This guide provides natural language examples for querying your FOCUS cost data through Claude Desktop, Gemini CLI, or other MCP-compatible agents.
Sign in to like and favorite skills
This guide provides natural language examples for querying your FOCUS cost data through Claude Desktop, Gemini CLI, or other MCP-compatible agents.
Once MCP Toolbox is configured, you can ask questions in plain English. The AI assistant will:
| What You Ask | Tool Used |
|---|---|
| "What did we spend..." | |
| "Show me costs by service..." | |
| "Which regions..." | |
| "What are the daily costs..." | |
| "Find cost anomalies..." | |
| "Forecast next month..." | |
| "When will we hit budget..." | |
| "Show me untagged resources..." | |
Simple Queries:
What did we spend last month?
Show me our total cloud costs for the past 90 days
What's our spending trend over the last 6 months?
Advanced Queries:
Break down our spending by month for the last year and show me the savings from discounts
Compare our spending this month versus last month
Expected Response:
Based on your FOCUS data, here's your spending for the last 30 days:
- Total Billed Cost: $15,234.56
- Total Effective Cost: $12,987.45
- Total Savings: $2,247.11 (14.7% discount rate)
- Number of Billing Accounts: 1
Simple Queries:
What are our top 5 most expensive services?
Show me costs by service for last month
Which services cost the most?
Advanced Queries:
Show me all services costing more than $100 last week
What percentage of our budget goes to Compute Engine versus BigQuery?
List our services ordered by cost with the number of charges for each
Expected Response:
Here are your top 5 services by cost (last 30 days):
- Compute Engine: $4,523.12 (35.8% of total)
- Cloud Storage: $2,145.67 (17.0% of total)
- BigQuery: $1,876.23 (14.8% of total)
- Cloud SQL: $1,234.45 (9.8% of total)
- Cloud Load Balancing: $876.34 (6.9% of total)
Simple Queries:
Which regions are we spending the most in?
Show me costs by region for the past month
Where are we deploying resources geographically?
Advanced Queries:
Compare costs between us-central1 and us-east1
Show me which regions have the highest average cost per charge
List all regions we're using with their associated services
Expected Response:
Your spending by region (last 30 days):
Region ID Region Name Services Total Cost us-central1 Iowa 12 $5,432.10 us-east1 South Carolina 8 $3,210.45 europe-west1 Belgium 6 $2,145.32
Simple Queries:
What are my daily costs for the past week?
Show me the spending trend over the last 30 days
What was our average daily spend last month?
Advanced Queries:
Show me daily costs with a 7-day moving average for the past 60 days
Are our costs trending up or down this month?
Which days had unusually high spending?
Expected Response:
Daily costs for the past 7 days:
Date Daily Cost 7-Day Avg Services 2024-10-24 $425.67 $398.23 15 2024-10-23 $412.34 $395.12 14 2024-10-22 $389.45 $392.45 16
Simple Queries:
Find any unusual cost spikes
Show me cost anomalies from the past week
Which services had unexpected costs recently?
Advanced Queries:
Analyze the last 30 days and find services with costs more than 2 standard deviations above normal
Show me moderate and high severity cost anomalies from yesterday
What caused the cost spike on October 23rd?
Expected Response:
Found 3 cost anomalies in the past 7 days:
High Anomaly:
- Service: Compute Engine
- Date: 2024-10-23
- Cost: $1,234.56 (avg: $425.67)
- Z-Score: 2.8
Moderate Anomalies:
- BigQuery on 2024-10-22: $876.23 (z-score: 1.7)
- Cloud Storage on 2024-10-21: $654.32 (z-score: 1.6)
Simple Queries:
What will we spend next month?
Forecast our costs for the next 30 days
Predict next month's cloud bill
Advanced Queries:
Forecast next month using the last 90 days of data
What's the expected cost and growth rate for next month?
Show me monthly forecasts for the next 3 months
Expected Response:
Forecast for November 2024:
- Expected Cost: $14,567.89
- Based on: 90 days of historical data
- Daily Growth Rate: 0.23%
- Confidence: Historical average daily cost is $456.78
Simple Queries:
Forecast costs by service for next month
Which services will cost the most next month?
