| name | ai-cost-dashboard |
| description | Build and manage a Snowflake AI Cost Dashboard using stored procedures and Streamlit. Use for **ALL** requests involving: AI cost, AI spending, cost dashboard, spending dashboard, Snowflake AI credits, Cortex cost tracking. DO NOT attempt AI cost dashboard work manually - invoke this skill first. |
AI Cost Dashboard
Deploys and manages a Cortex AI Cost Dashboard tracking credit usage across Snowflake AI services via stored procedures and a Streamlit app.
Source Files
- Setup SQL:
cortex_ai_cost_dashboard/setup.sql (template — uses {{WAREHOUSE}} placeholder)
- Streamlit App:
cortex_ai_cost_dashboard/streamlit_app.py (dynamic — derives DB/schema from session context)
- Semantic Model:
cortex_ai_cost_dashboard/cortex_ai_costs_model.yaml (template — uses {{DATABASE}}/{{SCHEMA}} placeholders)
- Snowflake CLI config:
cortex_ai_cost_dashboard/snowflake.yml (template — uses {{DATABASE}}/{{SCHEMA}}/{{WAREHOUSE}} placeholders)
- Deploy Skill:
cortex_ai_cost_dashboard/skills/ai-cost-dashboard-deploy/SKILL.md
Architecture
SNOWFLAKE.ACCOUNT_USAGE (usage history views)
-> MERGE (incremental)
<DATABASE>.<SCHEMA> (detail tables + unified + user-level)
-> Queried by
Streamlit App (CORTEX_AI_COST_DASHBOARD)
All object names (database, schema, warehouse, role, credit rate) are configurable at deploy time.
Key objects (created in user-specified database.schema):
- Procedures:
REFRESH_CORTEX_AI_COSTS(DAYS_BACK), REFRESH_USER_AI_COSTS(DAYS_BACK)
- Tables:
CORTEX_AI_UNIFIED_COSTS, USER_AI_COSTS, plus detail tables
- Task:
REFRESH_CORTEX_AI_COSTS_DAILY (CRON 0 6 * * * UTC)
Configurable Parameters
| Parameter | Description | Default |
|---|
| Connection | Snowflake CLI connection name | — |
| Database | Target database | ADMIN_DB |
| Schema | Target schema | AI_COSTS |
| Warehouse | Warehouse for queries and tasks | COMPUTE_WH |
| Role | Deployment role | ACCOUNTADMIN |
| Credit Rate | Current standard $/credit rate | 2.88 |
| Historical Rate | Previous $/credit rate (before current) | 3.01 |
Intent Detection
| Intent | Triggers | Load |
|---|
| SETUP | "set up", "deploy", "create", "install", "build dashboard" | setup-deploy/SKILL.md |
| ANALYZE | "analyze", "report", "show costs", "insights", "spending breakdown" | analyze-report/SKILL.md |
Workflow
Step 1: Identify Intent
Ask the user what they want to do:
What would you like to do with the AI Cost Dashboard?
1. Setup / Deploy - Deploy stored procedures and Streamlit app from scratch
2. Analyze / Report - Analyze current AI cost data and generate insights
Route based on selection:
- Option 1 -> Load
setup-deploy/SKILL.md
- Option 2 -> Load
analyze-report/SKILL.md
If intent is clear from context, skip the question and route directly.
Prerequisites
- Snowflake account with ACCOUNTADMIN or role with ACCOUNT_USAGE access
- Warehouse for query execution
Output
Deployed AI cost dashboard or cost analysis report depending on intent.
