AI Policy Profiles

Predefined policy rules for every field

Choosing allow, block or controls for 18 categories and 165 subcategories across 18,548+ AI tools can be daunting. Every profile below gives a recommended verdict for every AI tool domain, so a Shadow AI Audit finding tells you not just what was found, but what to do about it.

15profiles
165subcategory rules
18,548domains with a verdict
13risk flags checked
Run a Free Audit Pricing
Choose a profile

Each profile assigns a Block, Allow with controls, or Allow verdict

Based on category, subcategory, risk flags and the vendor's training terms. Every profile is included with every Shadow AI Audit report.

Choose a profile. The counts, category defaults and sample rules below change with it.
Business MSP K-12 Universities Healthcare Banking Insurance Legal Government Defense Accounting Manufacturing Software Agencies Pharma
Showing profile

Defense & critical infrastructure

Defense contractors, CMMC and ITAR environments, utilities and operators of critical infrastructure.

Controlled information never touches a public model. Deny by default, allow by exception, log everything.

Read the related guide →

12,038domains to block
6,508allow with controls
2allow
Agents & Automation
Defense default: Block
AI workflow automation: Block (6260) General autonomous agents: Block (4390) Personal AI assistants: Block (3363) RPA with AI: Block (2463) Agent-building frameworks: Block (1430) MCP servers & integrations: Block (1346) +2 more subcategories
Audio, Voice & Music
Defense default: Block
Text-to-speech: Block (694) Speech-to-text & transcription: Block (496) Real-time voice agents & phone AI: Block (367) Music generation: Block (321) Podcast tools: Block (311) Voice cloning: Block (288) +4 more subcategories
Business & Vertical
Defense default: Block
Finance & accounting: Block (1071) E-commerce: Block (971) HR & recruiting: Block (677) Healthcare & med-tech: Block (490) Government & public sector: Block (314) Legal: Block (233) +6 more subcategories
Code & Development
Defense default: Block
Code assistants & autocomplete: Allow with controls (1049) Documentation: Allow with controls (1030) Autonomous coding agents: Block (863) API dev assistants: Block (669) Debugging & error monitoring: Block (481) Testing & QA: Block (463) +6 more subcategories
Customer Support
Defense default: Block
Support chatbots & AI helpdesk: Block (1217) Conversational commerce & messaging bots: Block (671) Ticket routing & agent-assist: Block (533) Voice & call-center agents: Block (393) Review management: Block (64)
Data, Analytics & Research
Defense default: Block
Market & competitive intelligence: Block (2977) Predictive analytics & forecasting: Block (2370) BI & natural-language analytics: Block (1390) Document data extraction: Block (1258) Data cleaning & prep: Block (950) Survey & feedback analysis: Block (860) +4 more subcategories
Design, 3D & Creative
Defense default: Block
UI/UX design assistants: Block (1337) AR/VR content: Block (683) Game-dev AI: Block (609) 3D model & asset generation: Block (603) Fashion design & virtual try-on: Block (306) Font & typography: Block (255) +1 more subcategories
Education & Learning
Defense default: Allow with controls
AI tutors & personalized learning: Allow with controls (1045) Course & curriculum creation: Allow with controls (454) Quiz, flashcard & study: Allow with controls (369) Grading & feedback: Allow with controls (264) Kids' learning & STEM: Allow with controls (234) Language learning: Allow with controls (204) +1 more subcategories
Image & Visual
Defense default: Block
Text-to-image: Block (1515) Graphic design & layout: Block (1027) Image editing & inpainting: Block (865) Illustration & character: Block (586) Avatars & headshots: Block (486) Background removal: Block (288) +7 more subcategories
Lifestyle & Entertainment
Defense default: Block
AI companions & character chat: Block (557) Gaming & AI-native games: Block (295) Dating & relationships: Block (241) Mental wellness & journaling: Block (220) Astrology & fun generators: Block (199) Health, fitness & nutrition: Block (120) +5 more subcategories
Marketing, Sales & SEO
Defense default: Block
Product marketing: Block (1857) SEO & GEO optimization: Block (1206) Sales prospecting & SDR agents: Block (1096) Ad creative & optimization: Block (1038) Social media management: Block (1026) Email marketing & personalization: Block (1006) +4 more subcategories
Models & Infrastructure
Defense default: Block
Hosting & inference platforms: Block (620) Foundation models & model APIs: Block (523) Fine-tuning & training: Block (436) LLM gateways & routing: Block (326) Observability & evals: Block (248) Local runtimes: Allow with controls (238) +6 more subcategories
Productivity & Collab
Defense default: Block
Meeting assistants: Block (2698) Document drafting & office copilots: Block (2248) Task & project management AI: Block (2134) Email management & triage: Block (587) Calendar & scheduling: Block (308) Presentation & slide generation: Block (278) +2 more subcategories
Robotics & Embodied AI
Defense default: Block
Robotics platforms & foundation models: Block (58) Computer-vision & edge vision: Block (30) Wearables & AI hardware: Block (24) IoT & smart-home AI: Block (17) Autonomous vehicles & drones: Block (16)
Search, Knowledge & Docs
Defense default: Block
AI search & answer engines: Block (433) Enterprise search & KM: Block (216) RAG platforms & retrieval infra: Block (177) Chat-with-documents & PDF: Block (105) Browser copilots & reading assistants: Block (50) Note-taking & second brain: Block (28)
Security & Detection
Defense default: Allow with controls
AI for cybersecurity: Allow with controls (315) Identity verification: Allow with controls (144) Fraud detection: Allow with controls (135) Privacy & PII redaction: Allow with controls (89) AI-content detectors: Allow (21) Deepfake detection & forensics: Allow (20) +1 more subcategories
Text & Language
Defense default: Block
General assistants & chatbots: Block (3358) Writing & long-form: Block (1752) Copywriting & marketing copy: Block (1397) Summarization: Block (977) Translation & localization: Block (974) Grammar, style & proofreading: Allow with controls (813) +6 more subcategories
Video
Defense default: Block
Text-to-video: Block (1107) AI video editing: Block (840) Animation & motion graphics: Block (529) Dubbing, lip-sync & translation: Block (440) Short-form clip repurposing: Block (440) Talking avatars & presenters: Block (281) +4 more subcategories
Block keep it off the network Allow with controls enterprise or SSO tier, no-training terms, logging Allow fine on default settings

Apply this profile to your own network

Run a free Shadow AI Audit and every tool it finds gets a verdict under the profile you pick.

Run a Free AuditSee plans
FAQ

Questions before you deploy

What is an AI policy profile?

A recommended action for every AI tool: Block, Allow with controls, or Allow. One profile per audience, from MSPs to healthcare, legal and defense.

How is the verdict decided?

From fields every AI tool domain already carries: category and subcategory, risk rating, data-sovereignty exposure, abusive-purpose flag, and the vendor's stated position on training with user data.

How does this connect to my Shadow AI Audit?

Every tool your audit finds already carries these same fields. Pick a profile and the report shows how many findings would be blocked, controlled or allowed under it.

My sector is not listed

Start from General Business. Contact us if you need a profile for your specific sector.