Agent Intelligence

FinOps Agent
Action Tree

Complete catalogue of agent-executable actions across all modules — read, write, export and autonomous.

168

Total actions

Read / Query

Write / Action

Autonomous

Show:

Read Query only — no approval

Write Mutates cloud — approval gate required

Export Read with side-effect

Auto Agent-initiated — no UI trigger

Finance Compass 5 sub-modules · 30 actions

Overview5 actions

ReadRealized vs unrealized savings (lifetime)

ReadSpend vs budget chart (configurable time range)

ReadTop 5 entities by spend

ReadTop 5 entities by absolute budget deviation

ReadVariance % per entity

Spend Breakdown6 actions

ReadTotal monthly spend ($277K)

ReadDrill by application (387 items)

ReadDrill by cloud provider

ReadDrill by entity

ReadDrill by service (35 items)

ReadProduction vs non-production split (1% / 99%)

Commitment Efficiency — Savings Plans6 actions

ReadSavings plan utilization %

ReadUsed vs committed hours per plan

ReadNet savings per savings plan

ReadFilter plans by zone / type / platform / offering / efficiency band

ReadIdentify underutilized savings plans (below configurable threshold)

ExportSavings plan utilization report

Commitment Efficiency — Reservations8 actions

ReadList all reservations with utilization % (e.g. 96.8% → 78.6%)

ReadUsed vs total hours + unused hours per reservation

ReadReservation cost, savings, and date range

ReadTotal cost / net savings / realized / unrealized across all reservations

ReadFilter by zone / type / platform / offering / efficiency

ReadSearch reservation by instance name

ReadIdentify reservations expiring within 30 days

ExportReservation utilization report

Savings by Category6 actions

ReadCompute savings — realized / unrealized ($49,504)

ReadDatabase savings — realized / unrealized

ReadStorage savings — unrealized ($430)

ReadSnapshot savings

ReadCompression-driven savings

ExportFull savings report (CSV / JSON)

Technology Compass 7 sub-modules · 47 actions

Alerts6 actions

ReadList active alerts with count and last refresh time

ReadFilter by alert type / service type / cloud provider

ReadSearch by alert type / asset ID / cloud account

ReadGroup alerts by type / account / provider

WriteAcknowledge single alert

WriteAcknowledge all alerts (bulk)

Compute Recommendations — VM / LBVM / K8s (321 assets)13 actions

ReadList all assets sorted by savings potential

ReadFilter by department / application / cloud provider / tag

ReadCurrent server family per asset

ReadCurrent spend and potential savings per asset

ReadSafe profile thresholds (CPU ≤99% / Mem ≤100% / Net ≤99%)

ReadAlt profile thresholds (CPU ≤80% / Mem ≤100% / Net ≤80%)

ReadCurrent vs proposed family and spend comparison

ReadCurrent vs proposed utilization % comparison

ReadMax CPU / Memory / Network per asset (performance metrics)

ReadUptime % for proposed configuration

ReadAsset tags (Account, Application, CostCenter)

WriteAccept Safe recommendation → trigger VM resize

WriteAccept Alt recommendation → trigger VM resize

DB Recommendations5 actions

ReadList databases with right-sizing opportunities

ReadCurrent DB engine and instance class

ReadDB utilization metrics (CPU / memory / connections)

ReadRecommended DB instance family and projected savings

WriteAccept DB right-sizing recommendation

Unattached Disks (5 assets)7 actions

ReadList unattached disks (ID / type / provisioned size / cost/month)

ReadFilter by department / provider / tag key+value

ReadSafe storage recommendation (Object Cool/Hot tier)

ReadAlt storage recommendation with trade-offs (backup time, throughput)

ReadSnapshot storage recommendation and cost

ReadFull tag set per disk (CitrixCustomerId, CostCenter etc.)

WriteAccept storage tier recommendation

Idle Disks (21 assets)6 actions

ReadList idle volumes with idle %, host asset, provisioned size, cost/month

ReadBlock storage Safe recommendation (E1 family)

ReadBlock storage Alt recommendation (E2 family)

ReadSnapshot storage recommendation

ReadTags per idle volume

WriteAccept idle disk recommendation

IOPS Recommendations (48 assets)8 actions

ReadList volumes with IOPS over-provisioning

ReadProvisioned vs required IOPS per volume

ReadCurrent cost / Safe cost / savings delta

ReadRecommended storage type (e.g. s15)

ReadSupportable IOPS at recommended tier

ReadVolume tags (Account, CostCenter, NS_Application, Name)

WriteAccept Safe IOPS recommendation

WriteAccept Alt IOPS recommendation

Application Recommendations (313 apps · 395 assets)6 actions

ReadList assigned applications with resource overview

ReadFilter by dept / cloud provider / account / asset type

ReadPer-app VM / DB / Storage / IOPS count

ReadTotal current monthly cost per application

ReadPotential monthly savings per application

WriteAccept VM recommendation scoped to application

Tag Compass 5 sub-modules · 32 actions

Overview — compliance dashboard10 actions

ReadOverall compliance rate (56% compliant / 34% partial / 9% missing)

ReadCompliance breakdown by cloud provider (AWS / Azure / GCP)

ReadTotal assets and weekly delta (+4 this week)

ReadFully-compliant / partial / missing counts with trend

ReadDrift events in last 24h (18 events, last scan 4m ago)

ReadPer-asset compliance score (0–100%)

ReadMandatory tag policy (Environment, CostCenter, Owner, Domain, DataClass)

ReadOptional tag policy (Department, Region, BillingCode, Tier, Compliance)

ExportCompliance report

SyncTrigger tag re-scan across all cloud accounts

Tag Filters — policy and asset search7 actions

ReadAllowed values per mandatory tag (e.g. Environment: dev / staging / prod / sandbox)