What will Compute Engine cost next month?
Advanced Queries:
Show me next month's forecast for each service with confidence intervals
Compare forecasted costs to current month actuals
Which services are projected to grow the fastest?
Expected Response:
Service-level forecasts for next month:
Service Forecast Low Estimate High Estimate Compute Engine $4,678.23 $4,234.12 $5,122.34 Cloud Storage $2,234.56 $1,987.34 $2,481.78 BigQuery $1,965.43 $1,678.12 $2,252.74
Simple Queries:
When will we hit our budget limit?
Are we on track to stay within budget this month?
How many days until we exceed our $10,000 budget?
Advanced Queries:
Given a monthly budget of $15,000, what's our current burn rate and projected variance?
Calculate our budget runway and tell me if we'll go over
Show me month-to-date spend and forecasted month-end total
Expected Response:
Budget Runway Analysis:
- Monthly Budget: $10,000.00
- Month-to-Date Spend: $6,234.56
- Avg Daily Burn Rate: $312.50
- Days Remaining: 12
- Forecasted Remaining: $3,750.00
- Month-End Total: $9,984.56 ✅
- Projected Variance: $15.44 (under budget)
- Status: On track
Over Budget Example:
⚠️ Budget Alert:
- Monthly Budget: $10,000.00
- Month-End Forecast: $12,345.67
- Days Until Budget Exhausted: 4 days
- Action: Reduce spend by $195/day to stay within budget
Simple Queries:
How much are we saving from commitments?
Show me our CUD savings for the past 3 months
What's our commitment discount utilization?
Advanced Queries:
Break down commitment savings by type (spend vs usage)
Calculate our average savings percentage from commitments
Show me which commitments are providing the most value
Expected Response:
Commitment Savings (Last 3 Months):
Month Type Commitments Savings Savings % Oct 2024 Usage 2 $1,234.56 23.4% Oct 2024 Spend 1 $876.23 15.6% Sep 2024 Usage 2 $1,189.45 22.8% Total Savings: $3,300.24 over 3 months
Simple Queries:
Show me untagged resources
Which resources don't have proper cost allocation tags?
Find resources missing team or environment tags
Advanced Queries:
Show me the top 20 most expensive untagged resources from the past week
Calculate how much we're spending on untagged resources
List all Compute Engine instances without tags
Expected Response:
Top Untagged Resources (Last 7 Days):
Resource Type Resource ID Service Region Cost instances instance-12345 Compute Engine us-central1 $456.78 disks disk-67890 Compute Engine us-east1 $234.56 datasets dataset-abc BigQuery us $189.23 Total Untagged Cost: $2,345.67 (18.6% of total spend)
Simple Queries:
Show me costs by team
How much did the engineering team spend?
Break down costs by environment (prod, dev, staging)
Advanced Queries:
Show me costs for each team, sorted by spend
Compare production vs non-production environment costs
Which cost-center has the highest cloud spend?
Expected Response:
Costs by Team (Last 30 Days):
Team Resources Cost Services platform-engineering 45 $4,567.89 8 data-analytics 32 $3,456.78 6 machine-learning 28 $2,987.65 5 web-applications 19 $1,876.54 7
Simple Queries:
Compare Compute Engine pricing across regions
Which region is cheapest for storage?
Show me cost per unit by region for BigQuery
Advanced Queries:
For Compute Engine, show me the cost per unit in each region where we have usage
Which regions offer the best value for our workloads?
Compare us-central1 vs europe-west1 for all services
Expected Response:
Compute Engine Pricing by Region (Last 30 Days):
Region Total Usage Unit Total Cost Cost/Unit us-central1 12,345 hour $2,345.67 $0.1900 us-east1 8,976 hour $1,795.20 $0.2000 europe-west1 5,432 hour $1,195.04 $0.2200 💡 Insight: us-central1 is 5% cheaper than us-east1
Simple Queries:
What are our most expensive resources?
