Why 30% Untagged Spend Is a Structural Problem, Not a Data Quality One
Untagged cloud resources accumulate because tagging discipline degrades at the edges of organizational control — ephemeral workloads, developer sandboxes, legacy migrations, and resources spun up before your tagging policy existed. By the time $25,000–$55,000 per month in cloud spend is untagged, the backlog is too large for manual remediation and too distributed across AWS, Azure, and GCP for any single team to own. Apptio Cloudability surfaces the problem; ServiceNow CMDB has the ownership data to potentially solve it — but connecting them at scale requires matching logic that doesn't exist in either system. The FinOps chargeback program stalls because the foundational data isn't there.
Inference-Based Tagging With Human Approval at Every Step
An AI Labor Company agent mines Apptio Cloudability untagged spend data against ServiceNow CMDB ownership records to infer correct cost center, team, application, and environment tags for untagged resources. It generates bulk remediation plans — specific tag assignments for specific resources — and routes them to the appropriate cloud team lead for approval before any tagging action is applied. Terraform generates the approved changes; no resource is retagged without explicit human sign-off. Slack surfaces pending approvals and tracks remediation progress. The agent continuously monitors for new untagged resources, catching drift before it accumulates rather than auditing after the fact.
What Fixing Tagging Is Actually Worth
Moving untagged spend from 30% to under 5% — the typical outcome range, achieved in approximately 4 weeks — unlocks the chargeback program that's currently stuck. That has direct financial value: business unit leaders get accurate cost attribution, chargebacks become defensible, and engineering teams face real accountability for their cloud consumption. The downstream effects often include meaningful spend optimization, because teams that see their actual cloud costs make different purchasing decisions. Illustratively, a Fortune 500 with $1M+ per month in cloud spend recovering full attribution on 25% of it gives FinOps the data to identify and act on the largest cost reduction opportunities. The tagging remediation itself is the foundation — accurate attribution is what makes everything else in the FinOps program work.
Does the agent work across all three clouds simultaneously?
Yes. The agent ingests untagged spend data from AWS, Azure, and GCP through Apptio Cloudability and applies cloud-specific tagging conventions for each provider. Remediation plans are generated per cloud with the appropriate tagging API for each.
What if the ServiceNow CMDB ownership data is incomplete or out of date?
CMDB gaps are surfaced as a separate remediation queue — the agent flags resources where it cannot infer ownership with confidence and routes them to the FinOps team for manual assignment. Improving CMDB coverage is typically a parallel workstream that the tagging program accelerates.
How does the agent prevent new tagging drift after the initial remediation?
The agent runs continuously, monitoring for newly untagged resources and flagging them within days of creation rather than quarterly audits. Slack notifications and a persistent remediation queue keep the cloud teams accountable on an ongoing basis.