The AI Readiness Toolkit

AI is already in the room. Readiness is what keeps it useful, safe, trusted, and real.

Move from AI curiosity, scattered use, and unclear risk into practical, trusted, responsible adoption. Sixteen tools and seven premium appendices — inventory current use, prioritize real use cases, tier risk, set guardrails, prepare for agents, pilot responsibly, and build your 30/60/90-day path.

The assessment shows the pattern. This toolkit helps build the readiness.

AI doesn't improve work on its own. People do. Systems do. Decisions do. Trust does. Readiness is what allows all of it to work together.

This toolkit is built to be used, not filed away. Work through it in order, or jump straight to the tool that matches where you're stuck — a Start Here map is below. Every tool ends in a tiny, usable output.

The goal isn't to chase every new AI feature. It's to build the conditions — direction, judgment, guardrails, and trust — that let AI become part of how work actually gets done.

Start Here — Match Your Situation

If This Is Your Situation	Start With
Just finished the AI Readiness Assessment	Tool 1 · AI Readiness Debrief
You suspect people are already using AI quietly	Tool 2 · Current-State Inventory, Tool 5 · Risk Tiering
Leadership wants AI use cases	Tool 3 · Use-Case Map, Tool 4 · Prioritizer
Teams have AI tools but don't know how to use them well	Tool 7 · Rules of the Road, Tool 9 · Verification
Considering AI agents or automation	Tool 8 · Agentic Boundary Map, Appendix D · Security

A note on pace: don't try to do all sixteen tools in a week. Most teams move through Debrief → Inventory → Use-Case Map in week one, then pick up the pace from there.

Debrief & Inventory

Tool 1 — AI Readiness Debrief

Use this after completing the AI Readiness Assessment. The goal isn't to defend the result — it's to use it as a mirror, together.

Result Area	Response
AI Readiness Zone
AI Readiness Score
Strongest Area
Biggest Readiness Gap
Primary Pattern Flag

What part of this result feels accurate?

Where are we more ready than we think — what are we already doing well?

Where are we less ready than we say — tools but not habits, excitement but not ownership?

Name the one risk we shouldn't ignore.

Choose one primary next move (not five).

Tiny Output: Our AI readiness is strongest in ____________, but we need to strengthen ____________ before AI can become part of how work gets done.

Debrief & Inventory

Tool 2 — AI Current-State Inventory

Identify what AI use already exists before creating new plans. Most organizations start as if nothing is happening yet — that's rarely true.

Tool / Platform	Who Uses It?	For What Work?	Approved?
ChatGPT
Claude
Copilot
Embedded AI (CRM, HRIS, helpdesk...)
Other

What is the first shadow AI signal we should investigate? (public tools used because approved ones are unavailable, teams inventing their own rules, sensitive work entering unapproved tools...)

What finding needs immediate attention?

Tiny Output: AI is already showing up in ____________, and the first place we need more clarity is ____________.

Map & Prioritize

Tool 3 — AI Use-Case Map

Find where AI could create meaningful value. A good use case starts with work friction, not tool excitement.

What team, function, process, or workflow are we examining?

What does "better" mean here?

Friction Type	Where does this show up?
Time Drag
Rework
Search Burden
Decision Delay
Communication Load

Definitions: LLM Use (generate/summarize/analyze), Embedded AI (already inside a tool you use), Automation (rules complete a repeatable task), Agentic Workflow (AI can trigger, update, escalate, or coordinate across steps).

Name 3-5 use cases ready for prioritization.

Tiny Output: Our most promising AI opportunity is ____________, because it addresses real friction, not just excitement about the tool.

Map & Prioritize

Tool 4 — AI Use-Case Prioritizer

Score each use case 1–5 across eight criteria. Total 34–40 is a strong pilot candidate; below 20, don't prioritize now.

Use Case	Value	Feasibility	Risk (reversed)	Readiness	Total

"Are we fooling ourselves?" check: are we choosing this because it's valuable, or because it's easy or a leader likes it?

Tiny Output: Our best first AI use case is ____________ because it has strong value, manageable risk, and enough readiness to test responsibly.

Risk, Guardrails & Agents

Tool 5 — AI Risk Tiering Matrix

Not every AI idea needs the same review. When everything is treated as high risk, adoption slows. When everything is treated as low risk, trust breaks.

Risk Tier	Description	Review Path
Tier 1 — Low Risk	Brainstorming, generic drafting, non-sensitive summaries	Team-level guidance
Tier 2 — Moderate Risk	Internal data summaries, customer response drafts	Manager or SME review
Tier 3 — High Risk	Employee, financial, legal, clinical, regulated use	Formal review
Tier 4 — Prohibited	Sensitive data in public tools, high-impact automation without approval	Not allowed

Our Use Case	Tier (1–4)	Why

Tiny Output: Our clearest Tier 4 (prohibited) line is ____________.

Risk, Guardrails & Agents

Tool 6 — Responsible AI Guardrails Builder

Build allowed, cautious, and prohibited use boundaries — plus transparency rules for when AI use should be disclosed.

Where should AI use be disclosed, and to whom?

Example: "AI was used to support drafting, summarizing, or organizing this content. A human reviewed and approved the final version."

