Audit report

The AI Spend Agent generates an audit report for each connected provider: a self-contained, printable one-pager built for a finance reader. It condenses the agent’s reconciliation data and findings into the headline numbers, a plain-English spend story, the issues worth a look, the structural savings on the table, and an honest statement of what the billing reconciliation can and can’t confirm. Every figure in the report is computed from the same reconciliation rows and findings that power the dashboard. Nothing is hand-entered except the negotiated-rate market benchmark, which the report labels as a declared benchmark rather than a computed figure.

​

Opening the report

From the agent’s Overview section, click Audit report next to the provider tabs. The report opens in a new browser tab as a standalone page, scoped to the provider tab you have selected (or your first connected provider if All providers is active). The button appears once at least one connected provider has produced usage. Because the report is a plain HTML page, you can use your browser’s print dialog to save it as a PDF or print it, and forward it as-is — it has a print-friendly layout built in. If the agent hasn’t reconciled any usage for that provider yet, the report renders a “No audited spend yet” state instead of fabricating numbers. Run a sync and audit first — see Sync now and Refresh insights.

​

What the report contains

​

KPI band

Up to four headline numbers across the top:

• Audited spend — Total billed cost across the reconciled window. The window is the full date range of reconciled usage, shown in the report header, not a fixed 30 days.
• Growth — Spend growth from the first month to the last complete month, compared at per-day rates so a partial first or current month can’t distort the figure.
• Biggest month — The highest-spend complete month. A partial in-progress month is never counted as a record.
• Recent pace — Average spend per day over the trailing window (up to 14 days). This same trailing window is the basis for the annualized savings estimates, so the two never disagree.

​

Spend narrative

A plain-English paragraph that walks through monthly spend, the current month’s daily pace, weekend-vs-weekday mix (steady weekend spend suggests production workloads; a weekend drop suggests interactive use), and which models carry most of the bill. Project counts appear only for providers that expose a project dimension.

​

Worth your attention

The highest-impact open findings, each with its own dollar figure over its own period (a daily figure, a monthly rate, or an annualized pace). Duplicate instances of the same rule are collapsed to the highest-impact one, and the list is capped at a handful so the page stays a one-pager. The report also calls out new model adoption: models that first appeared partway through the window and are already running at a material annual pace, so you can confirm the rollout was deliberate.

​

Savings opportunities

The genuinely annual, structural levers, with a combined headline range:

• Batch pricing — Provider batch APIs are a flat 50% off for async-tolerable work. The report estimates this from your annualized embeddings line and only shows the item when the saving is material.
• Negotiated rates — At list price with a sufficient annual run rate (above $100,000 a year), a 10–20% discount against a 12-month commitment is commonly reported market practice. The report states this as a market benchmark, not a computed entitlement.

The two items are made disjoint — the negotiated-rate estimate excludes the batch-eligible spend — so the headline range never double-counts the same dollars.

​

Billing integrity

A statement of what the reconciliation confirmed for this window: whether billed totals reconcile against provider-reported usage at the account level, and whether any billing discrepancy is confirmed. Anything confirmed links into the Recovery flow.

​

Per-provider differences

The report says different things per provider because the providers report cost differently, and the report only makes claims its data can back:

• OpenAI — OpenAI’s cost API is model-blind: it reports cost per project, so per-model billed cost is an apportioned estimate. An OpenAI report therefore never claims a confirmed overcharge. Its integrity statement stays at “no material provider overcharge confirmed,” and any recovery is invoice-side — discrepancies must be validated against your provider invoice first. This matches the Confirmed vs. Potential split in Recovery.
• Anthropic — Anthropic reports real per-model cost, so its report can state per-model reconciliation and can surface a confirmed billing discrepancy when one exists.

Findings are scoped to the report’s provider. An Anthropic discrepancy never affects the OpenAI report’s integrity statement, and customer-level findings that aren’t tied to a provider are left to the dashboard views.