Show me the top 10 resources by cost
Which VMs cost the most?
Advanced Queries:
List the top 20 most expensive resources with their daily average costs
Show me resources that have been running for more than 30 days
Which resources cost more than $100 last month?
Expected Response:
Top 10 Resources by Cost (Last 30 Days):
Type Resource ID Service Region Total Cost Avg Daily instances prod-web-01 Compute Engine us-central1 $876.54 $29.22 instances data-proc-cluster Compute Engine us-west1 $654.32 $21.81 datasets analytics-warehouse BigQuery us $543.21 $18.11
Simple Queries:
What credits did we receive?
Show me all discounts and credits for last month
How much did sustained use discounts save us?
Advanced Queries:
Break down all credit types with amounts for the past 3 months
Which services received the most credits?
Compare promotional credits vs usage discounts
Expected Response:
Credits Received (Last 3 Months):
Month Credit Type Count Total Amount Services Oct 2024 SUSTAINED_USE_DISCOUNT 245 -$1,234.56 5 Oct 2024 COMMITTED_USE_DISCOUNT 89 -$876.23 3 Oct 2024 PROMOTIONAL 12 -$234.56 4 Total Credits: -$2,345.35 (credits are negative = savings)
Simple Queries:
How are our costs growing month over month?
Show me cost growth trends
Which services are growing fastest?
Advanced Queries:
Calculate month-over-month growth for the past 6 months by service
Show me services with more than 10% monthly growth
Compare this month's growth to last month's growth
Expected Response:
Month-over-Month Growth (Top Services):
Service This Month Last Month Change % Change BigQuery $2,456.78 $1,987.65 +$469.13 +23.6% 📈 Cloud Storage $2,123.45 $2,045.32 +$78.13 +3.8% ↗️ Compute Engine $4,567.89 $4,678.23 -$110.34 -2.4% ↘️
These queries may require the AI to use multiple tools:
Cross-Analysis:
Show me our top 5 services, then forecast what they'll cost next month
Find cost anomalies and then show me the resources driving those costs
Compare our spending by region and tell me which region has the best pricing
Comparative Analysis:
Show me how much we're spending on untagged resources versus tagged resources
Compare our actual spend this month to what we forecasted last month
Calculate our commitment savings and show what percentage of our bill they represent
Trend Analysis:
Show me daily costs for the past 30 days and identify any anomalies
Calculate our monthly growth rate and project when we'll hit $20,000/month
Show me which teams are growing fastest and forecast their costs for next quarter
✅ Good: "Show me costs for the last 30 days" ❌ Vague: "Show me recent costs"
✅ Good: "What did Compute Engine cost last month?" ❌ Generic: "How much did we spend?"
✅ Good: "Compare this month to last month" ✅ Good: "Which region is cheaper?"
✅ Good: "Show me costs with percentages of total" ✅ Good: "Include confidence intervals in the forecast"
After getting results, you can ask:
You: "What did we spend last month?"
AI: Uses
with days_back=30get_total_spend
You spent $12,456.78 last month with $1,234.56 in savings from discounts.
You: "Which services drove that cost?"
AI: Uses
with days_back=30get_cost_by_service
Your top services were: Compute Engine ($4,567), Cloud Storage ($2,345), and BigQuery ($1,876).
You: "Are those costs normal or did we have any spikes?"
AI: Uses find_cost_anomalies
I found one anomaly: BigQuery had an unusual spike on Oct 23rd ($876 vs normal $425).
You: "Forecast what we'll spend next month"
AI: Uses forecast_next_month
Based on trends, you'll likely spend $13,245 next month (6.3% increase).
You: "Will that exceed our $15,000 budget?"
AI: Uses
with monthly_budget=15000check_budget_runway
No, you're projected to stay under budget by $1,755. You have 28 days of runway remaining.
To create your own effective prompts:
Template:
Show me [METRIC] for [TIMEFRAME] [FILTERS] with [CONTEXT]
Examples:
Happy querying! 🚀