AI Evaluator secondary / Enterprise operations manager
Enterprise-approved AI / Copyable install pack / 8:55 PM
Approved AI configurator
Copyable install pack / Company-data safe
archetype, role, workflow, and review prompts
profile translated 8:37 PM
Policy lockoutNo company data leaves the approved system.
AIdentity can prepare the profile, prompts, and setup language. It does not ingest confidential work data or bypass enterprise controls.
AI Operator with AI Evaluator review habits
36 prompts plus 10 gates
admin-only actions are drafted, not executed
After-hours guardQueue improvements without crossing company data boundaries.
The configurator holds sensitive ideas until the user can check policy, manager expectations, or IT approval.
Real customer, employee, financial, or strategy examples do not enter AIdentity.
Prompts are rewritten with placeholders before use in the approved system.
The next day starts with the safest prompts and the clearest review checklist.
AIdentity stores the profile and prompt pack, not company work data.
Current setup state / Policy-safe backlog
Install pack waiting now
36 prompts are organized for AI Operator adoption habits. 10 gateskeep policy-sensitive work human-reviewed. Admin-only changes are drafted, not executed.
ProfileReady
Prompts36
Tools6
Gates10
Assessment to approved-tool install mapAI Operator / 36 prompts / Human approval hold
ASSESS
PROFILE
PROMPTS
TOOL
REVIEW
ID01
ID02
TL01
TL02
TL03
TL04
TL05
TL06
PR01
PR02
PR03
PR04
PR05
PR06
GT01
GT02
GT03
GT04
FP01
FP02
Ready to copy into approved tool Needs policy, manager, or admin review Blocked from AIdentity or admin-only
Configurator board
Targets for the first install pass.
The pack measures practical readiness: what the user can copy now, what needs review, and what must wait for IT or policy approval.
Tool-ready identity block
AI EvaluatorRole-specific starter pack
12 ready nowTool-specific setup text
copyable pathsHuman approval before risk
policy-firstDraft request, do not bypass
IT laneKeep data in approved system
boundary heldMeeting, brief, risk scan
first proofCan user apply it alone?
needs walkthroughExplain safe use
optionalRole and tool changes
laterSetup roster
12 ready / 10 gatedAvailable configuration by work lane.
Archetype, secondary signal, confidence, collaboration style
Copilot, Claude, ChatGPT Enterprise, Gemini, Glean, internal assistant
Role prompts organized by work moment and review risk
Company secrets, customer data, regulated data, and admin-only actions
One week of useful output tests inside the approved system
Connector, agent, sharing, retention, or workspace requests
Sensitive context parked before it reaches AIdentity
ID01AI Operator profile / AIdentity
Profile block
ReadyAIdentityID02AI Evaluator review habit / AIdentity
Review rules
ReadyAIdentityTL01Copilot instructions / Copilot
Tenant settings vary
ReviewCopilotTL02Claude project instructions / Claude
Project pack
ReadyClaudeTL03ChatGPT Enterprise project / ChatGPT
Workspace controls
ReviewChatGPTTL04Gemini agent prompt / Gemini
Agent access varies
ReviewGeminiTL05Glean assistant prompts / Glean
Assistant path
ReadyGleanTL06Internal assistant setup / Internal AI
Needs policy owner
BlockedInternal AIPR01Meeting prep prompts / Library
Low risk
ReadyLibraryPR02Weekly operating brief / Library
Role context
ReadyLibraryPR03Project risk scan / Library
Review required
ReadyLibraryPR04Customer follow-up draft / Library
No customer data in AIdentity
ReviewLibraryPR05Policy summary prompt / Library
Source required
ReviewLibraryPR06Spreadsheet review prompt / Library
Approved tool only
ReviewLibraryGT01Company data filter / Guardrail
Boundary
ReadyGuardrailGT02Admin request / Guardrail
IT lane
ReviewGuardrailGT03Manager explanation / Guardrail
Safe-use note
ReadyGuardrailGT04No-bypass rule / Guardrail
Locked
BlockedGuardrailFP01First output sample / Proof
Compare quality
ReadyProofFP02Prompt tuning queue / Proof
Human edit
ReviewProofPrompt queue
Prompt packs sorted by usefulness and risk.
PROMPT-001Prepare me for a recurring meeting using my role goals, open questions, and decisions needed
ReadyMeeting pack / Copy into approved AI with meeting details supplied inside the company tool
PROMPT-006Create a weekly operating brief from allowed notes and flag what needs human judgment
ReadyOperator pack / Use role context and review rules from the profile block
PROMPT-011Review a project plan for risks, missing owners, and decisions that should not be automated
ReadyEvaluator pack / Ask for evidence and assumptions before recommendations
PROMPT-014Draft customer-facing language without exposing customer data to AIdentity
ReviewComms pack / Only run inside the approved system with company policy in view
PROMPT-019Summarize a policy and identify what the user is allowed to do next
ReviewPolicy pack / Require source citation and human confirmation
PROMPT-023Analyze spreadsheet trends from an approved internal file
ReviewAnalysis pack / Keep the file inside the enterprise AI environment
PROMPT-027Explain a difficult AI concept in the user's preferred learning style
ReadyLearning pack / Use only public or user-provided non-sensitive context
PROMPT-031Create an admin request for enabling projects, agents, or connectors
ReviewEnablement pack / Send to IT or AI enablement, not directly to system settings
PROMPT-034Review AI output against the user's trust boundaries before sharing
ReadyReview pack / Check facts, policy, tone, and human decision points
PROMPT-036Refresh the prompt library after the user's role or company AI tool changes
LaterRefresh pack / Re-run after first proof or 90 days
Approval gates
Actions that need a yes/no.
Company confidential data
AIdentity creates placeholders; real data stays inside the approved enterprise tool
Customer or employee data
Prompt pack explains the boundary, but does not collect or store the content
Regulated or sensitive information
Use company policy and qualified review before asking AI to process it
Admin-only tool setup
AIdentity drafts a request; IT, security, or enablement approves the actual change
Connectors and agents
Configurator marks them optional because tenant settings may block them
Manager-visible outputs
Private coaching content stays private; only user-approved work artifacts are shared
Source-sensitive answers
Prompts require citations, assumptions, and confidence before relying on output
Policy ambiguity
When the rule is unclear, the pack tells the user to ask before proceeding
Automation requests
No autonomous actions are triggered by the prototype
Employer approval claims
The pack never claims company approval unless the employer has actually approved it