> ## Documentation Index > Fetch the complete documentation index at: https://docs.blubash.io/llms.txt > Use this file to discover all available pages before exploring further. # 360° View and AI quality analysis > Open a conversation in the 360° View and generate an AI analysis with quality, sentiment, outcome, compliance risks and coaching suggestions. The **360° View** brings everything about a conversation together on one screen: messages from the customer, the AI agent and the team, transfers and conversation metrics. From there, managers generate an **AI Analysis** that evaluates the service, human or automated, and suggests improvements. 360° View of a hybrid conversation with the AI analysis: quality 85/100, positive sentiment and topics ## When to use * To review conversations handled only by the AI agent and confirm it answers well. * To understand why a conversation was transferred or ended with an unhappy customer. * To prepare feedback sessions with team members, using concrete examples. * To spot risks, such as requests for personal data without a record of consent. ## Open the 360° View In **Reports & Insights**, click **View report** on the **Conversations** card. In **View mode**, choose **Detailed** and use the filters to find the conversation. The 360° View opens with the full conversation. Use **Back to list** to return to the report. The 360° View is available to **Admins** and to **Managers**, for conversations of the teams they belong to. ## 360° View tabs | Tab | Content | | - | - | | **Messages** | The full conversation, with messages from the customer, the AI agent and the team, transfers and internal notes | | **AI Analysis** | The **Conversation metrics** and the AI-generated analysis | ### Conversation metrics | Metric | What it shows | | - | - | | **Messages** | Total messages exchanged | | **Transfers (total)** | How many times the conversation changed hands | | **Transfer to AI** | Transfers from the team to an AI agent | | **Taken back by agent** | Times a team member took over a conversation the AI was handling | | **Duration** | Time from opening to closing, or until now | | **Satisfaction survey** | The Satisfaction Survey score, when answered | ## Generate the AI analysis On the **AI Analysis** tab, click **Generate AI analysis**. The analysis takes a few moments (**Analyzing…**). To redo it after new messages, use **Generate new analysis**. | Item | What it shows | | - | - | | **Quality** | A score from 0 to 100, rated **Excellent**, **Good**, **Average** or **Poor** | | **Sentiment** | **Positive**, **Negative**, **Neutral** or **Mixed**, and the **Sentiment evolution** from start to end | | **Outcome** | How the conversation ended, such as **Resolved**, **Unresolved**, **Escalated**, **Abandoned** or **Successful engagement**, with the **Outcome confidence** | | **Summary** | What happened in the conversation, in a few lines | | **Topics** | Subjects covered, such as pricing, support or sales | | **Identified intentions** | The customer's **Main intent** and **Sub-intents** | | **Reason for transfer** | Why the conversation moved from the AI agent to the team, when there was a transfer | | **AI evaluation** | How the service went: strengths and **Concerns** | | **Risks & compliance** | Whether the conversation **Has risks** and which points need attention | | **Coaching suggestions** | What the team and the AI agent can improve | AI evaluation with strengths, concerns and coaching suggestions ## Example: hybrid conversation In the conversation above, the AI agent Lia qualified an accounting lead and transferred it to the Sales team. The analysis gave a quality of 85/100 and positive sentiment, and pointed out two improvements: confirm the phone number or email before the transfer, and give a concrete callback time. Both became adjustments to the agent's training and instructions, and the next conversations were better right away. ## From diagnosis to action | What the analysis showed | Recommended action | | - | - | | The AI could not answer a topic | Add the content to the AI agent's training | | The AI transferred too early | Review the transfer instructions and the **Transfer to Team** skill | | The team member was slow or gave no timeline | Use the case in a feedback session and create a quick reply | | Compliance risk | Review the data collection script with the team and the AI agent | [Train the AI agent →](/ai-agents/train-ai-agent) · [Configure skills →](/ai-agents/configure-skills) The risks and compliance analysis is an alert for the manager to review, not a legal opinion. In conversations still in progress, the analysis only covers what has happened so far. For a complete evaluation, generate the analysis after the conversation is closed. ## Best practices * Analyze a weekly sample of conversations handled only by the AI, not just transferred ones or those with low scores. * Compare the AI evaluation with the customer's Satisfaction Survey score. Differences often reveal process problems. * Turn recurring suggestions into AI agent training, quick replies or team scripts. ## Next steps Find the conversations to analyze. Apply the improvements the analysis points out. This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.