Dashboard sections

The AI Spend Agent dashboard is organized into sections you navigate from the left rail. Each section answers a different question: how much are you spending, where is it going, what’s wrong, what to do about it, and what’s been recovered.

​

Overview

Overview is the headline view: how much your team is spending on AI, how that’s trending month over month, and the anomalies worth your attention. A date-range picker (7, 30, or 90 days) scopes the KPI strip, the trajectory row, and the daily spend chart. Next to it, a provider tab strip (“All providers” plus one tab per connected provider that has usage) scopes every metric on the page to that provider. Only providers with usage get a tab — a connected provider that hasn’t produced any spend yet won’t appear.

​

KPI strip

• N-day spend — Total cost across your connected providers over the selected range.
• Tokens — Total tokens consumed in the range, split between input and output.
• Cache-read rate — Share of input tokens served from a prompt cache, plus the raw read vs. written token counts. High cache-read rates mean you’re paying less for repeated context.
• Models tracked — How many distinct models the agent has seen usage from in the window.
• Price / 1M tokens — Blended cost across input and output tokens in the window, expressed per million tokens. Useful as a headline unit price you can compare across periods or after switching models.
• Cost / request — Average cost per provider-reported request. The average is taken only over rows the provider reports a request count for, so the cost basis is the request-reporting spend rather than total spend. Anthropic’s usage API is token-only, so if all your usage is Anthropic this shows — with provider does not report request counts.

​

Spend trajectory and budget

A second row of tiles projects where spend is heading:

• Projected monthly run-rate — Spend extended to 30 days at the selected range’s daily average.
• vs prior N days — Momentum: how the selected range compares against the equal-length window immediately before it. The tile shows — on the 90-day range, because no equal-length prior window fits inside the agent’s 90-day fetch.
• Budget — Budget vs actual against the spend target you set in Configure: actual spend to date against the target amount, with a straight-line projection to the end of the period and how far over or under target that projection lands. Budget targets are measured against aggregate spend across all providers, so this tile stays account-wide even when a provider tab is selected (the label says so). If no target is set, the tile prompts you to set a monthly spend target from Configure — see Set spend targets.

​

Monthly spend trend

A server-computed chart of spend per month, using the same metric layer as the audit report, so the two never disagree. The growth headline compares the first complete month to the last complete month at per-day rates, so a partial first or current month can’t distort the figure. Partial edge months render muted in the chart and never drive the headline. The biggest complete month is called out alongside the growth figure. Where any insight figures are estimates (for example OpenAI’s model-blind cost API), a caveat note renders below the chart naming the estimate basis.

​

Daily spend chart

Per-day spend bars for the selected range with a 7-day moving average line to smooth out day-to-day noise. Overview is the fastest way to spot a spend spike or a sudden drop in cache efficiency.

​

Anomalies at a glance

The highest-impact open findings (up to four), each with a dollar figure normalized to a monthly rate. The section header states the analysis window the figures cover (for example last 90 days, as monthly rates) because the agent’s audit window is independent of the page’s date-range picker. Click All anomalies to open the findings explorer, where each finding pairs with its recommended action.

​

Audit report

Next to the provider tabs, Audit report buttons open a generated, printable per-provider report in a new tab. A report is always scoped to one provider, so on All providers each provider with usage gets its own labeled button — nothing is picked silently. An optional Report window start/end date pair scopes the report to exact dates; leave it blank and the report follows the date range you’re viewing. See Audit report.

​

Usage

Usage has two views behind a Breakdown | Trends toggle at the top of the section.

​

Breakdown

Breakdown decomposes the selected date range (7, 30, or 90 days) by dimension. Each section shows a spend-mix donut alongside a ranked table.

• Spend by model — Every model that produced cost in the window, with input, output, cache-write and cache-read tokens per model plus each model’s share of total spend.
• Spend by provider — The same breakdown grouped by Anthropic vs. OpenAI.
• Spend by API key — Which keys are driving cost, useful for isolating a runaway integration.
• Spend by team member — Cost attributed to each team member, ranked. Click a row to open that member’s drill-down page — see Team member drill-down.

Use Breakdown when Overview surfaces a change and you need to attribute it: which model got more expensive, which provider is driving the shift, and whether cache usage moved along with it.

​

Trends

Trends shows monthly spend series aggregated on the server, so the range can reach past the 90-day window Breakdown works from. Pick a range of 3m, 6m, or 12m.

• Monthly spend by provider — Stacked monthly bars per provider.
• Monthly spend by model — Stacked monthly bars for the top models, with the remainder grouped as other.

When OpenAI usage is in the mix, a note below the charts flags that per-model figures include apportioned estimates, because OpenAI’s cost API reports totals rather than per-model cost. The toggle stays available even when the recent 90 days have no usage, so months of older history remain reachable from the Trends tab.

