Multi-cloud spend,
last month
"What did we spend last month across AWS, GCP, and Azure? Break it down by provider and show the month-over-month change."
What Billy returns
A provider-level table with current-month spend, prior-month spend, and absolute and percentage change. Includes a chart you can drop into a Finout dashboard with one follow-up.
- FinOps
- Finance
- Exec
- Multi-cloud
Month-to-date spend by environment
"Show me month-to-date spend grouped by environment (prod, staging, dev) across all my accounts."
What Billy returns
Spend by environment, sorted descending, with the share of total cost each environment represents. Untagged spend is called out separately.
- Engineering
- DevOps
- FinOps
- Multi-cloud
Top 5 cost-driving services this week
"Which 5 services drove the most cloud cost this week, and what changed from last week?"
What Billy returns
The top 5 services by spend, week-over-week delta for each, and a short explanation of the largest mover. Click-through to a MegaBill view.
- Finance
- FinOps
- Exec
- Multi-cloud
Spend mix: compute, storage, networking, AI
"What's my cloud spend mix between compute, storage, networking, data transfer, and AI services for the past 90 days?"
What Billy returns
A stacked bar by category with totals and percentages, comparing the latest 90 days to the prior 90. AI breaks out separately rather than being lumped into compute.
- FinOps
- Engineering
- Multi-cloud
Account-level spend ranking
"Rank my AWS accounts by spend for last month, and tell me which three grew the fastest."
What Billy returns
Account ranking with absolute spend, month-over-month growth, and the top three accounts by growth rate - so you catch fast-growing small accounts before they get big.
Spend by region with anomaly flags
"Break down my cloud spend by region for the past 30 days, and flag any region where spend deviates from the 90-day baseline by more than 20%."
What Billy returns
A regional spend breakdown with anomaly markers on outlier regions. Each flagged region links to the underlying anomaly detection record in Finout.
- Engineering
- FinOps
- Multi-cloud
Quarter-close reconciliation
"Reconcile our Q1 cloud spend by business unit, including credits, refunds, and marketplace charges. Surface anything that looks misallocated."
What Billy returns
A reconciled BU-level breakdown using Net Amortized cost excluding credits, with marketplace and refund lines called out separately. Misallocated resources surfaced.
- Finance
- FinOps
- Multi-cloud
Daily cost trend with driver analysis
"Show me my daily cloud cost for the past 14 days and tell me what's driving the trend."
What Billy returns
Daily cost line chart plus a narrative driver analysis - which services, accounts, or usage types account for the trend direction.
- FinOps
- Engineering
- Multi-cloud
Audit-ready quarterly reconciliation
"Generate a fully reconciled quarterly cloud cost statement, including credits, refunds, support charges, marketplace, and discounts, split by legal entity, with footnotes on any reclassifications."
What Billy returns
A finance-grade reconciliation document with auditable line items, entity-level totals, and footnotes for reclassifications. Exportable for inclusion in the close package.
- Engineering
- FinOps
- Multi-cloud
Explain my virtual tag structure
"Explain how my virtual tags are structured and show me how cost flows through them."
What Billy returns
A diagram of your vTag hierarchy, the rules behind each vTag, and an example of how a representative resource gets attributed.
"We used this instead of writing a 20-page internal vTag explainer doc." - Enterprise insurance customer
- FinOps
- Engineering
- Multi-cloud
Tag coverage and untagged spend
"What percentage of my cloud spend is properly tagged, and what's the dollar value of untagged resources by service and account?"
What Billy returns
Tag coverage percentage overall, untagged spend by service and account, and a ranked list of the largest untagged resources you should fix first.
- FinOps
- Engineering
- Finance
- Multi-cloud
Resource-level
tag lookup
"What tags does resource i-0abc123def456 currently have, and which virtual tags does it fall under?"
What Billy returns
All native cloud tags on the resource, every vTag that picks it up, and the cost center it ultimately maps to.
- Engineering
- DevOps
- Multi-cloud
Build a virtual tag for application owner
"Create a virtual tag that groups workloads by application owner, using the app_id tag as primary key and falling back to account ownership where app_id is missing."
What Billy returns
A draft vTag rule with the fallback logic, a preview of how many resources will be matched, and an estimate of how much currently untagged cost the new vTag will resolve.
Showback by business unit
"Generate a showback report by business unit for the past month, including both direct cost and allocated shared services."
What Billy returns
Per-BU cost with direct and allocated portions broken out, share of shared services each BU absorbs, and a downloadable view ready to send to BU owners.
- FinOps
- Finance
- Multi-cloud
Top anomalies this week with root cause
"Show me the top cost anomalies from the past 7 days and what's driving each one."
What Billy returns
Detected anomalies ranked by impact, with contributing service, resource, and a short root-cause narrative for each. Action suggestions where applicable.