ReadWildcard search assets by tag key or value (env:prod* / owner:*corp)

ReadFilter asset list by cloud / status (compliant, partial, missing) / resource type

ReadPer-asset mandatory + optional tag presence matrix

ReadAsset compliance status (Compliant / Partial / Missing)

WriteToggle mandatory tag active/inactive in policy

WriteApply wildcard filter and scope remediation to matching assets

Drift Monitor — change detection8 actions

ReadList all drift events in scope (18 events)

ReadAggregate counts — tags added / removed / values changed

ReadPer-event actor (service account / pipeline / user)

ReadBefore vs after tag diff per event

ReadEvent source (CloudTrail / Azure Activity Log / GCP Audit)

ReadEvent status (COMPLIANCE_VIOLATION / informational)

ExportCopy drift event JSON payload

WriteCreate ITSM ticket from drift event (Jira / ServiceNow)

Action Required — remediation queue (28 assets)7 actions

ReadList assets in remediation queue sorted by severity

ReadFilter by severity / cloud / missing tag type

ReadMissing mandatory tags per asset

ReadMissing optional tags per asset

ReadDrift reason and annotation per asset

WriteFix single asset — apply missing tags with values

WriteBulk remediate — apply tags to all flagged assets

Trend Report4 actions

ReadWeek-over-week compliance trend

ReadDrift rate trend over time

ReadTop violating services / accounts historically

ExportTrend report

Databricks Compass 3 sub-modules · 19 actions

InfraOps — cluster management9 actions

ReadList all Databricks clusters (name, type, state)

ReadCluster utilization % (CPU / memory / DBU)

ReadIdentify idle clusters (running but no active jobs)

ReadCluster cost per hour and per month

ReadCurrent auto-termination settings per cluster

ReadSpot vs on-demand instance mix per cluster

WriteTerminate idle cluster

WriteSet / update auto-termination timeout

WriteResize cluster — scale down worker count

QueryOps — query cost analysis6 actions

ReadTop expensive queries by DBU / cost

ReadQuery execution time and scan bytes per query

ReadQuery author / notebook / job attribution

ReadIdentify full-table scans with no partition pruning

ReadPer-workspace / per-team query spend

ExportQuery cost report

Job cost tracking4 actions

ReadList Databricks jobs by cost (DBU consumed)

ReadJob run history and average cost per run

ReadIdentify failed / retried jobs inflating spend

ReadJob-to-cluster assignment

AI Spend Ops 3 sub-modules · 15 actions

AI spend observability7 actions

ReadTotal AI / LLM API spend by provider (OpenAI / Anthropic / Bedrock / Vertex)

ReadDrill spend by model (gpt-4o / claude-3 / gemini etc.)

ReadDrill spend by team / application / cost center

ReadToken usage — prompt vs completion — per model

ReadCost per 1K tokens per model

ReadSpend trend (daily / weekly / monthly)

ExportAI spend report

Model optimization4 actions

ReadIdentify teams using expensive models where cheaper models suffice

ReadLatency vs cost trade-off per model

ReadCache hit rate and repeated prompt savings

ReadDetect anomalous AI spend spikes

Budget and limits4 actions

ReadAI budget vs actual per team

ReadQuota / rate-limit utilization per API key

WriteSet / update AI spend alert threshold per team

WriteFlag API key for review if spend exceeds budget

Autonomous Agent Capabilities 6 domains · 28 actions · no UI trigger required

Scheduled scanning and monitoring5 actions

AutoPoll cloud APIs hourly for new idle or unattached resources

AutoRun nightly tag compliance scan across all accounts

AutoMonitor reservation utilization daily — flag if below threshold

AutoTrack savings plan coverage % — alert if new spend is uncovered

AutoDetect new cloud resources deployed without mandatory tags

Anomaly detection6 actions

AutoDetect spend spikes >20% vs 7-day moving average

AutoDetect new resource types with no budget allocation

AutoFlag VMs running at <5% CPU for 7+ consecutive days

AutoFlag disks with 0 read/write IOPS for 14+ days

AutoDetect tag drift triggered by IaC / automation scripts

AutoDetect reservation expiry within 30 days and alert for renewal

Auto-remediation (policy-gated)5 actions

AutoAuto-apply default tags to new resources (from account / team policy)

AutoAuto-terminate Databricks clusters idle beyond configurable timeout

AutoAuto-snapshot then delete confirmed idle unattached disks

AutoAuto-downscale dev / sandbox VMs outside business hours

AutoAuto-create ITSM ticket on COMPLIANCE_VIOLATION drift event

Cross-cloud correlation4 actions

AutoCorrelate spend spike with recent deployment via git commit / release tag

AutoMap untagged resources to owning team via network / VPC context

AutoAttribute Databricks job cost to application via tag inheritance

AutoReconcile Finance Compass entity spend with Tag Compass owner field

Proactive reporting4 actions

AutoGenerate weekly FinOps digest (top savings, new alerts, compliance delta)

AutoGenerate monthly commitment efficiency report (utilization trends)

AutoPush spend-vs-budget alert to Slack / Teams when entity exceeds threshold

AutoSurface reservation purchase recommendation when coverage drops

Approval workflow management4 actions

AutoQueue write actions for human approval with full context and rollback plan

AutoAuto-approve low-risk actions below defined savings / risk threshold

AutoEscalate high-impact actions (bulk resize, bulk tag) to senior approver

AutoLog all agent actions with actor, timestamp, before/after state to audit trail

FinOps AgentAction Tree

LeanKloud FinOps agent action tree — 168 actions across 6 modules

FinOps Agent
Action Tree