Does this replace Cost Explorer or our FinOps tool?

No. Agents read from Cost Explorer, the Cost and Usage Report and tools like Kubecost. They add the part those tools leave to people: finding the change and the owner behind the number.

What access does it need?

Read-only access to billing data and the accounts you choose. Applying a saving needs a separate, scoped permission that you grant per account.

Can agents change resources on their own?

Only where your policy allows it. Most teams start with agents explaining and proposing, then let them apply low-risk savings like deleting unattached volumes.

What if our tagging is incomplete?

Agents also map spend to owners through accounts, repositories and the Terraform that created a resource. Gaps are reported so you can fix the tags.

AWS Partner

SOC 2 Reports

Frontier Investigationfor Cloud Cost

Every cost spike traced to the change that caused it, with a fix.

Connect your billing data and the accounts behind it. Agents pick up each cost anomaly, line it up against deploys, scaling events and config changes, and hand the owning team the change that caused it. Waste comes with a fix ready to approve, sized in dollars per month.

Book a Demo See how it works

AWS Cost Anomaly Detectioncause found

NAT Gateway spend up 340% in us-east-1

Spend: $412/day against a 30-day average of $94/day
Change: Terraform apply at 14:05 moved batch workers to private subnets
Cause: S3 traffic now routes through NAT. No gateway endpoint
Owner: data-platform, tagged on 38 resources
Fix: Add an S3 gateway endpoint. Saves about $9,500/month

Investigated in 2m 10s. Waiting for owner approval

[The work behind every cost spike]

The bill shows what you spent.Finding why still takes days.

01The manual work

Investigate

A spike shows up in Cost Explorer days later. Someone slices by service, account and tag, then asks around until a team admits to the change.

02The agent handoff

Cause found

Frontier agents trace the anomaly to the deploy, scaling event or config change behind it, name the owner and prepare a fix.

03Your engineers’ role

Approve

Set the savings and risk limits. Review the proposed change and decide what ships.

[Where CloudThinker fits]

Your stack stays.Agents work inside it.

01Billing data

Already in your bills

AWS Cost Explorer and CURDaily spend by account and service
Azure Cost ManagementCost analysis and exports
Google Cloud Billing exportBilling data in BigQuery
Kubecost and OpenCostKubernetes cost by namespace

No change to how you bill

02Change and usage

Your system of record

TerraformInfrastructure as code history
GitHub and GitLabCommits and merge requests
Amazon CloudWatchUsage and utilization metrics
Datadog and GrafanaService metrics and dashboards

Source of truth stays put

03Investigation

CloudThinker

Spot the spikeAnomalies by account, service and tag
Trace the changeThe commit or config edit behind it
Price the fixMonthly savings before you approve
ProposeA reviewed change, not a guess

Read-only by default

04Response

Savings, not reports

SlackSpike and cause posted to the owning team
Microsoft TeamsApprove the fix from the thread
JiraTicket for the owner with the evidence
GitHubPull request with the change

Fix runs on approval

Logos show common stacks. CloudThinker connects to each one through read-only access you approve.

[Example scenario]

Thursday, 09:30. The NAT bill quadrupled.Nobody remembers changing anything.

A 60-engineer SaaS company on AWS, 14 accounts under one organization. Cloud spend reviewed monthly, a board update due Friday.

09:30
Anomaly detectedSignal
AWS Cost Anomaly Detection: NAT Gateway in us-east-1 at $412/day, against a 30-day average of $94.
09:31
Agent picks it upAgent
Pulls the Cost and Usage Report by resource, VPC Flow Logs and the last 7 days of Terraform applies in the account.
09:33
Cause, with proofAgent
A Terraform apply on Monday moved batch workers to private subnets. Their S3 traffic now goes through NAT. No gateway endpoint exists.
09:34
Owner found, fix sizedAgent
Tags point to data-platform. Merge request adds an S3 gateway endpoint. Estimated saving about $9,500 a month.
10:05
Owner approvesYour team
The data-platform lead checks the flow logs, approves the merge request and lets the pipeline apply it.
11:40
Saving verifiedAgent
NAT data processing back to baseline. Hourly spend confirms the drop. Finance gets a note with the before and after.

