Glad the spam is being dealt with.

I've posted two similar requests but both were marked as spam. If they are being reviewed please post one or the other but not both.

I didn't intend to create a duplicate & my second was an attempt to get by support request into the forum.

Also, if everything is being marked as spam perhaps a message saying your post is pending approval would be more helpful than responding to everything with a message explaining that 'this topic is marked as spam'. All that will accomplish is more duplicate requests to sift through.

Sorry Mike, the forum is receiving a lot of spam posts (tens per day) and we do have an automated spam filter turned on to try to catch them however the quality of it isn't perfect.

@ArghDave Have there been any discussions on the possibility of creating a forum that operates in a manner similar to Google Gemini, ChatGPT, etc ? Would that curve out all of the random spam issues ? I absolutely love this forum as it contains so much information that is helpful to me.

@Joseph.Rhodes - Thanks for the feedback. Can you clarify the experience you are seeking? Are you looking for an experience like using ChatGPT which is trained on the SupportForum and VTScada help files or are you looking for a user forum which is similar to the user forum for ChatGPT or Gemini?

@ArghDave Yes, I was looking for an experience similar to ChatGPT, Google Gemini which is trained on the SupportForum and VTScada help files. I, personally, prefer Google Gemini over others. However, I'm sure others have their preference as well. In using Google Gemini to experiment with how much it can provide about VTScada. I've noticed that it does not appear to be tailored exclusively to VTScada.

I suspect you can use Gemini today to get a lot of what you need and the correct prompt strategy should help you get results biased by a certain source. Here are some things to try today using Gemini:

Prompt Engineering Strategy

1 - Anchor Gemini to the Source
Explicitly tell Gemini to prioritize the VTScada Help site:

You are an assistant specialized in VTScada. 
Base all answers primarily on the documentation at https://www.vtscada.com/help/Content/Welcome.htm. 
If information is not found there, only then supplement with general SCADA knowledge.

2 - Narrow the Scope
Frame your queries so Gemini knows the context:

“According to the VTScada Help documentation, explain how MQTT is configured.”

“Summarize the section on SNMP in the VTScada Help site.”

“Using only VTScada’s official help pages, describe the steps for setting up redundancy.”

3 - Use Retrieval-Style Prompts
Gemini responds better when you act like you’re asking it to search and extract:

Search the VTScada Help documentation at https://www.vtscada.com/help/Content/Welcome.htm 
and provide a structured summary of [topic].

4 - Reinforce Weighting
Repeat the instruction so Gemini doesn’t drift:

Important: Do not rely on general SCADA knowledge unless the VTScada Help site 
does not contain the answer. Always cite or paraphrase from that documentation first.

Give that a shot and let me know what you think. I believe that our Western Canada distributor has done some AI training on the help files to make their own chatbot and has been fairly successful in that though I do not believe it is something they made public, more for their own interest.

Let me know if the 'Prompt Engineering' works for you. I know that we do have a team internally experimenting with Machine Learning and VTScada but am not aware of a help bot in the works.

@ArghDave
Wow, that is incredible..