Cloud cost allocation: route every cloud and AI dollar to a cost center

How allocation works

A rule matches spend on a dimension your cost data already carries and routes it to one cost center. Nothing has to be retagged first.

Cost centers

Named buckets for teams, products, or business units, nested up to five levels deep. Spend enters at the leaves and rolls up, so an Engineering parent shows Platform, Data, and Mobile combined.

Three rule types

Match a tag (key and value, such as team = platform), a service (all of Amazon EC2, or Snowflake Compute), or an entire integration (everything from one connected OpenAI account). Scope each rule org-wide or to a single integration. Matching is exact; there are no wildcards.

Virtual tags

Because rules match dimensions already on the cost row, they allocate spend that carries no usable native tag — the cost center a rule assigns is the virtual tag. Untagged legacy resources don't wait on a retagging project.

Precedence

Rules are checked in one global priority order and the first match wins. The Resolution order tab lets you drag rules into place and preview current versus projected spend per cost center before you apply the order.

AI and Snowflake

OpenAI projects (project_id tag) and Anthropic workspaces (workspace_id tag) route to the same cost centers as cloud spend. Snowflake compute is split per warehouse by daily credit share, marked estimated, so a warehouse tag rule can allocate one warehouse.

One center per cost

Each rule sends what it matches to one cost center and the first match wins, so a shared line — a support fee, a shared database — goes to a single center. There are no percentage splits of one line across several teams.

What an allocation looks like

Specific rules first, broad catch-alls after. Spend none of them claims stays in Unallocated, where it is easy to find.

Tag coverage and showback

Allocation is only as good as the tags and boundaries under it. Tag coverage shows where the gaps are; the Allocation page and reports show each team its cost.

Tag coverage

The Tags view shows the share of cloud spend on resources carrying any tag, then coverage per tag key — env and Environment fold into one row — a Coverage by integration table, and the services with the most untagged spend. AWS tags, Azure tags, and Google Cloud labels share one namespace.

An honest headline

Azure's built-in resource_group, invoice_section, and cost_center keys don't count as tagged, so Azure can't read near 100% by default. AI and Snowflake spend stays out of the denominator: those platforms have no user-defined resource tags to govern.

Allocation page

Month to date by cost center: total spend, active rules, a coverage bar for the share mapped to a cost center, the Unallocated amount, and a Budget column with each center's budget pace. A parent row shows its whole subtree.

Showback reports

Cost Center is a report dimension: group a report by it, export to CSV, or schedule it by email, Slack, or Teams. A report reads one provider at a time, and the scheduled recap of a cost-center report summarizes by service — its link opens the full breakdown.

Who can edit

Contributors and above create and change cost centers and rules; viewers can read them. Agents connected over MCP can read cost centers, rules, and the allocation breakdown, but cannot change them.

Plans

Cost centers, allocation rules, and the per-tag-key coverage breakdown are on paid plans: Starter $99/month up to $50K of monthly spend, Growth $199/month up to $200K. The headline Coverage number is on every plan, Free included. The Tags view appears once AWS, Azure, or Google Cloud is connected.

What it unlocks

Once the bill sums to teams, every other control can be scoped to them.

• Budgets per cost center

  Scope a budget to a cost center; a budget on a parent measures its whole subtree with the same engine as the Allocation page, so the two reconcile. Thresholds default to 80% and 100%, delivered to Slack, Microsoft Teams, or email.

  Budgets docs →

• AI cost allocation

  A team's OpenAI project, Anthropic workspace, and tagged cloud spend land in the same cost center, so its AI line sits beside its infrastructure. Bedrock, Azure AI Foundry, and Vertex AI allocate with their cloud's tag rules.

  AI cost allocation by team →

• Showback or chargeback

  The Allocation page gives each team its share. Report it to the team (showback) or move it into the team's budget in your accounting system (chargeback) — CloudQuell produces the per-team numbers; the internal transfer happens in your finance system.

  Chargeback vs showback →

• Allocation an agent can read

  Over MCP, get_allocation_breakdown returns each cost center's own and rolled-up spend, its share of total, and the Unallocated bucket — for one integration or across all of them.

  MCP server (docs) →

Common questions

What is cloud cost allocation?

Assigning each cloud charge to the team, product, or business unit that caused it, so spend can be reported, budgeted, and owned. Bills arrive by service and account; allocation maps them to your org chart. In CloudQuell that means cost centers filled by rules that match a tag, a service, or an entire integration, with anything unmatched shown as Unallocated.

What is chargeback in cloud computing, and how is it different from showback?

Showback reports each team's cloud cost without moving money; chargeback bills that cost to the team's budget in your accounting system. Both run on the same allocation underneath. CloudQuell's Allocation page and cost-center reports supply the per-team numbers for either; the internal transfer for chargeback happens in your finance system.

How do I allocate untagged cloud costs?

Usually both of two ways. Fix tags at the source so future spend is attributed automatically, and bridge the rest with allocation rules — virtual tags — that match a service, an entire integration, or a tag the row already carries. Start broad, such as all of Amazon EC2 to Infrastructure, then add more specific, higher-priority rules for the exceptions.

What is tag coverage?

The share of your cloud spend on resources that carry tags. CloudQuell's headline Coverage counts any tag at all, so treat it as an upper bound; the per-key breakdown shows coverage for the keys allocation actually relies on, such as team or cost-center, and which services carry the most untagged spend. The headline is on every plan; the per-key detail is on paid plans.

Can I allocate AI costs like OpenAI and Anthropic to teams?

Yes. OpenAI cost rows carry a project_id tag and Anthropic rows a workspace_id tag, so one rule per project or workspace routes it to that team's cost center — the same center as its cloud spend. Spend in OpenAI's Default project or Anthropic's Default workspace carries no ID and stays Unallocated. Bedrock, Azure AI Foundry, and Vertex AI allocate with their cloud's tag rules.

Which plans include cost allocation?

Cost centers, allocation rules, and the per-tag-key coverage breakdown are on every paid plan: Starter $99/month for up to $50K of monthly spend, Growth $199/month up to $200K — flat, never a percentage of your bill, with a 14-day trial and no credit card. The Free plan, for spend under $10K/month, shows the headline tag Coverage number; the Allocation and Cost centers views stay visible but locked.