- Engineering
- FinOps
- Multi-cloud
Investigate yesterday's production anomaly
"Investigate yesterday's anomaly on the production environment — what resource or usage type caused the spike?"
What Billy returns
The specific anomaly record, contributing resources ranked by cost impact, and (when correlated data is available) the deployment or configuration change that aligns with the spike.
- Engineering
- DevOps
- Multi-cloud
Sustained run-rate changes, period over period
"Compare last week's daily run rate to the week before. Are there accounts or services with sustained reductions or increases?"
What Billy returns
Side-by-side weekly run rates, accounts and services with sustained changes (not one-off spikes), and an estimate of annualized impact if the change holds.
- FinOps
- Finance
- Multi-cloud
Anomaly impact attributed to team
or owner
"For the anomalies detected in the past 30 days, which teams or owners are responsible for the most impact?"
What Billy returns
Anomaly impact aggregated by team / owner / vTag, ranked by total dollar impact. Surfaces which team is generating the most cost surprises.
End-of-month projection from current burn rate
"Based on month-to-date spend and the past 90 days of patterns, project where my cloud spend will land at end-of-month."
What Billy returns
End-of-month projection with a confidence range and a breakdown of which accounts and services are most responsible for the projection.
- FinOps
- Finance
- Multi-cloud
What-if: 80% reservation coverage
"Model what our compute cost would look like if we moved from current commitment coverage to 80% coverage, mixing 1-year and 3-year terms."
What Billy returns
Projected savings vs. on-demand, payback period for the upfront commitments, and a recommended mix of 1-year and 3-year terms based on workload stability.
- Finance
- FinOps
- Multi-cloud
Budget variance with forward projection
"Show me budget variance by team for the current quarter, and project end-of-quarter overrun risk based on current burn rates."
What Billy returns
Variance per team (actual vs. budget), end-of-quarter projection from current run rate, and risk-flagged teams likely to overspend.
- Finance
- FinOps
- Multi-cloud
Multi-year cloud spend forecast
"Forecast our 2027 and 2028 cloud spend across AWS, GCP, and Azure, excluding marketplace and one-time charges. Assume 1% monthly organic growth."
What Billy returns
Monthly forecast through end-2028 by provider and major service category, with a high/base/low sensitivity range so Finance can plan with confidence intervals.
AI workload forecast at 40% growth
"If our AI inference workload grows 40% next year, what will it cost on Bedrock, Vertex, and direct Anthropic / OpenAI usage? Where will the new cost concentrate?"
What Billy returns
Forecasted AI cost by vendor and model at the 40% growth assumption, the marginal cost per additional unit of workload, and committed-use discount recommendations to cushion growth.
- Engineering
- Finance
- Multi-cloud
Azure forecast with savings plan scenario
"Forecast my Azure spend over the next 12 months assuming we shift 80% of consumption to a 3-year Compute Savings Plan starting month 3."
What Billy returns
Monthly forecast with the savings plan layered in starting from the cut-over month, total expected savings over 12 months, and the break-even point where the SP pays for itself.
Top 5 quick wins to act on this week
"What are the top 5 quick-win cost reductions I could act on this week? I want low effort, high impact."
What Billy returns
5 ranked actions - typically idle resources, unattached volumes, expired commitments, oversized non-prod, or untouched snapshots - each with the savings number and the team that owns the resource.
- FinOps
- DevOps
- Multi-cloud
Idle RDS instances last 30 days
"Find me RDS instances that have had near-zero connections for the past 30 days, and tell me what they cost."
What Billy returns
Idle or near-idle RDS instances with cost, connection metrics, and the account/team they belong to. Each row links to the CostGuard scan finding.
Rightsize underutilized compute
"Identify EC2 and GCE instances sustained below 20% CPU and 30% memory over the past 14 days, and recommend rightsizing."
What Billy returns
Underutilized instances ranked by potential savings, the recommended target instance type, and the estimated monthly savings if rightsized.
- DevOps
- Engineering
- Multi-cloud
Top savings opportunities ranked by value
"Give me my top 20 cost-saving opportunities right now, ranked by monthly savings potential."
What Billy returns
A ranked list spanning rightsizing, idle resources, expired commitments, untouched snapshots, and orphaned volumes — sized in monthly savings and tagged with the resource owner.
Non-prod auto-shutdown savings
"Estimate how much we'd save if we auto-shutdown all non-prod workloads outside business hours (7 PM to 7 AM weekdays, all weekend)."
What Billy returns
Annualized savings estimate, workloads contributing most to savings, and workloads where shutdown isn't safe (overnight jobs, off-hours users) so you can carve out exceptions.
- DevOps
- FinOps
- Multi-cloud
Expiring reservations in the next 60 days
"Which AWS reservations are expiring in the next 60 days, and what would those workloads cost on-demand?"
