Following the official launch of ALF v2, you can choose to use either ALF v1 or v2 from October 31, '25 through November 27, '25, and you can change the version at any time.
Starting November 28, '25, ALF will be integrated into v2, and only v2 will be available for use.
Please click the [Use ALF v2] button in [AI Menu] > [ALF v2].
Once the version change is complete, the ALF v2 tab will display 'ON' and the v1 tab will display 'OFF'.
Please click the [Use ALF v1] button in [AI Menu] > [ALF v2].
Once the version change is complete, the ALF v1 tab will display 'ON' and the v2 tab will display 'OFF'.
Important Notes When Changing Versions
Please double-check the ALF settings of the version you are switching to, as the change will be applied immediately to your chats.
All newly started ALF chats will use the changed version for responses.
ALF chats that are already in progress will continue to use the previous version for responses.
The Deploy menu allows you to review all ALF settings that will be applied to the customer support(chats) in one place, and use the preview screen to test the actual ALF's support experience. You can also view ALF settings for Channel Talk Chat as well as for each external service.
You can set the basic profile for the Customer ALF that will operate in chats through Channel Talk message, and review and manage the rules, knowledge, and tasks applied when providing responses, all in one place.
Here, you can set the basic profile for the Customer ALF that will be applied to chats through Channel Talk message, as well as whether to display response sources to customers.
Profile / Name : You can set ALF's profile image and name.
Customer-facing source labels: You can set whether to display the source of the knowledge ALF referenced when providing a response to the customer.
When you turn on this option, it helps reduce misunderstandings and increase the credibility of ALF’s answers.
Sources are shown to customers only when the answer references published Space articles or website knowledge.
Even if the option is turned off, all referenced knowledge will still be visible in the Inbox on the Channel Talk admin panel.
You can view at a glance all the rules, knowledge, and tasks that ALF uses when responding to customers in Channel Talk chats. You can check which settings are currently applied, and click any item to go to its settings page for further edits.
Quick Deploy is a feature that allows you to use Customer ALF directly in customer support without setting up a workflow. When a customer starts a chat, a message greeting them and a chat input field for conversing with ALF will appear immediately. Once you click the [Deploy] button, the Quick Deploy will be applied right away.
Greeting message
Enter the first greeting message that will be shown when a customer clicks the chat button. If you don’t enter one, the default greeting set for the channel will be displayed automatically.
By using [Translate], you can automatically show the greeting in the customer’s language.
Allow chat to open
When a customer clicks the [Not helpful] button or requests to connect with an agent during a conversation with ALF, you can choose whether to actually connect them to a live agent.
ON: Displays a button for the customer to connect with an agent.
OFF: Shows a message informing the customer that agent connection is not available.
Advanced Settings
Filtering - Chat InfoYou can add chat start page conditions to make the Quick Deploy feature appear only on specific pages. Try setting different Quick Deploy visibility options based on the customer’s access page.
Example: Enable Quick Deploy on the main page, but disable it on the payment page.
Filtering - User InfoYou can set whether to apply Quick Deploy based on specific customer conditions. Provide a customized ALF experience for different customer segments.
Examples:
Membership status (member / non-member)
Customer grade (VIP, regular, etc.)
Customer tags (purchase history, items in cart, etc.)
Language (device or browser language)
Scheduling:
You can set up Quick Deploy based on business hours.
Example: During business hours (09:00–18:00), run a workflow that connects customers to an agent; outside business hours, activate ALF Quick Deploy.
You can see and manage the workflows connected to and operating with Customer ALF at a glance. It is easy to check which workflows have the Customer ALF response set, and you can click to navigate to each workflow's settings screen.
For chats coming from external services such as KakaoTalk, Naver TalkTalk, or Instagram, you can set whether to display sources for Customer ALF’s answers. You can also review and manage all rules, knowledge, and tasks applied when ALF responds at a glance.
Customer-facing source labels:
You can set whether to display the source of the knowledge ALF referenced when providing a response to the customer.
When you turn on this option, it helps reduce misunderstandings and increase the credibility of ALF’s answers.
Sources are shown to customers only when the answer references published Space articles or website knowledge.
Even if the option is turned off, all referenced knowledge will still be visible in the Inbox on the Channel Talk admin panel.
You can easily review the rules, knowledge, and tasks applied when Customer ALF responds for each external service, all in one place. You can view the currently applied settings, and clicking on them will take you to the corresponding settings page where you can make further edits.
You can see and manage the workflows connected to and operating with Customer ALF at a glance according to each external service. It is easy to check which workflows have the Customer ALF response set, and you can click to navigate to each workflow's settings screen.
On the Preview screen in the Deploy menu, you can test the ALF chat with the current settings for free. After selecting the service you want to preview (Chat or Call), you can check if ALF is responding correctly by looking at the Customer view and the Chat log.
You can see the screen that is actually shown to the customer when ALF responds to your test inquiry.
In the chat log screen, you can check the rules applied when ALF generated the response and the Agentic Search* Log from the time the answer was created.
*What is Agentic Search?
Agentic Search is a method where the AI acts not as a simple search engine, but as an 'active agent that performs the search,' autonomously finding and processing the necessary information. It means the AI understands and reconstructs the user's question, performs additional searches if necessary, and explores, organizes, and summarizes multiple sources to create the final answer.