Claymore
Get in touch
Claymore®

AI Query Examples

Ask questions in plain English. Claymore's AI Analyst generates ClickHouse SQL queries and returns results with linked session replays.

<100ms
Query Execution
Natural
Language Input
SQL
Transparency
Linked
Replays

How It Works

1

Ask in Plain English

Type your question naturally, like you would ask a colleague.

2

AI Generates SQL

The AI understands your schema and generates optimized ClickHouse SQL.

3

Results with Context

Get answers instantly, with linked session replays for each result.

Revenue Impact

“Show me sessions where checkout errors cost us over $100”

Find high-value abandoned carts caused by technical issues

Filters by errorCode IS NOT NULL AND cartValue > 100
“Which errors caused the most lost revenue this week?”

Aggregate revenue impact by error type

Groups by errorType, sums cartValue, orders by total
“Find sessions from Germany with cart value over €500 that failed”

Geo-targeted high-value failure analysis

Filters by country, cartValue, and session outcome

User Behavior

“Show me sessions where users abandoned checkout”

Find users who reached checkout but never completed purchase

Filters by pageViews containing /checkout without purchase event
“Which pages have the highest rage-click rates?”

Identify frustrating UI elements causing user distress

Aggregates click events with rapid succession timestamps
“Find sessions where users scrolled but never clicked”

Detect content that fails to engage users

Filters by scrollDepth > threshold AND clickCount = 0

Error Analysis

“Show me all 500 errors from the payment API this week”

Track payment gateway reliability issues

Filters by errorCode = 500 AND url LIKE payment%
“Which browsers have the most JavaScript errors?”

Find browser compatibility issues in your frontend

Groups by browserFamily, counts JS errors
“Find sessions with network timeouts over 5 seconds”

Identify slow API responses affecting user experience

Filters by networkDuration > 5000

Geographic & Device

“Compare conversion rates between mobile and desktop in the US”

Segment performance by device type and region

Groups by deviceType, country, calculates conversion rate
“Show me sessions from Berlin where users got errors during checkout”

Drill down to city-level issue analysis

Filters by city = Berlin AND page LIKE checkout%
“Which countries have the slowest page load times?”

Find regions with performance issues

Groups by country, averages loadTime

Funnel Analysis

“Where do users drop off between cart and payment?”

Identify friction points in the purchase funnel

Counts sessions by funnel stage, calculates drop-off rates
“Show me sessions that visited pricing but never signed up”

Find users who showed intent but did not convert

Filters by pageViews containing /pricing without signup event
“What percentage of users who start signup complete it?”

Measure signup form completion rate

Calculates ratio of signup_complete to signup_start events

SQL Transparency

Every AI query shows you the generated SQL. No black boxes. You can copy, modify, and run the SQL directly in our SQL Explorer.

Example: “Show me checkout errors costing over $100”

SELECT
  s.sessionId,
  s.userId,
  s.country,
  s.city,
  e.errorMessage,
  e.errorCode,
  s.cartValue,
  s.timestamp
FROM sessions s
JOIN events e ON s.sessionId = e.sessionId
WHERE e.errorCode IS NOT NULL
  AND e.page LIKE '%checkout%'
  AND s.cartValue > 100
ORDER BY s.cartValue DESC
LIMIT 50
47ms23 sessions found with linked replays
Why ClickHouse? The AI Analyst works because ClickHouse can query raw, non-aggregated data in milliseconds. No pre-computation needed. Ask any ad-hoc question and get instant answers.Learn more about our architecture →

Ask, Don't Watch

Traditional tools: Watch 100 session replays to find the one with the bug.

Claymore: Ask “Show me sessions with payment errors over $100” and get exactly the sessions you need, with replays ready to play.

Turn hours of watching into seconds of asking.