GHOST TAX
GHOST TAX
INTELLIGENCE BENCHMARK
How ungoverned AI adoption creates financial exposure. Detection patterns, cost impact data, and practical governance frameworks.
Multiple AI assistants (ChatGPT, Copilot, Claude, Gemini) active simultaneously across teams with overlapping capabilities.
Detection: Public tech stack analysis reveals 2+ AI tools with overlapping use cases. | Frequency: Very common
Individual teams adopting AI tools via personal credit cards or team-level subscriptions outside procurement.
Detection: Hiring signals + tool mentions in job postings indicate rapid, ungoverned adoption. | Frequency: Common in high-growth companies
API-based AI usage (OpenAI, Anthropic APIs) without consumption governance, leading to unpredictable and growing costs.
Detection: Multiple API provider references in public documentation or engineering blogs. | Frequency: Growing rapidly
Using expensive frontier models (GPT-4, Claude Opus) for tasks that smaller, cheaper models handle equally well.
Detection: Inferred from AI tool footprint size and engineering team growth signals. | Frequency: Nearly universal
Procurement Gate
Require approval for AI tool purchases above 50 EUR/mo per user. This catches 80% of shadow subscriptions without creating excessive friction.
Approved Tool Registry
Maintain a curated list of sanctioned AI tools by category (code assistance, content, analytics, automation). Teams pick from the list instead of discovering independently.
Quarterly Utilization Review
Review active AI subscriptions quarterly. Identify tools with <30% utilization or >50% capability overlap with another tool. Consolidate or cancel.
Cost-Per-Outcome Tracking
Shift from cost-per-seat to cost-per-outcome. A 20 USD/seat AI tool that saves 4 hours/week is well-governed. A 20 USD/seat tool with 10% adoption is shadow waste.
Shadow AI refers to AI tools and services adopted by employees or teams without formal IT procurement approval. This includes personal ChatGPT subscriptions, unapproved API integrations, and team-level AI tool purchases that bypass governance workflows.
For a 200-person technology company, shadow AI typically adds 2,000–18,000 EUR/month in unmanaged spend. The cost is compounded by redundancy — multiple teams often pay for overlapping AI capabilities independently.
Ghost Tax detects shadow AI through public signal analysis: job postings mentioning AI tools, tech stack disclosures, hiring velocity patterns, and industry-calibrated heuristics. No internal system access is required at the detection stage.
Effective shadow AI governance requires three components: (1) a lightweight procurement gate for AI tool purchases above a threshold, (2) an approved AI tool registry with clear category coverage, and (3) periodic utilization review to identify and consolidate redundant subscriptions.
Related research
Detect your shadow AI exposure
Enter your domain. The system detects AI tool redundancy and governance gaps automatically.
RUN DETECTION