Assessment · Governance · Assess and know

6Data sensitivity, privacy and residency

Can sensitive data reach a model it should not, or leave the region?

What data each AI system can reach, where it is processed and on what terms.

AutomatedStakeholder interviews 4 controls 3 engines

How we collect

  • AutomatedSharePoint access reports: what Copilot and agents can reachRuns today
  • AutomatedSensitivity labels and SharePoint sites from GraphRuns today
  • AutomatedVendor privacy, DPA and trust pages captured with the exact quoteRuns today
  • Stakeholder interviewsPrivacy and data owners: data classes, authoritative sources, what AI may readRuns today

Status labels show what runs today, what needs an export or setup from you, and what is planned.

What you get

  • Copilot readiness: overshared sites and content
  • Data class per system and per workflow step
  • Vendor terms graded: published only, contracted or configured

Questions we ask

  • Which systems hold patient, card or legal data?
  • Which regions are allowed?

We can start from

Data classification, DPIAs, records of processing

Data sources

SourceWhat we readStatus
SharePoint access reportsData access governance reports showing which sites and files are shared widely, the content Copilot and agents can reach.Partly automated
Microsoft Graph and Entra IDService principals, OAuth grants, app registrations, sign-ins to AI apps, Copilot usage reports. Read-only scopes, metadata only.Runs today
QuestionnaireThe short interview for what data cannot show: purpose, owners, decisions AI informs, regulations in scope.Runs today
AI gateway telemetryLiteLLM, Cloudflare AI Gateway or AICost CostWall logs: cost per team, workload, agent and outcome.Partly automated

Built on governance engine modules

P14 PlannedPrivacy, Confidentiality & IP

Whose data trains what, and who owns the output?

Assess & Govern
P5 PlannedData & Knowledge Readiness

Is your data ready for the AI you want to run on it?

Architecture & Build
P8 BetaCopilot Readiness & Value

Is your tenant safe for Copilot, and are the seats earning?

Value & Optimization

AICost governance controls in this module

AICost-GovControl301Partly automated
Data sensitivity behind each AI system
What it is

Which data each AI system can read, retrieve or send, how sensitive it is, where it came from and who owns it.

Why it is needed

Most AI exposure comes from the data it can reach, especially copilots and retrieval over shared drives with loose permissions.

Why enterprises need it

Before rolling out a copilot or RAG assistant, teams need to fix oversharing so AI does not surface what people should never see.

What you get:Oversharing fixed before rolloutSensitive data kept out of promptsClear data owners per AI use
Run by
AI Data and Knowledge Readiness engineKnowledge Search (RAG) engineRe-indexing Cost engine
Data
SharePoint access reports, Microsoft Graph and Entra ID, Questionnaire
Evidence produced
Data exposure table: system, data source, sensitivity, sharing scope, owner, action.
Maps to
NIST AI RMF: Map; ISO/IEC 42001: data for AI systems; GDPR
AICost-GovControl302Partly automated
Privacy duties when using AI
What it is

Lawful basis, minimization, retention, cross-border transfer and explanation duties for personal data used by AI.

Why it is needed

Privacy law applies to prompts, training data and outputs just as it does to databases.

Why enterprises need it

Customers and regulators ask where personal data goes when an AI vendor processes it, and for how long.

What you get:Vendor retention and training settings checkedTransfers documentedExplanation duties met where they apply
Run by
Privacy Decision engineVendor Facts engine (ToolsInfo)
Data
Questionnaire, Microsoft Graph and Entra ID
Evidence produced
Privacy decision table: system, personal data, lawful basis, residency, retention, vendor training setting, decision.
Maps to
GDPR; CCPA/CPRA; EU AI Act: data governance
AICost-GovControl303Wave 2
Rights to training data and AI output
What it is

Rights to the data used to train or tune models, copyright exposure, and who owns what AI produces.

Why it is needed

Using data you have no right to, or shipping AI output you cannot own, creates legal and commercial risk.

Why enterprises need it

Marketing, product and legal teams need clear rules before AI content goes to customers.

