Store/Scripts/AI (GPT) Autoresponder

This is a Playground Script for Personal Bots to enable autoresponding to IMs. It operates based on ChatGPT and follows the instructions you provide. Tokens are used during its operation, more below.

How it works

1. You purchase the script. 2. The system redirects you to the Playground, where you can select the bot to which you want to add this script. This bot will respond automatically. 3. You need to draft instructions for your bot. You can find information about instructions below 4. Autoresponder consumes tokens. For more details on what tokens are, please refer to this link. 5. That's it! The script is activated & working. You can review the message history. Learn more about message history here. 6. You can deactivate the script at anytime.

What are instructions?

Instructions are the brains AI autoresponder, include all the information that a visitor might inquire about. How much does your service cost? How to access your library? What other products do you offer? etc. The instructions can be modified at anytime.

Since the AI autoresponder is powered by smart ChatGPT, you can even specify in the instructions the tone in which your AI autoresponder will communicate. You can instruct it to "Respond in a playful and charming manner" or "Answer with a confident and authoritative tone," and your AI autoresponder will adopt the specified character. You can always update the instructions later.

Important! The text of the instructions affects token usage. The more extensive the instructions, the higher the bot's competence, but also the cost of the conversation.

What is a token?

Token is a "charge" of AI you own, to be utilized by SmartBots AI engine. Think of token as a short word (~4 characters). Longer words may consume more than 1 token.

SmartBots AI engine spends tokens while processing your bot queries.

Token consumption

Each request processed by AI includes:

• User message (what your visitor/customer says)
• instructions
• All previous user messages (to keep a context)

For example: 1st message from resident Resident: "Hello!" Request to AI: "Your instructions" + "Hello" Bot response: "Hi!" 2nd message Resident: "Who are you?" Request to AI: "Your instructions + "Hello" (1st message) + "Hi!" (1st response) + "Who are you?" (new message) Bot response: "I am a bot"

How to purchase tokens:

Tokens can be acquired by visiting this page.

Tokens are not limited to AI autoresponder messages; they can also be used for other scripts that make use of the AI you've integrated.

Аccess message history

The dialogue history with each user is stored for a duration of 24 hours.

What this means:

• If you send a message to the AI autoresponder and it responds, the system saves both question & answer.
• If you engage the same bot again within 24 hours, the bot will remember your prior conversation.
  • If engaged outside of this timeframe, bot will NOT remember the prior conversation.
• The cost of each response comprises the expenses associated with the question, answer, and instructions.

Be aware that messages from all your AI autoresponders are aggregated here. You can connect multiple bots, which is particularly significant if you're managing multiple projects. Messages are presented in chronological order, based on when they were initially received.

The logs for each message contain all the necessary information:

• Bot name (the name of the bot that responded)
• Script name (the script that initiated the use of AI)
• Message (the initial message from the user in the SL, typically a line without hyphens)
• Sender (the name of the avatar who wrote the message)
• Instructions (the guidelines for the bot on how to respond, which may include a substantial amount of text spanning several paragraphs)
• Response (the bot's reply)
• History entries (indicating how many messages were retrieved from the communication history of this bot and this user)
• Tokens:
  • Prompt tokens: the number of tokens expended in the user's request
  • Response tokens: the number of tokens utilized in the bot's response
  • Total tokens: the overall number of tokens spent
  • Tokens remaining: the tokens that are left after the interaction
• Date

It's also possible to access the message history for 30 days, and the corresponding table is located on a separate page. This allows you to sort messages by username, facilitating your review of their conversations.