- Version
- Download 4
- File Size 408.00 KB
- File Count 1
- Create Date May 26, 2026
- Last Updated May 26, 2026
ITAM and AI Token Consumption
AI token consumption is now the unit of measure ITAM and SAM practitioners cannot afford to misread. In the three weeks before this paper went to press, GitHub Copilot rewrote the multiplier on Claude Opus 4.7 from 7.5× to 27×, OpenAI introduced a new $100/month ChatGPT Pro tier, Anthropic shipped Opus 4.7 with a tokeniser that inflates effective cost by 12–35% without changing the headline rate, and Microsoft 365 E7 went generally available at $99/user/month. None of these required a new contract from the buyer.
This SAM Charter whitepaper compares how the major AI products — ChatGPT, Claude, Gemini, Microsoft 365 Copilot, GitHub Copilot, Cursor and OpenAI Codex — meter and bill consumption as of late May 2026. Written for ITAM, SAM and procurement professionals who must forecast AI cost, govern usage, retire shadow subscriptions and challenge optimistic business cases, it is vendor-agnostic and grounded in published rates.
What you get when you download the paper:
An at-a-glance comparison table covering ten product families across seat, message, credit, token and compute-unit billing models, with the specific consumption risk each one introduces to SAM controls.
A role-by-role recommendation matrix — executive, knowledge worker, software developer (general and heavy agentic), data scientist, researcher, creative, ITAM practitioner, risk and compliance, customer-facing automation — each with token-cost rationale and procurement red flags.
A four-part ITAM governance framework covering entitlement, consumption, optimisation and governance disciplines, with operational practices mapped to existing process kits including HAM, SAM, Cloud and the new AI Process Kit aligned to ISO/IEC 42001.
Procurement red flags that matter most for SAM teams: mid-contract repricing, model-multiplier rewrites, tokeniser changes that act as stealth price hikes, Copilot Studio agent proliferation, and shadow personal subscriptions used for organisational work.
The whitepaper closes by importing lessons from FinOps and cloud cost management into AI licensing, and argues that the most valuable ITAM intervention is making visible a bill that is currently invisible — shadow subscriptions, ungoverned API keys and Copilot Studio agents created outside the seat owner's awareness.
Written by Rory Canavan, ISO 19770-1 WG21 member and founder of SAM Charter, with input from the ITAM Accelerate AI Process Kit.