What you get:Training data rights checkedContract terms for AI outputFewer IP disputes
Run by
Vendor Facts engine (ToolsInfo)
Data
Questionnaire
Evidence produced
IP review: system, training or tuning data, rights basis, output ownership terms, decision.
Maps to
EU AI Act: GPAI copyright duties; Copyright law
AICost-GovControl304Runs today
Data residency and model origin
What it is

Where each model runs and where its provider is based, matched to your residency and sovereignty rules.

Why it is needed

Some data cannot leave a region, and some buyers will not accept models from certain jurisdictions.

Why enterprises need it

Public sector, health and finance buyers ask for this in every security review.

What you get:Allowed regions per data classModel origin risk knownApproved routes for sensitive work
Run by
Privacy Decision engineAI Architecture Decision engineModel Origin Risk engineLocal or Cloud engine
Data
Questionnaire, AI gateway telemetry
Evidence produced
Residency table: system, data class, model, provider country, processing region, allowed or blocked.
Maps to
GDPR transfers; Sector sovereignty rules

Frameworks

GDPR and CCPA/CPRALawful basis, minimization, transfers, retention, explanation duties

Stakeholders

Privacy officeCISOData owners
← 5. Regulatory applicability All modules 7. Identity, access and agent permissions →
📖 Data sources & methodology 172 text models · 9 embeddings · 44 vision · 62 audio · 8 vector DBs across 10 vendor pages · last verified 2026-10-18

Methodology

  • All prices are USD per 1 million tokens, current as of 2026-10-18.
  • Vendor-published values have no mark. Inferred/extrapolated values are marked with * and listed below.
  • Batch API discounts are 50% off standard rates across providers that offer Batch mode.
  • Prompt caching discounts vary by provider (typically 80-90% off cached input tokens).
  • Regional data-residency surcharges (Anthropic 1.1x, OpenAI 1.1x, Google regional tiers) are NOT included in base rates.
  • Long-context pricing tiers apply when input exceeds model threshold.
  • Embedding prices are input-only (no output tokens generated).

Primary sources

Last-verified date is the most recent successful daily snapshot (aicost_pricing_snapshots) or, when no snapshot exists yet, the latest successful crawler run (aicost_crawler_runs). 10 of 10 vendors are currently verified. Aggregator services (TokenCost, AI Pricing Guru, etc.) are not listed.

Anthropic
2026-10-18
https://www.anthropic.com/pricing
Daily snapshot since Sep 2023 · 704 days captured
Anthropic Docs
2026-10-18
https://platform.claude.com/docs/en/about-claude/pricing
Daily snapshot since Sep 2023 · 704 days captured
OpenAI
2026-10-18
https://openai.com/api/pricing/
Daily snapshot since Sep 2023 · 705 days captured
Google AI
2026-10-18
https://ai.google.dev/gemini-api/docs/pricing
Daily snapshot since Dec 2023 · 680 days captured
Google Vertex
2026-10-18
https://cloud.google.com/vertex-ai/generative-ai/pricing
Daily snapshot since Dec 2023 · 680 days captured
DeepSeek
2026-10-18
https://api-docs.deepseek.com/quick_start/pricing
Daily snapshot since May 2024 · 619 days captured
xAI
2026-10-18
https://x.ai/api
Daily snapshot since Nov 2024 · 537 days captured
Mistral
2026-10-18
https://mistral.ai/pricing
Daily snapshot since Dec 2023 · 678 days captured
Cohere
2026-10-18
https://cohere.com/pricing
Daily snapshot since Sep 2023 · 704 days captured
Voyage AI
2026-10-18
https://docs.voyageai.com/docs/pricing

Inferred values (marked with * in calculator tables)

Derived from industry conventions, not directly published by the vendor. Typical conventions: cached input = 10% of base (90% off), Batch API = 50% of base (50% off).

