See the whole experience
Bring conversations, tool activity, generative UI, and application state together so product teams can understand what people were trying to do.
See how people use your agents, where the experience gets in their way, and which workflows your product should support better.
Start with the patterns. Follow them back to the conversation.
The agent returns a separate answer for each plan. The user needs a comparison they can work with.
Conversation is a new source of product evidence. The requests people repeat and the work they take over show where the experience can improve.
“Can you compare these plans side by side?”Illustrative scenario
The agent gives separate answers, but the user keeps asking for a comparison. Follow the interaction to see the gap between the response and what the person actually needs.
Give the comparison a place in the product: an interactive table, useful criteria, and a clear next action. Use the evidence to decide what to build.
Bring conversations, tool activity, generative UI, and application state together so product teams can understand what people were trying to do.
Look at repeated requests, corrections, and moments where users step in. Trace an insight back to the interactions that support it.
Improve the interface, workflow, or available tools. When the opportunity is better agent behavior, turn the lesson into a Skill.
Rich Threads preserve messages, tool activity, generative UI, and application state. Intelligence helps you connect the patterns to the experiences behind them, so the team can decide what deserves attention.
Explore the documentationWhere do users get stuck? What do they ask for that your product cannot help them do? Review the interactions, choose a change, and look for improvement in the next round of use.
Explore Product AnalyticsCopy the setup prompt into your coding agent to start building with CopilotKit.