Shadow AI Inventory Spreadsheet: Discover Unsanctioned AI Tools Without a CASB

Spreadsheet-first discovery, domain list, risk scoring, and AI-006 / AI-001 integration.

Resource guide · Updated 2026 · 16 min read

Shadow AI adoption often outpaces formal approval: engineering teams on Claude Code, marketing pasting customer data into ChatGPT, support using unapproved summarizers, finance uploading spreadsheets to AI analytics tools.

Traditional playbooks recommend CASB deployment, DLP, and domain blocking. For startups and mid-market organizations, that stack is not always practical on day one — yet governance teams still need a defensible answer to a common audit question: which AI tools are in use, and who approved them?

This guide documents a spreadsheet-first shadow AI discovery method: discovery tactics, an AI domain starter list, a risk-scoring register, and integration with AI-001 and AI-006 — without requiring enterprise-grade tooling to begin.

Legal disclaimer

Operational guidance only. This guide supports shadow AI discovery planning. It is not legal advice. Discovery methods must comply with employment law, privacy notices, and local monitoring regulations. Engage counsel before employee device or communication monitoring.

Definition

Shadow AI inventory is a spreadsheet-based register of AI tools used for work without formal approval. In practice, it’s also described as an unsanctioned AI tools register, an employee AI usage audit, or an AI shadow IT spreadsheet.

What Counts as Shadow AI?

Classify as shadow AI when all four apply:

  1. Uses generative AI, ML, or AI-assisted decision support
  2. Used for work (even occasionally)
  3. Has not completed formal approval
  4. Not on the organization’s Approved AI Tools Register
Usually shadowUsually approved
Personal ChatGPT for work draftsEnterprise ChatGPT with SSO
Personal Claude for codeApproved Claude Team deployment
Unreviewed Merlin/Harpa extensionSecurity-reviewed extension
Dept-paid AI without vendor reviewVendor-reviewed platform + DPA

Include: LLMs, generative tools, user-data training. Borderline: Grammarly (if generative), Smart Compose on confidential docs, Notion AI, M365 Copilot. When in doubt, include it — under-inclusion is the bigger risk.

Quick Self-Assessment

  • We can name every AI tool used for work
  • We discover new AI tools as they emerge
  • We classify tools by risk tier
  • We have an AUP covering AI tools
  • We can show auditors how we manage unsanctioned AI

Why Spreadsheet-First Beats “Wait for CASB”

ApproachTimeCostBest for
CASB / SSE4–12 weeksHigh ($50k+)Enterprise security teams
Network DLP + proxy2–6 weeksMediumMid-market with stack
Spreadsheet discovery<1 day$0Startups, scale-ups, pilots

Augment — do not replace — the spreadsheet when the register exceeds 50 tools, critical findings repeat, or audit pressure increases. Lightweight SSO exports and DNS alerts typically precede full CASB investment.

Common shadow AI findings

  1. Personal ChatGPT for customer comms
  2. Claude / Claude Code for development
  3. Otter, Fireflies, Fathom on customer calls
  4. Notion AI, GrammarlyGO, Jasper in SaaS
  5. Browser extensions (Merlin, Harpa, Monica)
  6. Perplexity / You.com for research
  7. Midjourney / DALL·E for marketing
  8. Excel Copilot, Tableau AI on business data

Shadow AI risk tiering

TierCriteriaResponse
CriticalPII/PHI/customer data; training default; unknown policyEscalate + block + policy update (24h)
HighInternal data retained; opt-out unconfirmedRegister + AUP + quarterly review
MediumPublic data; reputable vendor policyRegister + annual review
LowPersonal use; no work dataOptional entry + awareness

Quick Start: First 5 Tools in 1 Hour

  1. SSO logs (15 min): 90-day export; filter ai, OpenAI, Claude, Copilot
  2. Expenses (15 min): OpenAI, Anthropic, ChatGPT, Claude, Midjourney
  3. Slack/Teams (15 min): search ChatGPT, Claude, Copilot
  4. Poll (15 min): 3-question survey to eng + marketing
  5. Document: add rows to CSV below with risk tier

Discovery Tactics (No CASB)

1. SSO / identity logs

Filter Okta, Azure AD, or Google logins for AI domains. Note: Copilot may appear as GitHub; Notion AI as Notion — cross-check URLs.