What Billy returns
Expiring RIs and Savings Plans with a renewal recommendation (renew, modify, or let expire), the on- demand cost the workloads would absorb, and a target purchase amount.
Commitment coverage rate and gap analysis
"What's my current commitment coverage rate by service, and where am I leaving savings on the table?"
What Billy returns
Coverage rate per service, on-demand spend that could move under commitments, and a recommended commitment purchase plan with payback period.
- FinOps
- Finance
- Multi-cloud
Optimal RI / Savings Plan mix for next quarter
"Recommend an optimal mix of RIs, Savings Plans, and Reserved capacity for next quarter, factoring in workload stability and growth."
What Billy returns
A recommended portfolio by commitment type and term, expected savings vs. on-demand, and an explanation of why each commitment was chosen.
- FinOps
- Finance
- Multi-cloud
Multi-vendor AI spend rollup
"Roll up our spend on AWS Bedrock, GCP Vertex, direct Anthropic, OpenAI, Cursor, and GitHub Copilot for the past month."
What Billy returns
A single table with each AI vendor's spend, the share of total AI spend, and month-over-month trajectory. Includes unattributed AI cost as a separate line.
- FinOps
- Engineering
- Exec
- Multi-cloud
Bedrock spend by model
"Break my Bedrock spend down by model — Claude Opus, Claude Sonnet, Titan, others — for the past 90 days."
What Billy returns
Per-model spend, input vs. output token cost split, and the application or account each model's usage maps to. Highlights any model where cost grew disproportionately to call volume.
Cursor and Copilot spend per engineer
"What's the average Cursor and GitHub Copilot cost per engineer over the past 14 days, and who are the heaviest users?"
What Billy returns
Average per-engineer spend, top 10 users by cost, and a comparison to peer-company benchmarks where available. Useful for seat-allocation policies.
- Engineering
- FinOps
- Multi-cloud
AI training vs inference breakdown
"Split our AI cost between training and inference for the past quarter, and tell me how the split is trending."
What Billy returns
Training vs. inference cost over time, with the inflection point where one started growing faster than the other.
- Engineering
- Product
- Multi-cloud
Token cost trends with input/output split
"Show me the trend in cost per input token and cost per output token across our top three LLM providers for the past 90 days.
What Billy returns
Cost per million tokens — input and output separately - trending over time per provider. Surfaces provider pricing changes or model swaps that shifted unit economics.
- Engineering
- FinOps
- Multi-cloud
Kubernetes cost by namespace across all clusters
"Show me Kubernetes cost grouped by namespace across all my EKS, GKE, and AKS clusters for the past
30 days."
What Billy returns
Namespace-level cost ranked by spend, with cluster attribution for namespaces that exist in multiple clusters. Idle namespaces called out separately.
- Engineering
- DevOps
- Multi-cloud
Namespace cost with utilization
"For the payment-processing namespace, show me cost alongside CPU and memory utilization for the past 14 days."
What Billy returns
Daily cost overlaid with CPU and memory utilization. Reveals over-provisioned namespaces and rightsizing opportunities. Recommends a request/limit profile.
- Engineering
- DevOps
- Multi-cloud
Multi-cluster cost-per-pod comparison
"Compare cost per running pod across my production clusters. Are there clusters significantly more expensive on a per-pod basis?"
What Billy returns
Per-cluster cost-per-pod metrics, highlighting clusters where unit economics are worse than peers. Reasons (smaller instances, fewer spot nodes, less bin-packing) called out where detectable.
- Engineering
- FinOps
- Multi-cloud
Workload-level cost attribution to owner
"Attribute Kubernetes cost down to the deployment / workload level for the past month, and roll it up to the application owner."
What Billy returns
Workload-level cost rolled up to team or application owner using your existing labels or vTags. Identifies workloads not yet attributed to an owner.
- Engineering
- FinOps
- Multi-cloud
Cost per customer,
by feature
"Calculate cost per customer for each of our top 5 product features over the past month."
What Billy returns
Per-customer cost broken down by feature, with the unit-cost trend over the past 6 months. Highlights features where unit economics are deteriorating.
- Product
- Finance
- FinOps
- Multi-cloud
Cost per transaction, trended
"Show me cost per transaction for our core API, trended monthly for the past 12 months."
What Billy returns
Monthly cost-per-transaction trend, the inflection points, and the architectural or scale changes that drove each shift.
- Engineering
- Product
- Multi-cloud
Fully-loaded cost
per feature
"Give me a fully-loaded cost per product feature - compute, storage, networking, AI, and shared services allocated proportionally."
What Billy returns
Per-feature cost with cost breakdown by category and the share of shared-services allocation. Surfaces features where allocated cost dominates
direct cost.
