Insight search

Insight search lets you query your entire agent dataset using natural language. Ask about errors, costs, user behavior, performance, trends - anything. No SQL, no filters, no dashboards to navigate. Just ask and get answers backed by real data, with links to the specific runs behind every finding. You can be as broad or as specific as you want. Available through:

• The dashboard search bar
• The Slack bot (@The Context Company)
• MCP and API integrations

​

Example queries

​

General

Frustration analysis

Find the 3 biggest reasons users were frustrated with the agent in the past week and tell me potential reasons for why they’re happening.

Cost trends

How have my agent costs changed over the past month? Which models and tools are driving the most spend?

Weekly health check

Give me a summary of how my agent performed this week compared to last week. What got better, what got worse?

Topic breakdown

What are users asking about most? Which topics have the highest failure rate and which have the most positive feedback?

User satisfaction

Are users generally happy with the agent? Show me the ratio of positive to negative feedback and what the negative feedback is about.

​

Specific

Tool ordering issues

Find runs where the search tool was called after the summarize tool for users on the enterprise plan where the final response didn’t answer the user’s question.

Multi-turn failures

Show me sessions with more than 5 back-and-forth messages where the user eventually gave up. What was the agent getting stuck on?

Cost outliers by metadata

Find the most expensive runs from the past two weeks where the customer_tier metadata is set to “free”. What models were used and how many tool calls did they make?

Tool error chains

Find runs where a tool call failed immediately after another tool succeeded. Which tool pairs fail together most often?

Regression detection

Compare error rates for the generate_report tool this week vs last week. Did a specific model start failing more?

Slow paths

Find runs that took over 30 seconds where the user had to wait for more than 3 sequential tool calls. What tools were in the chain?

​

Browsing results

Every answer includes a table of relevant runs or sessions. Click any row to open the full trace - the complete prompt, the agent’s response, all tool calls and their results, token usage, costs, and any errors.

​

Chat history

Your searches are saved locally so you can return to previous investigations. Start a new chat anytime or continue where you left off.