieval and results

Generally with using help docs and vector search they can be unstructured but you might need to use metadata or another method for associating the tickets + comments correctly

RRomanito

Thanks!
We have documentation, but it does not cover all functions of app. On other hand we have lot of tickets with knowledge hidden there.
I want to present demo AI helper to colleagues asap and I dont have time (and knowledge) to create doc covering whole project. So I had a idea of "shortcut" by using existing questions and answers. But I dont know if gpt could handle this type of data corectly.

Add a reply

Find answers from the community

Hello, I would like to use AI at our