AI apps connected to Microsoft 365
Review supported application inventory, owners, consent, permissions and credential metadata returned by the customer-approved connection.
AI + COPILOT GOVERNANCE
Find AI applications. Establish access. Ask what is approved.
SCOUTz reviews supported AI-application inventory, ownership, permissions and credential metadata using approved tenant evidence.
EVIDENCE · INSIGHT · PROOF
What you can know before they let you in, and after.
01
Before consent
Published domain, DNS, email authentication, certificates and external evidence.
02
With read-only consent
Supported Microsoft 365 identity, app, permission, licensing and configuration evidence.
03
Never inferred
No message bodies, private files, chats or prompts. Unread is not clean.
EVIDENCE · INSIGHT · PROOF
A real review is a starting point—not a readiness verdict.
01
Observed evidence
Retain the finding, source and date.
02
Customer-confirmed
Record purpose, owner and agreed context.
03
Gap
State what needs review or an approved change.
04
Not assessed
Show unread sources and work not tested.
Application presence is not proof of active use. A permission grant is not proof of misuse. Keep observed evidence, customer-confirmed context, documented gaps and questions not assessed distinct.
This does not provide complete Shadow IT coverage, AI-agent activity or delegation-chain visibility, embedded-agent discovery, or access to private prompts and content.
WHAT'S NEW IN SCOUTz · OPEN BETA
AI Center: Shadow AI and Copilot preparation evidence
Review AI questions for one client or across your client portfolio in one place. AI Center separates connected applications, vendor-documented AI features, identifiable agents, managed-device evidence and Copilot licensing, so a product feature is not mistaken for observed use.
Business apps with AI features
See AI features described in the vendor's own documentation. A vendor feature is context; it does not prove the client enabled or used it.
AI agents with their own sign-in
Review agent identities where the supported source exposes a separate sign-in. This does not reveal actual agent activity, delegation chains or every agent embedded inside other software.
AI apps on managed devices
Use available Intune managed-device application evidence for computers, phones and tablets. Coverage depends on the connected source and permissions; unmanaged devices and missing reads stay outside the conclusion. SCOUTz does not install an endpoint agent.
Copilot seats and readiness review
Keep license assignment, supported configuration and customer-confirmed prerequisites separate. Use observed evidence, customer-confirmed context, gaps and not-assessed questions—not an unsupported readiness score or an automatic “ready” verdict.
Honest by design: every count names its source. A check SCOUTz could not read stays “Not read”—not “none” or “all clear”. Microsoft 365 reviews use client-consented, read-only access and do not read message bodies, files, chats, documents or prompts.
AI DISCOVERY IS NOT ONE THING
Name how the application was observed before naming the risk.
Application presence, entitlement, consent, activity, network traffic, data movement, runtime inspection, and enforcement are different layers.
AI APPLICATIONHow was it observed?
INSTALLED?LICENSED?CONSENTED?SIGNED IN?OAUTH?ENDPOINT?
What does that evidence actually prove?AI APP OBSERVEDpresence or relationship in a supported source≠PROMPT CONTENT OBSERVED
DISCOVERY
Which services or applications appear?
IDENTITY
Who or what is related?
APPLICATIONS
Publisher, owner, assignment, and state.
PERMISSIONS
What access was granted?
USAGE
What supported activity evidence exists?
DATA MOVEMENT
What a separate source can establish.
RUNTIME
Prompt, agent, or MCP inspection under a different boundary.
ENFORCEMENT
Allow, educate, monitor, constrain, replace, or block.
SCOUTz CURRENT FOCUSDiscoveryAssessmentGovernanceNot runtime prompt surveillance or autonomous enforcement.
REAL APP CAPTURE · CLIENT DETAILS REMOVED
Discovery before recommendation.
“AI discovery” can describe several different evidence layers. SCOUTz names the layer, preserves its boundary, and helps the MSP decide what deserves investigation. It does not turn an application name into a predetermined product pitch.
Evidence layerQuestion it can support
Application and OAuth inventoryWhat relationship, publisher, consent, and permission scope are visible?
Licensing and assignmentWho has an entitlement—not whether every assigned user is active?
Available activity or network evidenceWhat supported source records use or traffic within its stated window?
Deeper prompt, agent, or MCP inspectionWhat a separately authorized specialist source can actually observe?
Governance and enforcementWhat should be allowed, reviewed, educated, monitored, replaced, or blocked?
Finding an AI application is not the same as finding unapproved AI use. SCOUTz reviews supported application inventory, ownership, permissions, consent and credential metadata through customer-approved, read-only tenant access. Where an approved source supplies activity evidence, show its coverage and date. Otherwise actual use is not assessed. The customer confirms purpose and approval. This is not complete Shadow IT discovery, AI-agent runtime inspection or access to private prompts and content.
REAL APP CAPTURE · CLIENT DETAILS REMOVED
SCOUTz can help reviewSCOUTz does not claim to read
Enterprise application and publisher contextPrompts or model conversations
Consent and permission scopeEmail or chat message bodies
Owners, assignments, and available activity cluesDocuments, files, or browser history
Credentials, history, and evidence state where availableEvery unsanctioned AI service on every device
We use AI. We do not outsource truth to it.
AI can accelerate research organization, repetitive drafting, summarization, and translation. It does not establish what was observed, decide the product boundary, invent a capability, approve a conclusion, or replace the MSP’s judgment. The source and the responsible person remain visible.
Why authentic thought still wins →Your next prospect meeting is already on the calendar.
Bring dated evidence, clear limits, and useful questions.