Dashboard tour

Once a billing source is connected, the dashboards are where you read your spend. This page tours the main views so you know where to look for what.

The sidebar is organized into three groups: Dashboards at the top, Insights (the analysis views — Reports, Savings, Commitments, Anomalies, Tags, Alerting), and Manage (org-wide configuration — Allocation, Budgets, Cost centers). A scope picker at the top of each view selects the provider and integration the view reads from, and it stays put as you move between views. A Cost view control in the same bar switches every source to one cost basis — credits, taxes, or amortization — for the session; by default each integration keeps its own cost settings.

The Dashboards menu adapts to what you have connected. With one kind of source it goes straight to that dashboard; once your sources span more than one category — cloud platforms, AI platforms, or a data warehouse — it leads with an org-wide Overview, groups providers under clickable category headers (Cloud, AI), and lists each provider beneath.

The Overview (all sources)

When your organization has more than one category of source connected, landing on the dashboard opens the Overview — every connected source in one ledger roll-up. It surfaces:

Total spend (MTD) — across every source, compared to the same days last month, with a forecasted month-end.
AI platforms share — what portion of total spend is direct Anthropic / OpenAI platform usage.
Sources — how many sources are connected, with a “data through” freshness stamp per source.
Monthly and daily cost by source — stacked charts, one color per provider; click a segment to open that provider’s own dashboard.
Top services across sources — the biggest line items this month, labeled Provider · Service.

Spend that your cloud provider bills — like Claude on AWS Bedrock or Azure OpenAI — counts as cloud spend here (it’s on your cloud invoice), so every number reconciles against the bills you actually receive. Bedrock is also visible in the AI Spend view below, as a lens over those same AWS rows — shown there for comparison, never counted a second time.

The Cloud dashboard

The Cloud dashboard — opened from Dashboards (or automatically as your landing view when cloud is your only source category) — is the home of your day-to-day cloud cost picture. At organization scope it combines every connected cloud provider (AWS, Azure, Google Cloud); pick a provider in the Dashboards menu or the scope picker to focus on one. It’s available on every tier and pulls from the single cost ledger. It surfaces:

KPIs — headline numbers, including month-to-date spend versus the previous month.
Spend trend — how spend is tracking over time, so you can see direction at a glance.
Top movers — the services (and accounts) whose spend changed most, so the biggest swings surface first.
Per-account breakdown — spend rolled up by each linked account or subscription.

The AI platforms dashboard

With an Anthropic or OpenAI integration connected — or with Amazon Bedrock usage on your AWS bill — Dashboards also offers the AI Spend view: token-level cost by model, token type (including cache reads and writes), workspace/project, API key, and service tier. With more than one AI source it opens on All AI with a per-source split; the buttons narrow to one source.

Amazon Bedrock appears here as a lens, not as an integration. Bedrock is billed by AWS, so its rows already live in your cloud spend and are counted there exactly once. The AI Spend view reads those same rows so you can see Bedrock beside your direct Anthropic and OpenAI usage — by model, token type, tier, routing, linked account, region and tag. Nothing is ingested twice, and there is nothing to connect: the lens appears on its own once your AWS bill shows Bedrock usage. Because those dollars sit on your cloud invoice, the All AI total is an analytical figure and is never added to cloud spend. Selecting Bedrock or All AI suspends the scope picker — the AWS-billed half is always org-wide, and a total that narrowed only its direct half would be a half-truth. Pick a single direct provider to narrow again.

Azure OpenAI is Microsoft-billed and appears in cloud spend only.

Two numbers to read carefully:

Effective $ / 1M tokens divides only the spend that actually bought tokens. Non-token charges — server tools, provisioned capacity, guardrails — are in the total but not in that rate, and a note tells you how much of the spend the rate covers.
Data through dates a blended figure by its oldest source. Cloud billing data settles a day or two behind the direct provider APIs, so an All AI total is complete through the AWS side’s last day; later days are partial, not missing.

For the full reporting surface — pivoting by product family, usage type, or tag, and saving or exporting a view — head to Reports under Insights in the sidebar (see Cost reporting).

Where everything else lives

Each sidebar entry goes deeper on one part of FinOps.

Under Insights:

Reports — build detailed pivot reports, save them, and export to CSV. See Cost reporting.
Savings — rightsizing, idle-resource, and commitment recommendations with dollar estimates. See Savings.
Commitments — Reserved Instance and Savings Plan utilization and upcoming expiries. See Commitments.
Anomalies — spot unusual month-over-month and week-over-week swings and investigate the drivers. See Anomalies.
Tags — how much of your spend carries cost-allocation tags, and where the gaps are. See Tag coverage.
Alerting — rules and notification channels that push cost and budget events to Slack, Microsoft Teams, or email. See Alerts.

Under Manage:

Allocation — assign spend to cost centers with allocation rules. See Cost allocation.
Budgets — set targets and watch actual-versus-forecast pace against them. See Budgets.
Cost centers — define the cost centers that allocation assigns spend to. See Cost centers.

Organization-level administration — settings, members, billing, integrations, notification channels, and cost settings — lives behind the Admin entry at the bottom of the sidebar.

Reading spend with an agent

You can also query all of this in plain language by connecting an AI agent through the CloudQuell MCP server — ask “why did our RDS spend jump last week?” and the agent answers from your live data.