​

Cascading filters

The filter bar above the breakdowns is cascading: each dropdown only shows values that are still possible given the ones above it. Pick a provider and the Model dropdown narrows to that provider’s models; pick a model and API Key narrows to keys that used it; and so on down through Team Member. If an upstream change makes a downstream selection impossible (for example, switching to Anthropic while an OpenAI model is picked), the downstream dropdown resets to All on its own.

​

Team member drill-down

From the Spend by team member breakdown, click a row to open a per-member view for the last 90 days. It shows:

• Spend, Tokens, Cache-read rate, Requests — The member’s headline numbers, with team-median comparisons on cache-read rate and team-p90 comparisons on tokens per request so you can see whether they’re an outlier or in line with the rest of the team.
• Model mix — A donut of which models the member’s spend went to.
• Recommendations — Per-member optimization suggestions with an estimated annual saving where the agent can quantify one.
• Findings for this member — Every open finding attributed to that member’s email, so you can see anomalies and violations tied to their usage in one place.

The member view is a drill-down of the same data, not a separate audit — findings match what appears elsewhere in the dashboard.

​

Anomalies

Anomalies collect behavioral and efficiency findings the agent produced from your usage. Each row is one finding, tied to a specific rule (e.g. spend_spike_anomaly, cache_efficiency_outlier, model_overkill, stale_key_active). Categories covered here:

• Usage patterns — Spend spikes, off-hours usage.
• Governance — Stale keys, workspace concentration, subscription-cheaper opportunities.
• Cache efficiency — Cache-write overspend, cache-read opportunities, cache-efficiency outliers.
• Model efficiency — Over-powered model choices, low-output request bursts, tokens-per-request outliers.

The list is sorted by estimated dollar impact, biggest first. Date-range tabs above the table (All time, 90 days, 30 days, 7 days) scope the list by when each finding was detected. The date range is the only filter — there are no severity or category facets, and each row shows its rule’s category under the rule name. Dollar figures in the $ impact column are normalized so rows are comparable even though findings measure different periods:

• A finding measured over multiple days shows a 30-day rate (≈$X/mo) — the same basis the Overview Anomalies card and the audit report use, so a finding shows one figure everywhere.
• A single-day event shows $X on the day.
• A finding with no stamped measurement period shows its raw figure.
• Advisory findings with no recoverable dollar figure show —.

Rules with a recommended action show it inline on the row (Recommended: <action>), valued as worth ~$X/yr by annualizing that finding’s own figure over its own measured period. Click any row to expand it — you’ll see the human-readable explanation, a Likely driver line on spend-spike findings naming what moved versus the prior period, the attribution (model, key, workspace, or user where available), the pricing basis the agent used, a Confidence score (0–100%) for rules that produce one, and Measured over stating the finding’s measurement period and its raw dollar total. Confidence is separate from severity: a low-confidence high-severity finding is worth double-checking before acting; a high-confidence low-severity finding is usually safe to treat as real.

​

Anomaly report

Click View report above the list to open the anomaly report in a new tab: a printable document you can share with your engineering team. To save the report as a file, click Download PDF next to View report. The button opens the report and brings up your browser’s print dialog so you can save it as a PDF. The report’s print styles keep full colors, so the saved PDF matches the report on screen. Both buttons cover the date-range tab you have selected in the findings explorer. The report opens with a summary and an index table, then groups findings from the same rule into one numbered entry (one story, one action) with each affected scope carrying its own figure. Every entry states its evidence, a Recommended action, and How you’ll know it worked — a verification step for confirming the fix landed. Figures are never totalled across incompatible measurement periods.

​

Spend targets and spend alerts

Spend-overage findings under Usage patterns are gated on the daily, weekly, and monthly spend targets you set in the Configure modal. Targets are measured against aggregate spend across all connected providers.

• A window with a target produces spend alerts when aggregate spend in that window runs over the amount.
• A window with no target gets no spend alerts for that granularity, even if spend rises sharply.

To turn on spend alerts, set a target from Configure. See Set spend targets.

​

Violations

Violations are the disputable findings: billing and pricing discrepancies where the provider’s charge doesn’t match what pricing says it should have been. These findings can back a refund claim in Recovery. Categories covered here:

• Billing — Billing discrepancies (point-in-time and windowed), rate & discount summaries.
• Pricing data — Stale pricing entries, unknown-model observations.

Violations use the same date-range tabs, normalized impact figures, and expand-for-detail interaction as Anomalies. Each finding is categorized into either Anomalies or Violations — a single finding never appears in both.

​

Action Plan

The Action Plan is your prioritized to-do list. It takes every finding — anomalies and violations — and ranks them by estimated dollar impact, biggest first. At the top of the page you’ll see the total identified opportunity: the sum of estimated impact across every open finding. Each item shows the finding title, its severity, the human explanation, and its individual impact. Work top-down to pick up the highest-value fixes first.

​

Recovery

Recovery is the home for refund claims filed against billing violations. See Recovery and refund claims for the full walkthrough.