Analytics
The Analytics page helps you track your workspace performance easily. It provides four main reports, Resolution Report, Bot Performance, Conversation Traffic, and Complaints Report.
Resolution Report
This report shows a summary of interactions that handled either by the human agent or AI agent that you have previously configured.
Here's the breakdown of interaction mechanisms:
| Metrics | Description |
|---|---|
| Conversations → Resolved by Customer | Total number of conversations completed directly by customers without the help of human agents or AI agents until the end. |
| Conversations → Agent | Total number of conversations handled by human agents. |
| Conversations → Bot | Total number of conversations handled by AI agents. |
| Conversations → Resolved by System | Total number of conversations that were closed automatically by the system without direct intervention from customers, human agents, or AI agents. |
| Agent → Resolved by Customer | Total number of conversations handled by human agents and the conversations is ended by the customers. |
| Agent → Resolved by Agent | Total number of conversations handled by human agents and the conversations is ended by human agents. |
| Agent → Resolved by System | Total number of conversations handled by human agents and the conversations is ended by the system. |
| Bot → Resolved by Customer | Total number of conversations handled by AI agents and the conversations is ended by the customers. |
| Bot → Agent Takeover | Total number of conversations handled by AI agents and human agents taking over conversations. |
| Bot → Bot Handover | Total number of conversations that AI agents handed over to the human agents. |
| Bot → Resolved by System | Total number of conversations handled by AI agents and the conversations is ended by the system. |
| Agent Takeover → Resolved by Customer | Total number of conversations that human agents took over from AI agents and the conversations is ended by the customers. |
| Agent Takeover → Resolved by Agent | Total number of conversations that human agents took over from AI agents and the conversations is ended by the human agents. |
| Agent Takeover → Resolved by System | Total number of conversations that human agents took over from AI agents and the conversations is ended by the system. |
| Bot Handover → Agent Takeover | Total number of conversations where AI agents handed over the conversation and a human agent took over the interaction. |
| Bot Handover → Resolved by Customer | Total number of conversations handed over from AI agents to human agents where no human agent has taken over, and the conversation was ended by the customer. |
| Bot Handover → Resolved by System | Total number of conversations handed over from AI agents to human agents where no human agent has taken over, and the conversation was ended by the system. |
Summary
The summary gives you a quick overview of performance within your current workspace and selected time range.
| Metrics | Description |
|---|---|
| Source → Target | Represents the origin and destination of the event being measured. It shows where the data started (source) and where it ended up (target). |
| Average | The mean value of the measured metric across all records within the selected time range. |
| Count | The total number of records included in the calculation. This indicates the sample size used to compute the other metrics. |
| p50 | The value at which 50% of the data points are below it. |
| p75 | The value below which 75% of the data points fall. This shows the upper middle performance range and helps identify a significant portion of customers are experiencing slower performance than the median. |
| p90 | The value below which 90% of the data points fall. This metrics is commonly used to evaluate performance reliability. |
| p99 | The value below which 99% of the data points fall. This highlights extreme cases or worst case performance scenarios. |
You can copy and paste the summary table into Excel without changing the data, or download it as a .csv file for further analysis.
If you scroll down, there's a detailed table of all incoming conversations in your workspace. From this table, you can quickly see a conversation by clicking on one in the list.
The Today report filter is not functioning properly at the moment, so it is strongly recommended to use the Last 7 days filter instead for more accurate results.
Bot Performance
The Bot Performance tab displays AI agent's performance based on user interactions. This report helps you evaluate both auto-pilot and co-pilot modes.
- In auto-pilot mode, you can assess the quality of the bot's responses and determine whether they align with your expectations or the bot's defined persona.
- In co-pilot mode, you can review how often you used the suggested answer directly, modified it before sending, or chose to ignore it.
At the top of the page, you will find the Select a bot option, where you can choose which AI agent's data you want to review. You will also see a brief summary of your selected AI agent, including its latest active mode and the labels applied to the AI agent.
You can view the bot's performance percentage as well as the detailed breakdown of its messages from dashboard and message details.
Use this report to identify topics that your bot may struggle with. If there are many unresolved conversations, consider to add new articles to the Knowledge Base to help improve the AI agent's accuracy and performance.
Conversation Traffic
The Conversation Traffic report provides a detailed overview of your workspace engagement levels over your selected timeframe. This data helps you identify peak activity periods by tracking daily, weekly, or monthly conversation volumes. Additionally, you can monitor the number of unique customers interacting with both AI and human agents to better manage your team's workload.
Agent Performance
The Agent performance report provides a detailed analysis of agent responsiveness, resolution efficiency, and workload within the selected workspace and reporting period.
All metrics are calculated based on conversations that match the applied filters, such as workspace and time range. The report is displayed at the agent and web widget level, meaning that if an agent handles conversations from multiple web widget, each combination will appear as a separate row, even if the agent name is the same.
This report helps stakeholders evaluate service quality, monitor operational efficiency, and identify opportunities for performance improvement.
Metrics
| Metrics | Description |
|---|---|
| Average first response time | The average duration between the start of a conversation and the first message sent by an agent. This metric is used to measures agent responsiveness and reflects how quickly incoming conversations are acknowledged. |
| Average resolving time | The average duration from the start of a conversation until it's marked as resolved. This metric is used to evaluates how efficiently agents handle and complete conversations. |
| Total conversations | The total number of conversations handled by agents within the selected reporting period. This metric is used to represents overall workload and activity volume during the reporting timeframe. |
Filter
You can use filters to narrow down the agent performance. Filter options include date range and channels.
Download
You can also download this report in .csv format by clicking the download csv button on the right.
This is an example of a csv file.
Complaints Report
This report contains a list of complaints or negative feedback from customers. Customers can submit report about human agent, AI agent, and automation rules. You can review each report individually to find out:
- Frequently complained about by customers
- Bot error patterns
- Priority areas for improvement