#finops-alerts4 messages

AWS Cost Anomaly Detection09:30
Anomaly: NAT Gateway us-east-1, +$318/day above expected spend.
CloudThinker09:34
Cause: Terraform apply on Monday (run 5521) moved batch workers to private subnets. S3 traffic now routes through NAT. Fix: add an S3 gateway endpoint, MR !842. Saves about $9,500/month. Owner: data-platform.
Data platform lead10:05
Makes sense, we missed the endpoint. Approved and merged.
CloudThinker11:40
Endpoint live. NAT data processing down 82% in the last hour. Anomaly closed. Summary sent to the FinOps lead.

from anomaly to cause and owner: 4 min
monthly saving in this example: $9.5k
spreadsheets passed around: 0

An illustrative example. Team, systems and times are representative, not a specific customer.

[Frontier investigation agents]

Every anomaly gets an explanation.Every explanation gets an owner.

Agents read billing and infrastructure together, so a cost alert arrives with its cause and its owner, and a saving arrives as a change, not a report.

Trace spend to a change: Deploys, Terraform applies, scaling events and config edits lined up against the day the spend moved.
Find the owner: Tags, accounts and repositories mapped to the team that made the change, even where tagging has gaps.
Cut the waste: Idle resources, oversized instances and missing endpoints found with a monthly saving attached.
Hand over a fix: Each saving comes as a reversible change that runs only at the approval level you set.

[What changes]

Same team. Same tools.Far less of the work by hand.

Moment	Today	With frontier agents
Spotting a spike	At month end, on the invoice	The day the anomaly starts
Finding the cause	Slicing Cost Explorer by hand	Spend traced to the change behind it
Finding the owner	A message to every team channel	Owner named from tags, accounts and repos
Acting on savings	A recommendations report nobody owns	A reviewed change with the saving attached
Keeping it fixed	The same waste returns next quarter	Repeat patterns flagged before they cost

[Integrations]

Connects to the rest of your stack.Read-only to start.

AWS Cost Explorer
AWS Cost and Usage Report
AWS Cost Anomaly Detection
AWS Compute Optimizer
Amazon CloudWatch
Kubecost
OpenCost
Terraform
Datadog
Grafana
Slack
GitHub
GitLab
ServiceNow

[Adoption path]

One pilot.Then company-wide.

The rollout follows the four phases of the AWS Cloud Adoption Framework, so it fits the plan your cloud team already runs.

01Envision
Pick one costly account
Connect billing and one account read-only. Let agents explain the last month of anomalies and compare with what finance found.
02Align
Agree the savings policy
Decide which savings agents may apply alone, which need approval, and who signs off for each team.
03Launch
Roll out account by account
Add each business unit’s accounts on the same policies, owner mapping and audit trail.
04Scale
Make it the default
New accounts launch with cost investigation on. Findings feed budgets, tagging rules and architecture reviews.

[Customer proof]

F88.Results on the record.

A Vietnamese consumer-finance company with 800+ branches added continuous cost and resource optimization to its governed hybrid-cloud operating model.

Read the case study

lower AWS spend: 30%
of daily operations automated: 80%

[AWS guidance]

Built the way AWS says to run it.Reviews start from a known reference.

[Trust and control]

Agents do the work.Your team keeps control.

You approve every change: Agents propose. Nothing touches production until someone on your team says yes, and you set that rule per system.
Every action on the record: Each step is logged, attributed and reversible, ready for your auditors.
Certified for enterprise: SOC 2 Type II and ISO 42001, with reports in our trust center.
Runs where you need it: In our cloud, through AWS Marketplace, or inside your own account.

[Questions]

What teams askbefore they start.

Does this replace Cost Explorer or our FinOps tool?: No. Agents read from Cost Explorer, the Cost and Usage Report and tools like Kubecost. They add the part those tools leave to people: finding the change and the owner behind the number.
What access does it need?: Read-only access to billing data and the accounts you choose. Applying a saving needs a separate, scoped permission that you grant per account.
Can agents change resources on their own?: Only where your policy allows it. Most teams start with agents explaining and proposing, then let them apply low-risk savings like deleting unattached volumes.
What if our tagging is incomplete?: Agents also map spend to owners through accounts, repositories and the Terraform that created a resource. Gaps are reported so you can fix the tags.

[Go deeper]