← P402 Meter
Enterprise · AI Cost PlatformSynthetic Demo Live
> _ P402 METER · ENTERPRISE · AI COST MANAGEMENT PLATFORM

AI spend without
attribution
is just noise.

By 2026, every engineering team, marketing team, legal team, and finance team is using AI. The monthly invoice from OpenAI, Anthropic, and Google is growing. Nobody can tell you which department drove it, which project it belongs to, or whether the model tier used was even necessary.

P402 Enterprise is the metering layer that makes AI spend auditable — down to the employee, the session, the token, and the model tier choice. And it generates optimization recommendations that cut 30–70% of model costs by routing tasks to the cheapest tier that meets the quality bar.

30–70% model cost reduction via routing optimizationPer-employee · per-project · per-client attributionSOC2-ready onchain audit trail
01The Problem

Enterprise AI spend is growing and completely unattributed.

$0
Per-employee AI cost visibility at most enterprises
OpenAI, Anthropic, and Google bill the org. The bill is opaque. No department breakdown, no per-project attribution, no per-employee cost center. Finance cannot charge back AI costs. Engineering cannot optimize model tier choices without guessing.
30–70%
Model cost waste from misrouted requests
Engineers use Claude Opus for simple code completion. Marketing uses GPT-4o for tweet drafts. Neither task needs the premium tier. The routing decision is invisible so it defaults to the best model available. The waste compounds at scale.
$0
Audit trail for enterprise AI usage
When a compliance audit asks 'what AI did this employee use to produce this output?' the answer is nothing. No session log, no model log, no cost log, no receipt. SOC2, ISO 27001, and internal legal reviews increasingly require this.
Unknown
End-of-month AI spend projection for any department
Day 10 of the month. Engineering is at $45 of a $200 budget. Is that on track? Ahead of pace? Behind? Without per-day velocity tracking and a projection model, nobody knows until the invoice arrives.
02Attribution Hierarchy

Every token attributed to seven levels.

P402 Enterprise maintains a full attribution chain from the organization down to the individual token. Every LLM call is tagged at ingestion and attributed at billing.

Organization
Acme CorpTop-level budget owner. Gets the consolidated invoice.
Department
Engineering · Marketing · Legal · FinanceCost center. Each has its own budget cap and model routing policy.
Project
Project Phoenix · Series B DiligenceBudget sub-cap per active project. Chargeable to a client.
Client
Client A · InternalOptional passthrough. Law firms and agencies charge clients for AI compute.
Employee
Alice ChenIndividual attribution. Who drove this cost? Leaderboard + anomaly detection.
Session
sess_abc123A bounded unit of work. Maps to a task, document review, or conversation.
Request → Token
12,847 tok · claude-opus-4-5 · $0.097The atomic unit. Model chosen, tokens used, cost settled on Tempo.
03Platform Capabilities

Meter. Trace. Track. Receipt. Audit. Route. Optimize.

Meter

Every LLM call priced at token granularity. Cost calculated at request time, not billing time. Model tier, token count, and USD cost recorded per request.

Trace

Full request trace: which employee, which session, which model, which prompt pattern, what output, what cost. Traceable from the monthly invoice back to the individual keystroke.

Track

Real-time spend tracking across the hierarchy. Department dashboards, project burn rates, employee leaderboards. Budget consumption visible live, not at month-end.

Receipt

Immutable session receipts: model used, tokens consumed, routing decision rationale, cost, timestamp, Tempo settlement hash. Exportable as JSON or PDF.

Audit

Every session produces an audit artifact: who did what with which AI, when, at what cost, with what output. SOC2, ISO 27001, and internal legal review ready.

Route

Per-department routing policies. Engineering: quality-first for complex tasks. Marketing: cost-first for content generation. Legal: compliance-aware, minimum Claude Sonnet. Each policy enforced per request.

Optimize

Gemini Pro analyzes 30 days of routing history and task-type similarity scores. Identifies where premium models are used on tasks that economy models handle equally well. Quantifies the savings before you act.

Project

Budget projection engine: current velocity × remaining days = end-of-period forecast. Alert thresholds at 80% and 95% budget consumption. Projected overage surfaced 2 weeks early.

04Routing Optimization

The 30–70% cost reduction is real and measurable.

The optimization engine compares output quality scores across model tiers for each task type in your org. Where quality is equivalent, it routes to the cheaper model. The savings are projected before you commit to any routing change.

Task TypeCurrent ModelRecommendedMonthly Saving
Code completion (simple)claude-opus-4-5claude-haiku-4-5
$2.1294%
Content generationgpt-4oclaude-sonnet-4-5
$1.4389%
Financial Q&Agemini-2.0-progemini-2.0-flash
$0.8791%
Contract analysisclaude-sonnet-4-5maintain
97%
Code review (complex)claude-opus-4-5maintain
88%
Actionable savings (3 task types)$4.42/month total
05Budget Controls

Hard caps at every level in the hierarchy.

Department budget cap

Each department gets a monthly budget in USD. When projected spend (current velocity × remaining days) exceeds 90% of the cap, the department head is notified. When the cap is reached, requests are blocked or downgraded to economy tier — configurable per policy.

Project budget cap

Individual projects get their own sub-cap within the department budget. A client engagement can be capped at $50 regardless of the department's remaining budget. The cap is enforced at the session level before any LLM call is made.

Model tier policy

Each department can set a minimum and maximum model tier. Legal: minimum claude-sonnet-4-5 (no economy models for contract work). Marketing: maximum claude-sonnet-4-5 (no premium models for copy). Enforced at routing time, not billed after the fact.

Anomaly detection

Session costs 3× above the employee's 30-day average are flagged automatically. An employee who suddenly runs $10 of LLM calls in one session triggers a Sentinel review. The session is not blocked — it is surfaced for review within 60 seconds.

06Audit Trail

Every session is an immutable receipt.

The audit requirement is different in every industry. But the underlying need is the same: prove what the AI did, when, for whom, at what cost, with what output.

Session receipt — example
Receipt IDrcpt_2c4f8a1b...
Session IDsess_a7c3e9d2...
EmployeeAlice Chen
DepartmentEngineering
ProjectProject Phoenix
ClientInternal
Timestamp2026-05-05T14:32:11Z
Modelclaude-opus-4-5
Tokens (in)2,847
Tokens (out)10,000
Total tokens12,847
Cost USD$0.09635
Routing rationalecomplexity=7, dept_policy=quality-first
Tempo tx hash0x7a3f...c91e
Block14,829,441
Human reviewapproved by Alice Chen
Receipt is immutable. Settlement hash links to Tempo block explorer. Exportable as JSON, PDF, or CSV batch.
SOC2 Type IIAccess logs, cost records, and output artifacts satisfy CC6.1 (logical access) and CC7.2 (system monitoring) control requirements.
Financial controlsPer-department budget caps with blocking enforcement satisfy internal IT general control requirements for AI expenditure authorization.
Legal holdOnchain settlement hashes provide tamper-evident records. Session receipts can be produced in discovery or regulatory response without relying on vendor-controlled logs.
07Unit Economics

The savings are structural, not marginal.

Opaque
Status quo AI billing
Monthly invoice. No department breakdown. No per-project attribution. No routing visibility. No optimization lever. Costs compound invisibly.
30–70%
Model cost reduction via routing
Routing code completion to Haiku instead of Opus saves 87% on that task type. Across an engineering team of 10, that's $25–100/month per employee recovered.
Full trail
Per-token attribution + receipt
Every dollar attributed. Every session receipted. Every optimization quantified before action. Finance can close the AI cost center monthly with confidence.
Acme Corp · synthetic 30-day projection (100-person org)
Without P402Monthly OpenAI invoice: ~$2,400Zero attribution. Zero optimization lever.
With P402 (tracking only)Same $2,400 spend, now attributedDepartment breakdown. Per-project chargeback.
With P402 (+ routing optimization)Projected spend: ~$1,08055% cost reduction via model-tier routing.
Try the demo dashboard now

The dashboard is live with synthetic data.

Org KPIs, department breakdown, employee leaderboard, model mix, budget projections, routing optimization panel, and session log — all wired and running with Acme Corp synthetic data. Connect your P402 API key to see real org spend.

P402 Meter · Enterprise · Tempo Mainnet · MPPTempo × MPP × Gemini 3.1