SELECT user_login, app_name, event_time FROM login_events WHERE (app_name ILIKE ‘%ai%’ OR app_name ILIKE ‘%copilot%’ OR app_name ILIKE ‘%claude%’) AND event_time > CURRENT_DATE – INTERVAL ’90 days’ ORDER BY event_time DESC;

2. Expense & procurement

Search: OpenAI, Anthropic, Cursor, Midjourney, Jasper, Otter, Fireflies, Notion, Grammarly.

3. DNS / firewall logs

Top domains vs approved list — see domain starter list below.

4. Browser history (small teams)

Managed Chrome/Firefox exports for repeated AI domain visits.

5. Surveys & interviews

Non-punitive: “What AI tools do you use for work?” — engineering, marketing, support, finance.

6. API key scanning

grep -rE “(sk-[A-Za-z0-9]{20,}|sk-proj-|api\\.openai|anthropic)” ./src \ –include=”*.py” –include=”*.js” –include=”*.env”

Pair with prompt firewall / secret detection in CI.

7. Extensions & MDM (optional)

Merlin, Harpa, Monica — endpoint or MDM reports.

AI Domain Discovery Starter List (50+)

Copy into DNS filters, log analysis, or manual review. Update quarterly — new AI tools launch weekly.

# LLM APIs api.openai.com, platform.openai.com, claude.ai, anthropic.com, api.anthropic.com api.cohere.com, api.mistral.ai # Coding cursor.sh, github.com/features/copilot, codeium.com, tabnine.com # Writing jasper.ai, copy.ai, grammarly.com, notion.ai # Meetings otter.ai, fireflies.ai, fathom.video, tl;dv.io # Search perplexity.ai, poe.com, you.com, phind.com # Image midjourney.com, leonardo.ai, runwayml.com # Extensions merlin.fo, harpa.ai, monica.im # Enterprise embedded microsoft.com/copilot, salesforce.com/einstein, zendesk.com/ai, hubspot.com/ai

Shadow AI Register Schema

Core columns (aligns with AI-006):

ColumnPurpose
tool_ide.g. SHADOW-AI-001
tool_nameClaude Code (unapproved)
discovery_methodSSO, expense, survey, API scan
discovered_dateYYYY-MM-DD
primary_userTeam or individual
usage_scopeIndividual / Team / Dept / Org
data_exposure_riskPII, customer data, internal, public, unknown
vendor_data_policyTraining default, opt-out, no training, unknown
opt_out_confirmedYes / No / N/A
risk_tierCritical / High / Medium / Low
remediation_statusBlocked, AUP pending, approved w/ conditions, monitoring
next_review_dateYYYY-MM-DD

Extended: aup_reference, business_justification, approved_alternative, resolution_notes.

Copy-Paste CSV Template

tool_id,tool_name,discovery_method,discovered_date,primary_user,usage_scope,data_exposure_risk,vendor_data_policy,opt_out_confirmed,risk_tier,remediation_status,next_review_date SHADOW-AI-001,Claude Code (unapproved),SSO logs,2024-11-15,engineering-backend,Team-level,Internal code,Opt-out available,No,High,AUP acknowledgment pending,2025-02-15 SHADOW-AI-002,ChatGPT for customer emails,Expense scan,2024-10-22,support-team,Department-wide,Customer PII,Training by default,N/A,Critical,Blocked + policy update,2024-11-22 SHADOW-AI-003,Notion AI for meeting notes,Survey,2024-09-30,product-team,Team-level,Internal notes,No training,N/A,Medium,Approved with conditions,2025-03-30

