This component is used to validate the client's response after interacting with a Message or Template component.
Purpose and utility:
If used after a message with buttons, it will automatically generate an execution branch for each button.
If used after a message without buttons (free text), it will create a single branch called "Other" by default.
It is always possible to manually add more branches and configure the evaluation based on content.
Evaluation of entered content:
You can add a manual branch and compare what the user wrote with a specific value (e.g., date of birth, email) or with the message body. The value to compare can be:
A. Multimedia:
Allows formats such as image, audio, video, or document.
If you need to store the file in another repository, you must use Base64 or URL format, assign it a name, and use the HTTP Request component (POST method) to send it to your CRM or Database.
B. Message Body (Text): Features various logical conditions:
Starts with / Ends with
Is equal to
Greater than / Less than
Contains / Contains date / Contains time
Matches regex (regular expressions)
Is empty or has not been set
To evaluate this text you can use:
Keywords: Define exact words to direct the query. For example: if a user wants to pay, use "pay" or "make a payment". (Tip: Keywords do not filter exact numbers; for numbers, use regex).
Atom Magic (AI): Describe the scenario for the AI to process it. For example: "Here the user is likely looking to pay their invoice".
⏱️ The "No response timeout" option
Allows you to define how long the bot will wait before proceeding if the user does not respond.
Where is it? Within Evaluate Response, the "No response timeout" field is enabled. You can set minutes, hours, days, or weeks.
What is it for? To redirect the user (e.g., Flow Jump), close the interaction in an organized way, show an alternative message, or escalate to an advisor in case of abandonment.
Best practices: Use at least 2 minutes to allow reading time, redirect to a clear option (e.g., "Type 'Advisor' if you need help"), simulate the interaction to validate, and avoid infinite loops (do not redirect back to the same message).
🚨 Advanced considerations and common errors in Evaluate Response
🤖 Keywords not recognized: If you do not include a welcome message in free text flows, the bot will not respond until it receives an exact word. ✅ Recommendation: Add an initial message that guides the user (e.g., "Hello, type 'accept' to start").
🔁 False assignment promises: If you send the message "An agent will assist you shortly" but do not include the Assignment component, you create false expectations. ✅ Recommendation: Back each message with a functional component.
🔌 Silent failures in external APIs: If an API (e.g., identity validation) fails, the bot may continue without indicating the error. ✅ Recommendation: Implement fallback routes with error messages or automatic retries.
⚠️ Unassigned conversations: A very common error is the absence of a "no response" branch. If not included, the user remains in infinite pause. It also occurs if the assignment group only has inactive agents or if the flow jump limit has been exceeded. ✅ Recommendation: Always include the "no response" containment branch, review your agent groups, and test the flow before publishing it.
You have full control of the conversation! 🚀
Mastering response evaluation, intentions, and assignment allows you to create error-proof flows, ensuring that each customer receives the correct information or timely human attention. ✅