Draft the Guardrail Statement

Our organization/team uses AI to ____________.

We allow AI for ____________.

We use caution when ____________.

We do not use AI for ____________.

Humans remain accountable for ____________.

When something is unclear, we ____________.

Tiny Output: Our AI guardrails should help people use AI confidently while protecting ____________, ____________, and ____________.

Risk, Guardrails & Agents

Tool 7 — Everyday AI Rules of the Road

Practical team-level rules for the daily stuff: drafting, summarizing, brainstorming, research support, meeting notes.

Rule 1 — Use AI to support thinking, not replace it.
Rule 2 — Don't paste sensitive information into unapproved tools.
Rule 3 — Check before you trust.
Rule 4 — Use AI where it reduces drag, not to avoid accountability.
Rule 5 — Be transparent when it matters.
Rule 6 — Escalate the weird stuff.

We Can Use AI For…	We Won't Use AI For…	We Need Approval Before…

Drafting Prompt StarterHelp me draft a first version of [X] for [audience]. Keep it clear, practical, and aligned to this goal: [Y]. Ask me any questions before finalizing if the context is unclear.

Team Agreement: complete the statement in your own words and post it somewhere visible.

Risk, Guardrails & Agents

Tool 8 — Agentic Workflow Boundary Map

Agents may take action, update records, trigger tasks, or escalate issues. Define what AI can assist, recommend, prepare, or do — and where human approval must remain.

Name the workflow being considered for agentic support.

Level	AI Role	Human Role
1 — Assist	Drafts, summarizes, searches, organizes	Reviews and uses judgment
2 — Recommend	Suggests a next step	Decides
3 — Prepare	Prepares an action	Approves before execution
4 — Act Within Guardrails	Completes low-risk actions within defined limits	Monitors
5 — Escalate Exceptions	Stops or escalates when unclear or risky	Resolves exception

What actions are strictly No-Go for this workflow (high-impact, legally significant, irreversible, requires empathy or accountability)?

Agent Readiness Decision: Not ready / Assist only / Recommend only / Prepare with approval / Limited action / Ready for controlled pilot?

Tiny Output: In this workflow, AI may ____________, but a human must approve ____________.

Judgment & Trust

Tool 9 — Human Judgment + Verification Checklist

AI confidence is not the same as accuracy. This turns "be careful" into a practical review habit.

Check	Question	Finding
Accuracy	Is this factually correct? What needs verification?
Context	Does this fit our situation? What might AI be missing?
Judgment	Would I stand behind this if challenged?
Bias + Fairness	Could this disadvantage anyone?
Privacy	Does this include sensitive information?

Decision: Use / Edit / Verify / Escalate / Do Not Use?

Red flag rule: if AI output affects someone's employment, care, access, finances, or rights — pause, always.

Tiny Output: Before we use AI output, we'll always check for ____________, ____________, and ____________.

Judgment & Trust

Tool 10 — Team Trust Conversation Guide

This conversation isn't about convincing everyone to love AI. It's about understanding what people need to use it responsibly.

Questions are not resistance. Concerns are not negativity. Silence is not alignment.

Curiosity round: what could AI help us do better?

Concern round: what feels unclear or risky? What are people worried will happen?

What should remain human — where does judgment, empathy, or accountability matter most?

Choose one low-risk experiment to test.

Leadership Follow-Up Action	Owner	Date

Tiny Output: For our team to trust AI adoption, we need clarity on ____________, support with ____________, and boundaries around ____________.

Judgment & Trust

Tool 11 — Workflow Redesign Canvas

The question isn't "where can we add AI?" It's: how should this work change if AI can support part of it?

Name the workflow, its owner, and its current pain point.

Step	Who Does It Today?	Pain Point	AI Role (Assist/Recommend/Prepare/Act)
1
2
3

Define success: what measure, current state, and target?

Tiny Output: AI should improve this workflow by ____________, not simply make the old process faster.

Pilot & Learn

Tool 12 — AI Pilot Charter

A pilot shouldn't be "let's try AI and see what happens." Define scope, data boundaries, review, and success before launch.

Charter Element	Response
Pilot Name / Use Case
Problem We're Solving
Pilot Owner / Sponsor / Users
Data Allowed / Not Allowed
Human Review Required
Risk Tier
Success Measures
Stop / Continue / Scale Criteria

Tiny Output: This pilot will test whether AI can help ____________ without increasing ____________.

Pilot & Learn

Tool 13 — AI Learning Log

AI readiness improves through practice, not theory. Capture what worked, what didn't, and what risk appeared — every check-in.

Check-In Date	What We Tried	What Worked	What Didn't / Risk Seen

What decision trail follows — stop, continue, redesign, or scale?

Sustain & Scale

Tool 14 — 30/60/90-Day AI Readiness Plan

The goal isn't to do everything at once — it's to build enough readiness to move responsibly.

Phase	Focus	Our Actions
First 30 Days — Clarify	Shared view of where AI can help and where risk exists
Days 31–60 — Design	Basic structures for responsible adoption
Days 61–90 — Enable	Move from planning to responsible testing

Tiny Output: In the next 90 days, our AI readiness focus is to move from ____________ to ____________.

Sustain & