AcmeOps AI margin risk

Catch AI margin risk before it reaches the P&L. The tenants, features, and models putting AI margin at risk this month, with the recommended action surfaced at the top.

Open simulator Open demo walkthrough

Workspace customer

AcmeOps AI

AI workflow software for document-heavy operations teams. This demo analyzes one AI product line, not the full company P&L.

ARR: $18M
Employees: 180
Total AI spend: $85K / mo
Revenue under analysis: $41.5K
AI COGS under analysis: $24.55K
Projected AI COGS after action: $16.98K

Unprofitable tenant

Northstar Logistics

-17.8% gross margin

Risky feature

Contract Analyzer

98.1% of feature cost from unprofitable tenants

Cost driver model

Claude Opus 4.7

100% of model cost from unprofitable tenants

Triggered guardrail

Northstar Logistics

294.4% of monthly AI budget

Recommended action

Routing change

Route 90% of Contract Analyzer traffic from Claude Opus 4.7 to Claude Sonnet 4.6

Monthly revenue

$41,500

Across all active tenants

AI cost (MTD)

$24,545

Sum of seeded usage records

Gross margin

$16,955

40.9% blended

Unprofitable tenants

1 tenant where AI cost exceeds revenue

Triggered guardrails

1 tenant over 120% of budget

Triggered guardrails

Tenants that crossed the 120% AI budget threshold this month.

Critical guardrail at 120%

Northstar Logistics exceeded 120% of monthly AI budget (294% of $5,600).

Recommended action: Route 90% of Contract Analyzer traffic from Claude Opus 4.7 to Claude Sonnet 4.6.

Critical

Northstar Logistics budget used

294.4%

$16,485 spent

against $5,600 allowance

80%100%120%

Healthy under 80%Warning 80% to 100%Budget reached 100% to 120%Critical over 120%

Open simulator Open guardrails

Cost distribution

Where the workspace AI spend lands across features and models this month.

AI cost by feature

Total

$24,545

Contract Analyzer
$16,20066%
Document Q&A
$3,68515%
AI Support Copilot
$1,9457.9%
Research Assistant
$1,4135.8%
Summarization
$1,3035.3%

AI cost by model

Total

$24,545

Claude Opus 4.7
$15,40062.7%
GPT-5.4
$2,97312.1%
Claude Sonnet 4.6
$2,95012%
Llama 3.3 70B
$1,5606.4%
Gemini 2.5 Pro
$1,1634.7%
GPT-5.5
$5002%

Tenant margin risk

Sorted by gross margin. The most unprofitable tenants surface first.

Gross margin by tenant

Northstar Logistics

-17.8%

BrightCart

+67.9%

Atlas Health

+70%

Redwood Finance

+72.3%

Apex Legal

+74.6%

UnprofitableBelow targetHealthy

Tenant	Revenue	AI cost	Gross margin	Margin %	Budget used	Risk status
Northstar Logistics	$14,000	$16,485	-$2,485	-17.8%	294.4%	Unprofitable
Apex Legal	$2,800	$710	+$2,090	+74.6%	84.5%	Warning
BrightCart	$4,200	$1,348	+$2,853	+67.9%	106.9%	Budget reached
Redwood Finance	$6,500	$1,803	+$4,698	+72.3%	92.4%	Warning
Atlas Health	$14,000	$4,200	+$9,800	+70%	75%	Healthy

Feature margin risk

A feature is risky when it has material spend and a large share of that spend comes from unprofitable tenants.

Company AI cost by feature

Contract Analyzer

$16,200

Document Q&A

$3,685

AI Support Copilot

$1,945

Research Assistant

$1,413

Summarization

$1,303

Feature	AI cost	Share of company AI cost	Cost from unprofitable tenants	Status
Contract Analyzer	$16,200	66%	98.1%	Risky
Document Q&A	$3,685	15%	6.8%	Healthy
AI Support Copilot	$1,945	7.9%	6.7%	Healthy
Research Assistant	$1,413	5.8%	0%	Healthy
Summarization	$1,303	5.3%	15.7%	Healthy

Model margin risk

A model is risky when it has material spend and at least 40% of its cost comes from unprofitable tenants.

Model	Provider	AI cost	Share of company AI cost	Cost from unprofitable tenants	Status
Claude Opus 4.7	Anthropic	$15,400	62.7%	100%	Risky
GPT-5.4	OpenAI	$2,973	12.1%	6.9%	Healthy
Claude Sonnet 4.6	Anthropic	$2,950	12%	8.5%	Healthy
Llama 3.3 70B	Meta	$1,560	6.4%	8.3%	Healthy
Gemini 2.5 Pro	Google	$1,163	4.7%	0%	Healthy
GPT-5.5	OpenAI	$500	2%	100%	Low spend

Models below the materiality threshold are not marked risky even if their spend is concentrated in an unprofitable tenant.