I need to systematically check for data quality issues across all tables. Let me start by examining the sales_pipeline table more closely, then move to other tables. {"action": "tool_call", "reasoning": "Check for outliers and impossible values in sales_pipeline. I already know close_value ranges 0..30288 with 2089 nulls, and dates span 2016-2017. Let me look for: negative values (impossible), zero close_values on Won deals (suspicious), date logic issues (engage_date after close_date), and extreme outliers in close_value.", "tool": "query_sql", "purpose": "Identify data quality issues in sales_pipeline: impossible values, date logic violations, and suspicious patterns", "args": {"sql": "SELECT COUNT(*) as total_rows, COUNT(CASE WHEN close_value < 0 THEN 1 END) as negative_values, COUNT(CASE WHEN close_value = 0 AND deal_stage = 'Won' THEN 1 END) as won_with_zero_value, COUNT(CASE WHEN engage_date > close_date THEN 1 END) as engage_after_close, COUNT(CASE WHEN close_date IS NOT NULL AND engage_date IS NULL THEN 1 END) as close_without_engage, COUNT(CASE WHEN deal_stage = 'Won' AND close_value IS NULL THEN 1 END) as won_without_value, COUNT(CASE WHEN deal_stage = 'Won' AND close_date IS NULL THEN 1 END) as won_without_close_date FROM raw.sales_pipeline"}}