DepthData

About DepthData

DepthData is a system of record for company-wide AI spend. It connects multiple third-party AI tools into a single, audit-ready view that tracks adoption, usage, and cost. The tool enforces a strict rule: never display a figure it cannot verify from a vendor's own API, and it labels every number by its data source.

Review

DepthData addresses a specific pain point for organizations that have adopted several AI subscriptions but lack a consolidated picture of spending and seat utilization. The tool pulls billing and usage data directly from provider APIs where available, and it supplements gaps with user-supplied contract pricing. Its core differentiator is a verification label that travels with every number, so the provenance of each figure remains clear even months later.

Key Features

• Connects to AI vendor APIs (Anthropic, OpenAI, Cursor, and others) to pull actual dollar spend, usage counts, and seat data, then labels each figure as MEASURED, ALLOCATED, or MODELED depending on how it was obtained.
• Reads only metadata - seats, usage, and spend - and never accesses prompt content, keeping the system narrow in scope for security reviews.
• Displays cost per person, not just per seat, so teams can spot concentration on usage-priced tools where one heavy user may outspend the rest.
• Surfaces overlapping tool adoption: identifies people who hold active seats in two same-category tools and shows the cost of the second one, without making a judgment about which to keep.
• Includes an expense reconciliation section that matches vendor names from a finance CSV export against connected tools, showing unmatched spend as named gaps. The matching runs in the browser without uploading the file.

Pricing and Value

Pricing has not been publicly disclosed. The tool's value stems from its verification labeling and gap visibility, which help finance and IT teams trust the numbers they're reporting and avoid treating estimates as facts.

Pros

• Every number carries a source label (MEASURED, ALLOCATED, etc.) that persists through exports and reports, reducing the risk of mistaking an estimate for a hard figure.
• The metadata-only approach avoids touching prompt content, which simplifies security approval and limits exposure.
• Gaps are explicitly shown rather than hidden behind estimates, so teams know exactly what the APIs can and cannot report.
• Per-user cost breakdowns make it harder to overlook a single individual driving disproportionate spend on usage-based tools.
• The expense reconciliation feature allows a CSV upload to match shadow spend against known tools, all processed locally in the browser.

Cons

• Session-level or per-run cost tracing is not available because vendor reporting APIs don't expose that granularity; the tool only goes as deep as daily aggregates or minute buckets where supported.
• It is not well suited for organizations that need to audit prompt content or analyze the substance of AI interactions, since DepthData deliberately avoids reading prompts.
• No direct integration with SSO or identity provider logs (e.g., Okta, Entra) exists to cross-check actual login activity against vendor-reported usage, so the data relies solely on what each AI tool's API exposes.

DepthData fits finance and IT teams that require an auditable, cross-vendor overview of AI spend and seat utilization without touching sensitive prompt data. It is less useful for teams that need per-session cost attribution or content-level auditing, as those capabilities fall outside its scope. The tool's emphasis on labeling gaps rather than guessing over them makes it a practical choice for organizations where trust in the numbers is the primary requirement.

Open 'DepthData' Website