Usage Analysis

Provides backend monitoring data to understand user information.

How to view

Go to the left sidebar "AI" in "Usage Analysis", you can see the dashboard of how the AI handles conversations.

• Words in conversations

• Number of conversations

• Number of messages

• Average interactions per conversation

• User satisfaction

• Time filter (by year, month, day, hour)

Through these metrics, you can understand the AI assistant's usage from multiple perspectives.

Detailed explanation of statistics

1. Words Count (Words Count)

Calculation method:

• Text character count: Counts the number of characters in message content

• Image cost: Each image is calculated according to the configured image word-cost

• Total words = Message content character count + (Number of images × Image word-cost)

Statistical scope: Includes user input messages and AI assistant reply messages, supports daily and monthly time-granularity statistics

---

2. Conversations Count (Conversations Count)

Calculation method:

• Each time a new conversation is created, the counter +1

• When the first message in a conversation is created, it triggers an update to the conversation statistics

Update timing: When a user starts a new conversation (first message)

Statistical scope: Supports hourly, daily, monthly time-granularity statistics

---

3. Messages Count (Messages Count)

Calculation method:

• Each time a new message is created, the counter +1

• Includes user input messages and AI assistant reply messages

Update timing: Statistics are updated whenever each message is created

Statistical scope: Supports hourly, daily, monthly time-granularity statistics

---

4. Average Messages per Conversation (Average Messages per Conversation)

Calculation formula:

Average messages per conversation = Messages count ÷ Conversations count (displays 0 if Conversations count is 0)

Meaning explanation:

• Indicates how many messages each conversation contains on average

• Reflects the depth of interaction between users and the AI assistant

• A higher value indicates deeper conversations and more frequent interactions

Statistical scope: Supports hourly, daily, monthly time-granularity statistics

---

5. User Satisfaction Rate (User Satisfaction Rate)

Calculation formula:

User satisfaction = Likes ÷ Total feedback (displays 0 if Total feedback is 0)
Total feedback = Likes + Dislikes

Meaning explanation:

• Indicates the proportion of positive feedback (like rate)

• Value range: 0% ~ 100%

• A higher value indicates users are more satisfied with the AI assistant's replies

Feedback mechanism:

• Users can give the AI assistant's reply a “like” or a “dislike”

• Supports real-time statistic updates

Update timing: When users provide feedback (create, update, delete)

Statistical scope: Supports hourly, daily, monthly time-granularity statistics

---

Statistics update mechanism

Real-time update: When related events occur (new message, new conversation, feedback change), statistics updates are triggered immediately

Supported time granularity:

• Hourly: Suitable for real-time monitoring

• Daily: Suitable for daily report analysis

• Monthly: Suitable for monthly reports and trend analysis

Application scenario: Usage effectiveness evaluation

Observing trends in “User Satisfaction” and interaction counts can evaluate whether the AI assistant's response quality meets expectations and serve as a basis for optimization.