- Engineering
- FinOps
- Multi-cloud
Turn this answer into a dashboard
"Take the breakdown you just gave me and turn it into a Finout dashboard I can share with my team."
What Billy returns
A Finout dashboard created in your workspace with the same filters and grouping from the conversation, plus a shareable link.
- Product
- Finance
- FinOps
- Multi-cloud
Weekly cost digest, scheduled
"Set up a weekly digest that emails me every Monday with last week's cloud spend, top anomalies, and any new cost-saving opportunities."
What Billy returns
A scheduled report configured in Finout, sent to the address you specify, with cost summary, anomalies, and CostGuard findings each Monday.
Board-ready executive summary
"Build me a board-ready cloud cost summary for last quarter: total spend, year-over-year change, top drivers, savings actions taken, AI spend trajectory, and risks for
next quarter."
What Billy returns
A presentation-ready summary with the quarter's headline numbers, narrative bullets on top drivers and actions, AI spend trajectory chart, and a forward-looking risks section.
Underutilized GPU instances this month
Which of our GPU instances are underutilized this month?
What Billy returns
A ranked list of this month's GPU instances with GPU utilization, GPU memory utilization, and power draw shown next to cost, so you can see which instances are genuinely idle rather than just quiet. Instances already flagged by the GPU Idle scan are called out separately.
Cost vs. CPU for your top 5 EC2 instances
For our five most expensive EC2 instances this month, show cost next to average CPU. Which are expensive and underused?
What Billy returns
Your five highest-cost EC2 instances for the month, with average CPU utilization shown next to each one's cost, and a direct call-out on which instances are both expensive and underused.
- FinOps
- Engineering
- DevOps
- AWS
Verify a CostGuard idle flag before terminating
CostGuard flagged this instance as idle. Show me its CPU and network for the last 30 days before I terminate it.
What Billy returns
30 days of CPU and network utilization for the flagged instance, so you can confirm the idle finding against real usage before terminating it. Links back to the originating CostGuard scan record.
- Engineering
- DevOps
- FinOps
- AWS
Confirm headroom for a rightsizing recommendation
We have a rightsizing recommendation to downsize this instance. Show me peak and average CPU — is there headroom for the smaller size?
What Billy returns
Peak and average CPU for the instance over the recommendation window, with a direct read on whether the smaller instance size has enough headroom to safely absorb peak load.
Check if a resource ran continuously or on a schedule
Was this resource running continuously last month, or only during certain hours?
What Billy returns
The resource's runtime pattern for the past month, broken out by hour, showing whether it ran continuously or only during specific windows, and what that pattern is costing you.
- DevOps
- Engineering
- AWS
- GCP
Snowflake warehouse usage pattern, 60 days
Show me the usage pattern for this Snowflake warehouse over the last 60 days — is it bursty or steady?
What Billy returns
60 days of usage volume for the warehouse plotted against cost, showing whether consumption is bursty or steady, useful for deciding on auto-suspend settings or a warehouse size change.
- Engineering
- FinOps
- Finance
- Snowflake
Generate an API key and enable team tokens
Where do I get an API key, and how do I let my team create their own personal access tokens?
What Billy returns
Step-by-step instructions for generating an API key from your account settings, plus how to enable personal access tokens for teammates, pulled directly from Finout's documentation with a link to the full doc page.
Set up SSO groups
How do I set up SSO groups?
What Billy returns
Walkthrough steps for configuring SSO groups in Finout, sourced from documentation, with a link to the relevant setup guide.
Virtual tag permissions for basic users
Can basic users create and use virtual tags?
What Billy returns
A direct answer on which permission level is required to create and use virtual tags, drawn from Finout's documentation.
Schedule a recurring export
How do I schedule a recurring export?
What Billy returns
Step-by-step setup instructions for scheduling a recurring export, sourced from documentation, with a link to the export settings page.
Connect an S3 bucket for telemetry
How do I connect an S3 bucket for telemetry?
What Billy returns
Setup steps for connecting an S3 bucket for telemetry, pulled from documentation, including the permissions the bucket policy needs.
Exclude promotional credits from cost data
How do I exclude promotional credits from my cost data?
What Billy returns
Instructions for excluding promotional credits from your cost views, sourced from documentation, plus a note on where the setting applies, account-wide or per report.
MegaBill data lag explained
Why is my MegaBill data 24–48 hours behind?
What Billy returns
An explanation of why MegaBill data has a 24 to 48 hour lag, pulled from documentation, so you know it's expected behavior rather than an issue with your account.
Create a virtual tag and see your existing ones
How do I create a virtual tag — and show me which ones we already have?
What Billy returns
Documentation steps for creating a new virtual tag, combined with a live pull of the virtual tags already configured in your account, in a single answer.