Vendor / Model Field Why it’s inferred
Anthropic — Claude Sonnet 4.6 cachedInput Derived at 10% of input rate — Anthropic publishes 90% cache-hit discount on this tier.
Anthropic — Claude Sonnet 4.5 cachedInput Derived at 10% of input rate; same 90% cache-hit convention as Sonnet 4.6.
Anthropic — Claude Sonnet 4.5 batchInput Derived at 50% of standard input — Anthropic documents uniform 50% Batch discount.
Anthropic — Claude Sonnet 4.5 batchOutput Derived at 50% of standard output — Anthropic documents uniform 50% Batch discount.
Anthropic — Claude Haiku 4.5 cachedInput Derived at 10% of input rate — Anthropic 90% cache-hit discount convention.
OpenAI — GPT-5.4 Mini cachedInput Derived at 10% of input — OpenAI documents automatic 90% discount on cache hits across GPT-5.x tier.
OpenAI — GPT-5.4 Nano cachedInput Derived at 10% of input — OpenAI 90% cache-hit convention.
OpenAI — GPT-5.4 Nano batchInput Derived at 50% of input — OpenAI Batch API uniform 50% discount.
OpenAI — GPT-5.4 Nano batchOutput Derived at 50% of output — OpenAI Batch API uniform 50% discount.
OpenAI — GPT-5.4 Pro cachedInput Derived at 10% of input — OpenAI 90% cache-hit convention.
OpenAI — GPT-5.4 Pro batchInput Derived at 50% of input — OpenAI Batch API uniform 50% discount.
OpenAI — GPT-5.4 Pro batchOutput Derived at 50% of output — OpenAI Batch API uniform 50% discount.
OpenAI — GPT-5.2 cachedInput Derived at 10% of input; no residency uplift.
OpenAI — GPT-5.2 batchInput Derived at 50% of input.
OpenAI — GPT-5.2 batchOutput Derived at 50% of output.
OpenAI — GPT-5 cachedInput Derived at 10% of input.
OpenAI — GPT-5 batchInput Derived at 50% of input.
OpenAI — GPT-5 batchOutput Derived at 50% of output.
OpenAI — GPT-5.5 Pro cachedInput Derived at 10% of input — OpenAI does not publish a cached rate for *-pro models; using the family convention.
OpenAI — GPT-5.5 Pro batchInput Derived at 50% of input.
OpenAI — GPT-5.5 Pro batchOutput Derived at 50% of output.
OpenAI — GPT-5.2 Pro cachedInput Derived at 10% of input — pro-tier convention.
OpenAI — GPT-5.2 Pro batchInput Derived at 50% of input.
OpenAI — GPT-5.2 Pro batchOutput Derived at 50% of output.
OpenAI — GPT-5.1 batchInput Derived at 50% of input.
OpenAI — GPT-5.1 batchOutput Derived at 50% of output.
OpenAI — GPT-5 Pro batchInput Derived at 50% of input.
OpenAI — GPT-5 Pro batchOutput Derived at 50% of output.
OpenAI — GPT-5 Nano cachedInput Derived at 10% of input.
OpenAI — GPT-5 Nano batchInput Derived at 50% of input.
OpenAI — GPT-5 Nano batchOutput Derived at 50% of output.
Google — Gemini 3 Flash cachedInput Derived at 10% of input — Google caching discount convention ~90%.
Google — Gemini 3.1 Flash-Lite cachedInput Derived at 10% of input — Google caching convention.
Google — Gemini 3.1 Flash-Lite batchInput Derived at 50% of input — Google Batch API uniform 50% discount.
Google — Gemini 3.1 Flash-Lite batchOutput Derived at 50% of output — Google Batch API uniform 50% discount.
Google — Gemini 2.5 Pro cachedInput Derived at 10% of input.
Google — Gemini 2.5 Flash cachedInput Derived at 10% of input.
Google — Gemini 2.5 Flash-Lite cachedInput Derived at 10% of input — Google caching convention.
Google — Gemini 2.5 Flash-Lite batchInput Derived at 50% of input — Google Batch API uniform 50% discount.
Google — Gemini 2.5 Flash-Lite batchOutput Derived at 50% of output — Google Batch API uniform 50% discount.
xAI — Grok 4 (legacy) cachedInput Extrapolated at 25% of base.

Pricing is cross-verified against the LiteLLM community registry when available. Daily snapshots are kept in aicost_pricing_snapshots; every change is logged to aicost_price_changelog with old & new values for full audit trail. Read the full methodology →