How many times do i have to edit my model prompt ...
# support
t
I have edited the prompt right now almost the a thousand time followed https://gist.githubusercontent.com/bajajcodes/ff0bc6e126d4956fb422d47a397abeff/raw/67124092dfe0a193a9baa310a3e76a5c2f146524/prompt.txt but i still have to edit it every time adding a little tweeks every time ,its difficult because clients nolonger trust the process ,they get pissed off yet every scenario in make.com is correct but the prompt keeps missing the flow when booking appointment ,what is happening @Shubham Bajaj
v
@trx598 I'm sorry to hear that you're experiencing difficulties with your prompt and that it's affecting your clients' trust in the process. It sounds like you're facing a challenge with ensuring the prompt consistently handles the flow for booking appointments correctly. To better assist you, I would need to understand the specific issues you're encountering with the prompt. Could you please provide more details about the problems you're facing? For example, are there certain conditions or variables in the appointment booking process that the prompt is not handling correctly? Any examples of where it's going wrong would be very helpful. Once I have a clearer understanding of the issue, I'll do my best to offer guidance on how to improve the prompt for a more reliable performance.
You can continue this conversation by mentioning me in the message.
t
@Shubham Bajaj why is it when i say book an appointment after trigerring a tool it doesn't trigger the tool but says the appointmet was booked successfully,this is annoying and disheartening,i put gpt-4o-mini and the prompt is quite long most preventing things i already told it to prevent
s
Hey @trx598 , I completely get it that, it is disheartening when things like these happen because of that we are building blocks which will solve all the issues that you are facing right now. It will be mostly out by end of this month and of mid of next month. For now, all I can say is that you will need to work on your system prompt, use gpt-4o and use maxToken 0.3 to get the best result.
t
ok
4 Views