Risk scoring formula

# Weighted score (customize for organizational risk appetite) =IF(OR(G2=”PII”,G2=”PHI”),3,IF(G2=”Customer data”,2,IF(G2=”Internal docs”,1,IF(G2=”Unknown”,1,0)))) + IF(H2=”Training by default”,3,IF(H2=”Unknown”,2,IF(H2=”Opt-out available”,1,0))) + IF(I2=”No”,2,IF(I2=”Unknown”,1,0)) + IF(F2=”Org-wide”,3,IF(F2=”Department-wide”,2,IF(F2=”Team-level”,1,0))) # 9–11 Critical | 6–8 High | 3–5 Medium | 0–2 Low

AI-006 + AI-001 Integration

Shadow AI Register ↓ Vendor review (AI-011) + privacy (PRI-004) ↓ Block / approve with conditions / sanction ↓ Migrate approved tools → AI-006 System Register

Sample AI-001 language:

§3.2 Unsanctioned AI Tools Employees may not use AI tools for work unless listed in the Approved AI Tools Register OR approved in writing by Security + Legal. For approved tools: opt out of training where available; do not upload PII/PHI/customer data; report new tools within 5 business days.

Run vendor due diligence with the 30-question AI vendor questionnaire. For governed inventory columns, see the ISO 42001 register walkthrough.

Kickoff email (enablement, not punishment)

Subject: Helping You Work Safely with AI Tools We’re running a lightweight discovery initiative — not to restrict productivity, but to support valuable tools safely and protect company/customer data. If you use AI tools for work, reply or complete this 2-minute survey: [link] No one is in trouble. We want to help you work safely with AI.

Escalation triggers

  • PII/PHI/customer data without approval
  • Training default with no opt-out or unconfirmed opt-out
  • No security documentation (SOC 2, privacy policy)
  • Org-wide use without awareness

Program metrics (monthly)

MetricWhy
Total discovered toolsDiscovery effectiveness
Approved vs unapproved ratioEnablement progress
Critical findings (sustained zero)Risk control
Remediation SLA<7d Critical, <30d High
AUP acknowledgment rateCultural adoption

Limitations

Spreadsheet discovery will not reliably catch: BYOD without MDM, personal email accounts, local open-source models, encrypted destinations, or brand-new domains. Document these limitations in the organization’s risk assessment.

Shadow AI Discovery + Governance Toolkit

Spreadsheet register, AUP, and vendor intake from the AI Governance Kit.

  • AI-006 — system register (sanctioned tools)
  • AI-001 — acceptable use policy
  • AI-016 — policy acknowledgments
  • AI-011 — vendor intake for newly discovered tools
Get the AI Governance Toolkit →

FAQ

How do we avoid a “gotcha” culture?
Frame as risk management; use non-punitive surveys; emphasize approved alternatives; remediate high-risk first. Document in AI-001.
How often to run discovery?
Quarterly baseline; monthly in high-growth or regulated contexts. Automate SSO/expense scans where possible.
Personal AI on work devices?
Define in AUP: no work data, separated profiles, no interference with duties. Log as Individual + Low tier if compliant.
Critical tool already org-wide?
Escalate immediately; interim safeguards; fast-track sanction or approved alternative; document risk acceptance timeline.
Non-AI shadow IT?
Same register schema works — expand domain list and risk criteria.

Implementation checklist

  • Define shadow AI scope
  • Pick 2–3 discovery tactics (SSO + survey recommended)
  • Create register + risk formula
  • Run 1-hour quick start sweep
  • Classify and escalate Critical/High per SLA
  • Update AI-001 with §3.2-style clauses
  • Communicate kickoff email to affected teams
  • Configure SSO block / DNS where feasible
  • Quarterly re-discovery + metrics
  • Retain evidence for audits

This spreadsheet method provides a practical baseline for shadow AI discovery when enterprise tooling is not yet in place. Pair the register with AI-001 policy language and AI-006 for sanctioned tools to demonstrate governance maturity to auditors and leadership.