We provide a statistics feature that allows you to check, at a glance, the chat involvement rate, the resolution rate of meaningful answers to customer questions, the CX score, and the estimated cost savings achieved through ALF.
Existing ALF statistics can be found in ALF v1 and will only be provided until November 27, 2025, which is when ALF v1 is discontinued.
Please set the period you wish to view.
ALF statistics can be viewed for a minimum of 1 day and a maximum of 31 days.
You can view the statistics of the past ALF v1 by clicking the [Previous Data] button.
Downloading of ALF v2 statistics will not be provided until the statistics renewal scheduled for the end of December, 2025.
The percentage of total chats in which ALF was involved.
Involvement Rate Calculation
= [Number of chats ALF was involved in] ÷ [(Number of chats ALF was involved in) + (Number of chats ALF was not involved in)]
The percentage of chats ALF was involved in that did not get connected to an agent.
Use knowledge and tasks to increase answer accuracy, improve the resolution rate for inquiries, and effectively lower the agent connection rate.
Resolution Rate Calculation
= [Number of chats ALF was involved in that did not connect to an agentALF] ÷ [Number of chats ALF was involved in]
Compared to Total: The percentage of total chats that ALF resolved without agent response.
= [Number of chats ALF was involved in that did not connect to an agent] ÷ [Total number of chats]
A score that the AI evaluates itself for ALF's answers.
Score rated by AI. Out of 5, considering customer sentiment, resolution, and response quality.
This is the estimated cost saved because ALF responded instead of an agent.
Calculate the efficiency of ALF directly by substituting the average labor cost and the number of chats handled by the agent.
Period: Reflects the period selected in the statistics.
Chats resolved by ALF (A): he number of consultations actually resolved by ALF during the selected period, which is the same as the [Resolved by ALF] metric in the consultation processing funnel.
Agent cost per chat (B): Average agent salary ÷ (Average daily chats ÷ Number of agents × 30 days)
Average Agent Salary: Enter the average monthly salary of the agents to be used as the basis for calculation. The default value is 2,500,000 KRW.
Number of Agents: Enter the number of agents to be used as the basis for calculation. The default value is 5.
Average daily chats: The value obtained by dividing the actual number of consultations generated in the channel during the selected period by the number of days in that period, calculated as: '[Chats] in the consultation processing funnel ÷ [Number of days in the period]'.
The Chat Processing Funnel is a statistic that shows, step-by-step, how much ALF was involved in the total number of chats and whether that involvement actually led to resolution. It allows you to visually check the chat flow, making it easy to understand the performance of AI chat automation.
Aggregation Criteria
Chat: Chat in which an agent or ALF was involved. (During the CBT period, chats where only the workflow was executed are not aggregated.)
ALF Involvement: Chats where the customer sent one or more messages to ALF
ALF Not Involved: Chats where the customer has never conversed with ALF.
ALF Resolved: Chats where ALF was involved and an agent did not respond.
Agent Handled: Chats where an agent sent one or more messages.
Clicking each chat processing funnel allows you to view the list of chats included in that funnel.
Clicking each chat allows you to view the chat content on the right.
List by Selected Funnel
ALF Involved: Shows only chats where the customer has sent at least one message to ALF.
ALF Not Involved: Chats where the customer has not interacted with ALF.
ALF Resolved: Chats where ALF was involved and a manager did not respond.
Manager Responded: Shows only chats where a manager was connected after ALF's chat and sent at least one message.
Positive CX Score: Shows only chats with a CX Score of 4 or higher.
Neutral CX Score: Shows only chats with a CX Score of 3.
Negative CX Score: Shows only chats with a CX Score of 2 or lower.
Chats currently being evaluated for CX Score are not provided in the chat list.
Information Provided
User Chat: Customer name is displayed.
ALF Involvement: Indicated as 'Involved' if ALF responded at least once, otherwise 'Not Involved'.
Manager Response: Indicated as 'Responded' if a manager responded at least once after ALF's chat, otherwise 'Not Responded'.
ALF CX Score: Score is displayed from 1 to 5.
Chat Tags: Tags attached to the chat are displayed.
Team: The name of the assigned team is displayed. If not assigned, it shows "None".
Assignee: The name of the assigned manager is displayed. If not assigned, it shows "None".
Followers: The names of team members added as followers are displayed. If none are added, it shows "None".
Service: The channel through which the chat originated is displayed. E.g., Channel Talk message, Phone, Kakao.
TOP 20 Most Referenced Knowledge
Check the frequency and resolution rate of knowledge and tasks utilized by ALF to further enhance ALF. By increasing ALF's answer accuracy, you can boost inquiry resolution rates and effectively lower manager connection rates.
Clicking each chat allows you to view the chat content on the right.
Knowledge: The title of the referenced knowledge.
Only knowledge referenced at least once is listed. Click the [Go to Knowledge] button to go directly to that knowledge.
Reference Count: Accumulated when the knowledge is referenced in an ALF chat.
Resolved Chat Ratio: The ratio of chats categorized as [ALF Resolved] among chats where the knowledge was referenced.
Only counts where the knowledge was referenced within the inquiry period
are aggregated.
Example: If the inquiry period is 25/12/1~25/12/10, If the chat with ALF started on December 10th, but the knowledge was referenced on December 11th, that count is not included.
Click [See More] to view up to 20 items.
Information Provided
Task: The title of the executed task. Only tasks executed at least once are listed. Click the [Go to Task] button to go directly to that task.
Execution Count: Accumulated when the task is executed in an ALF chat.
Resolved Chat Ratio: The ratio of chats categorized as [ALF Resolved] among chats where the task was executed.
Only counts where the task was executed within the inquiry period are aggregated.
Example: If the inquiry period is 25/12/1~25/12/10, If the chat with ALF started on December 10th, but the task was executed on December 11th, that count is not included.
Various filters provided in the detailed chat list Detailed chat list data download provided