RT 6.0.3 AI plugin expectations

Hi,

We created a test instance of RT and upgraded it to 6.0.3 in order to test the AI plugin.

RT 6.0.3 is much nicer, but the AI plugin is not really useful at this point. It is true that it simplifies text proofreading and similar tasks, but the implementation is not very clean. For example, the “Adjust tone/voice” option takes your text and processes it, but in the process it makes a big mess of the formatting (losing spaces, line breaks, etc.).

At this point, it is simpler to copy the text to your preferred AI chat and then paste the output back into RT.

The AI query feature for searches is similar in this context. It takes a lot of AI processing to create a simple query that would have been easier to create using the old tools.

What would it take to use the real power of AI? For example, could we perform a full scan of the ticket database so that an AI agent could provide suggestions for responses based on previous tickets? That would be much more useful.

The summary and sentiment options don’t seem to be very practical either. They generate a large summary that essentially describes everything included in the ticket, rather than providing a useful summary.

Has anyone else been testing these AI features? Am I missing something that could actually be useful?

I haven’t started looking at the AI module yet, but we already have plans for it. I see the shipped module as a starting point and expect to have to modify the prompts to better suit our exact use case. For the summary method you mentioned, you may want to modify the prompt to ask for more brevity/conciseness.

I’ll add more once I’ve had a chance to evaluate it myself.