I've been hearing so much about Chat GTP recently and how the AI can naturally answer questions, etc.
It got me thinking, would it be possible to integrate Chat GTP or some other AI into osTicket?
I have no idea at all how to even begin doing something like that, but there is so many answered questions in the osTicket database from customers that surely it would be enough info to educate the AI about how to answer queries.
Anyway - just wondering if anyone has thought about it?
Thank you 
Chat GPT is still very new but something we have on our radar for v2.0. Currently, osTicket doesn't have a robust API at all so it'd be very difficult to integrate with such systems. However, in v2.0 we will have a fully RESTful API that will make it much easier to integrate with external systems. Please stay tuned!
Cheers.
Nice! Thanks for letting me know, I'm really excited about this now! 
After more than a month, is there more information about API integration?
There has been no change on this.
loginid To be quite honest, I don't think such an integration would be earth shattering in the context of general user support - unlike for example an integration into Course material on Moodle for students to get almost tutor-like explanations on different topics.
In this space you would find very little separating a ChatGPT integration from a solid Heuristic or Fuzzy search engine on top of a well maintained Knowledge Base - which I think OST handles well enough.
The earth shattering bit will come when users can have a localised extension of such an integration that could actually perform the corrective/support tasks for the user. Or in remote systems, it is connected to the supported platform to perform the aforementioned tasks - thereby stepping past being informative into active. We do this now with intent driven bots and bot platforms, but what ChatGPT brings is just a smoother, more natural interaction between user and platform.
My 2cents...
We have over 10,000 closed support tickets from which ChatGPT can learn how to identify the symptoms described by the customer, consider the possible diagnosis, and prepare a suggested response in a professional and courteous manner. We already use the knowledge base and canned responses, which is useful for recurring issues of the same type, but ChatGPT can assist with pattern matching. Suggested responses, with links to potentially relevant past tickets, can support the training and productivity of customer support staff.
I can work with the dev team to provide a corpus of data and assist with beta testing.
I think this would be a fantastic addition that could cut down common 'simple' requests - or even just to reply and route tickets to the appropriate place.
Hi ntozier , @KevinTheJedi do you know if the version 2 is coming anytime soon?
You can read my recent comment on this subject here:
Cheers.
I believe that integration can be achieved step by step.
Full Integration: One thing is for the operator to be able to get an answer tailored to the customer’s specific problem directly, and yes, this case requires a complex integration work, and it would be necessary to train the AI with the entire archive of questions/answers.
Light Integration: Another thing, however, is to put the operator in a position to respond more quickly with advanced tools.
My project/suggestion:
Given that the help desk I manage caters to an international clientele that uses our e-commerce service, our operators often have to deal with complaints in different languages that address sensitive issues like "money."
For this reason, waiting for version 2.0 of OsTicket, I created a script that, when text is selected, displays a small menu next to the selected text where the operator can:
The very nice thing is that by using the innerHTML method, the operator can select and translate the content of the user's ticket into their own language without copying/pasting into external translators. 
Demo:
I wanted to package the code in a plugin, but unfortunately I can't make $ost->addExtraHeader(<script>) work in a plugin.
Integrating ChatGPT or another AI into osTicket sounds like a cool idea! I’ve worked on something similar where we used AI to handle support tickets in a different system. It was a bit of a project, but getting AI to learn from existing FAQs and customer interactions was super effective.
For osTicket, you could start by exporting your ticket data and using it to train an AI model. There are tools out there that can help bridge the gap between AI and ticket systems. A manufacturing ERP solution I worked with also used AI to streamline support, which made the whole process smoother and more efficient. It’s definitely worth exploring—if you need more detailed steps, checking out some integration guides or forums might give you a clearer path.
