> ## Documentation Index > Fetch the complete documentation index at: https://docs.sketricgen.ai/llms.txt > Use this file to discover all available pages before exploring further. # Conversations & Insights > Monitor AI agent performance with conversation history, trace analysis, credit usage dashboards, and platform distribution analytics in SketricGen. ## Overview The **Conversations** and **Insights** tabs provide a complete monitoring layer for your agents. They help you **review, analyze, and optimize** performance by showing real user interactions, credit usage, and traffic distribution across platforms. These two features together close the feedback loop, letting you trace how your agents perform in real-world deployments and how your overall account resources are utilized. ## Conversations ### Viewing Conversations The **Conversations** tab lets you review all interactions across your deployed agents. To view them: 1. Open **View Conversations** from the sidebar. 2. Select the **AgentSpace** or specific workflow you want to review. 3. A list of all recorded conversations for that AgentSpace appears. 4. Click on any conversation to open its full chat history. Each conversation displays both user messages and agent responses exactly as they occurred, giving you a clear understanding of how your workflow behaves in production. Screenshot showing the conversations list interface with search, filters, and conversation history ### Viewing Traces for Conversations Within each conversation, you can open **detailed execution traces** for any message. * Click the **View Traces** button next to an agent’s response. * A detailed trace panel opens showing: * The **agents involved** in generating that response. * Any **handoffs** or **tool calls** executed. * The **time taken** and **credits consumed** for that message. * Structured input/output data for debugging. This feature provides the same deep transparency available in the Playground — now extended to real, live conversations. Animated GIF showing conversation details interface with message history, agent interactions, and trace information ### Exporting Conversations You can also **export conversations** for documentation, analysis, or training purposes. Click the **Export** button at the top right of any conversation to download the full dialogue in JSON format. **Use cases include:** * Reviewing customer queries for workflow improvements. * Sharing examples with team members. * Fine-tuning your agents’ prompts or tool configurations. ## Insights ### Overview The **Insights** tab helps you track and analyze how your account and agents are performing. It includes two main views: **Account Overview** and **Agent Analytics**. Screenshot showing insights overview dashboard with conversation metrics, performance statistics, and usage analytics ### Account Overview This section summarizes key metrics for your account: | Metric | Description | | ----------------------- | ---------------------------------------------------------------------- | | **Credit Usage** | Shows how many of your allocated credits have been consumed. | | **Agent Capacity** | Displays how many agent slots are currently being used. | | **Active Agents** | Indicates how many of your agents are live and responding. | | **Estimated Days Left** | Calculates how long your credits will last at your current usage rate. | You’ll also see a **Credits Overview** panel showing: * Total credits used and remaining. * Your monthly credit allocation. * Visual credit utilization charts. This gives administrators and teams a clear picture of account health and usage patterns. You can also view the credits consumption by hovering on your Profile Screenshot showing credits usage interface with consumption tracking, billing information, and usage limits ### Agent Analytics The **Agent Analytics** view provides performance insights for individual agents or AgentSpaces. You can select any workflow to view its usage metrics. Each analytics dashboard includes: * **Total Messages** — Number of messages processed. * **Conversations** — Total unique conversations handled. * **Credits Used** — Credits consumed by that agent or workflow. * **Last Active** — Timestamp of most recent activity. * **Activity Timeline** — Message and credit usage trends across the past few days or weeks. Animated GIF showing analytics dashboard with conversation trends, performance metrics, and usage statistics ### Platform Distribution When agents are deployed across multiple channels, you can track where your users engage most. * **Usage by Platform:** Breakdown of traffic across connected apps and integrations (e.g., WhatsApp, Shopify, Slack). * **Platform Performance:** Credit and message consumption per platform. This helps you identify which deployment channels bring the highest engagement or cost and where optimization efforts should focus. ## Conversation Memory and Statefulness Within the same conversation thread, agents retain full context of the interaction. This means follow-up messages in the same conversation can reference previous turns without repeating information. Key behaviors to be aware of: * **Within a conversation:** Stateful. Agents remember the full thread context. * **Across conversations:** No persistent memory. Each new conversation starts fresh with only the workflow instructions — no previous chat context carries over. * **Long conversations:** When context grows long, the system automatically summarizes and optimizes it to stay within model limits. ## How Conversations and Insights Work Together * **Conversations** let you drill down into the *quality* and *content* of agent interactions. * **Insights** provide a *quantitative overview* of performance and resource utilization. Combined, they give you full operational visibility — from individual message traces to system-wide analytics. ## Best Practices * Review **Conversations** weekly to identify recurring user queries or misunderstood intents. * Monitor **Insights** regularly to track credit efficiency and agent utilization. * Use **Agent Analytics** to identify high-performing workflows or underused tools. * Compare **Platform Distribution** metrics to refine your deployment strategy. * Export key conversations for retraining or documentation when refining prompts.