Create automation with AI 🤖, a feature designed for further optimizing our productivity and meeting the ongoing demand for efficient solutions for our users. 🔗 Find the video tutorial at the bottom of the page 🎥
The 4 key steps to succeed:
- Determine the trigger
- Configure the role you assign to AI
- Select the data from your workspace for AI to process
- Insert and select a new column to integrate the response.
To explore different scenarios of data processing by AI,
Start here ⬇️
1️⃣ Click on the menu → ⚙️ Automation
To access the list of various automations.
2️⃣ Select → Automation with Timetonic + AI
A new window opens, giving you a view of existing automations and the possibility of creating a new one.
3️⃣ Click the button → ＋ Create an automation
A new window opens:
⚙️ This is the configuration form for automation
-
4️⃣ Name the scenario
for automation, to easily identify it later
For example, "AI Response for client fields".
-
5️⃣ Select a trigger to start the automation.
When: the field is modified.
Typical examples of triggers:
- Status change
- If the checkbox is ☑️
- A formula based on today's date
- Others…
6️⃣ Add an additional condition if needed
7️⃣ Click on the condition → "Then"
to access the automation actions window
8️⃣ Select AI from the available choices
A new window opens:
⚙️ The form to connect your AI and configure its role and actions.
⬇️ Here are the details expected in each field, in order
9 🔑 Insert your API key to enable TimeTonic to interact with AI.
If you don't have one yet, visit these links: (API keys 🔑 Open AI) or (API keys 🔑 Mistral) to create a new key or retrieve your existing one if necessary. Then copy-paste the API key into the designated field. (Requires a paid Open AI or Mistral plan)
10 Insert your Organization ID for your AI account (Optional)
Retrieve your Organization ID in the account settings of OpenAI or Mistral
11 🤖 Choose the AI model
For Chat-GPT:
- Chat-GPT-4o: Prioritize the latest version available from OpenAI.
For Mistral:
- Mistral-tiny: Ideal for simple tasks that can be performed en masse, such as classification, customer support, or text generation.
- Mistral-small: Ideal for intermediate tasks requiring moderate reasoning, such as data extraction, document summarization, or writing product descriptions.
- Mistral-medium: Ideal for intermediate tasks requiring moderate reasoning, such as data extraction, document summarization, or writing product descriptions.
[For more details, consult the documentation of OpenAI or Mistral models and their specifications]
12 🤖 Describe what you expect from AI
Technical Language: Writing a Prompt
In technical language, you will often hear the term: "Writing a prompt."
Define the Role
Specify the role you are assigning to the AI, then qualify the objective, structure, verbal form, and style.
For example: "You are an expert in content creation for SEO optimization."
13 ❓ The Question
Explain the action the AI must perform, enriching it with elements from your field if necessary. For example, insert the field of an [email] to be processed.
-
Use the character "$" and select the field.
-
Or simply write the prompt for the action the AI should execute.
Example Prompt: "You will receive a text corresponding to the tutorial video information for TimeTonic. Based on this information, generate a concise and professional description and title."
14 📎 Attachment
Insert an attachment to extract data from an image.
15 🪙 Tokens
Define the value to control expenses on the subscribed plan. Set the maximum number of tokens the AI can use per response to manage costs.
Tokens can be considered as units of words for natural language processing. In English text, one token equals approximately four characters or 0.75 words.
16 & 17 Choosing the Response Format
ℹ️ Pre-configure the columns
Create fields of type: Single-line text, Multi-line text, SmartText, where AI-generated responses will be stored.
Selecting Text
A new field appears: Enter the character "$" or use the selector ⌄ to direct your response to this column.
Selecting JSON
Define the requested names to extract in JSON format within the "JSON Property Name" fields.
-
Then, insert the dynamic value character "$" or the selector ⌄ for "Destination Fields" from your table.
🚀 Once all fields are configured, validate your setup and perform a test extraction to ensure that data is correctly analyzed and stored in the appropriate fields.
⚠️ Field names must be written identically (uppercase, emoji, syntax) across the prompt, JSON property field, and destination field to avoid extraction errors.
View all available automation actions in TimeTonic ⚡️