Skip to main content

Save Field Component

Save Field Component

In this article we will explain how to save information within a conversational flow, what options exist to do so, and when it is convenient to use the Save Field component or a Save Field Smarton.

Correctly saving customer data is key to:

  • Personalize responses

  • Make decisions within the flow

  • Improve reporting and analysis

  • Avoid errors or inconsistent data

📌 What is Save Field?

The Save Field component allows you to store information within an information field, based on:

  • What the user says (input)

  • A fixed value defined by the flow

📍 Examples of information you can save:

  • Document / ID

  • City

  • Payment method

  • Status of a query

  • Internal identifiers

Configuration:

  • Select the field where the information will be stored.

  • Option to assign a specific value to all clients.

✍️ How can information be saved?

There are two main ways to save a field within a flow:

1️⃣ Save user input

The field takes the customer's actual response.

Example:

  • The bot asks: "In which city are you located?"

  • The user responds: "Panama City"

  • The flow saves that value in the City field

👉 Ideal when you need to capture explicit information from the user.

2️⃣ Assign a specific value ("burned" value)

The field always saves the same value, regardless of the input.

Example:

  • Field: City

  • Specific value: Panama City

👉 Useful when the data:

  • Is defined by context

  • Is fixed for the entire flow

  • Does not need to be asked

🔀 Save Field vs Save Field Smarton

Both options serve to store information, but are not used in the same scenarios.

🧩 Save Field Component

  • Saves a specific piece of data

  • Executes at a specific point in the flow

  • Ideal for:Direct inputsFixed valuesSimple and controlled logic

  • Direct inputs

  • Fixed values

  • Simple and controlled logic

🤖 Save Field Smarton

  • AI interprets the context

  • Decides what value to save and when

  • Ideal for:Open conversationsData that does not come in a single responseAutomatic saving without explicit questions

  • Open conversations

  • Data that does not come in a single response

  • Automatic saving without explicit questions

📌Tip: If the data is clear and structured → Save Field

📌 If the data arises from a more free-flowing conversation → Smarton

🗂️ Information Fields vs Flow Variables

Understanding this difference is key to not overloading the bot.

🟦 Information Fields

  • Persist over time

  • Are used for reporting and analysis

  • Impact performance if used in excess

🟨 Flow Variables

  • Only live during the conversation

  • Are not stored long term

  • Ideal for temporary logic or intermediate states

📌Practical rule:

If the data is only useful "at the moment", use a variable.

If you need history or reports, use a field.

📝 The importance of field descriptions

The field description functions as a prompt for the AI.

It indicates:

  • What to save

  • When to do it

  • What values are valid

A bad description can generate:

  • Incorrect data

  • Poorly populated fields

  • Errors in automatic decisions

📢 Best practices for Save Field

1️⃣ Define clear and specific descriptions

🔹 Avoid generic or ambiguous texts.

✅Correct example

  • Field: Payment method

Field: Payment method

  • Description: "Payment method chosen by the user. Possible values: 'cash' or 'credit'."

Description:

"Payment method chosen by the user. Possible values: 'cash' or 'credit'."

❌Incorrect example

  • "User's payment method."

2️⃣ Include clear storage conditions

🔹 Specify when the value should be saved and when not.

✅Correct example

  • Save only if the user mentions "cash"

  • Do not save if they mention "installments"

❌Incorrect example

  • "Save when the user mentions a payment method."

3️⃣ Adapt the field to the context of the conversation

🔹 Not all fields depend on a direct response.

✅Example

  • Field: Query made

Field: Query made

  • Description: "Save 'True' if the user made a query and received a successful response."

Description:

"Save 'True' if the user made a query and received a successful response."

4️⃣ Use structured and standardized values

🔹 Avoid long texts or free interpretations.

✅Correct example

  • Allowed values: "Resolved" / "Pending" / "Escalated"

Allowed values:

"Resolved" / "Pending" / "Escalated"

❌Incorrect example

  • "The user asked something and it was answered correctly"

5️⃣ Avoid saving unnecessary information

🔹 Not everything should be a persistent field.

🔹 Temporary data → flow variables.

📌 Real case:

A bot had more than 600 information fields, of which 30% could have been variables, affecting:

  • Reporting

  • Bot performance

  • Unnecessary data loading

Did this answer your question?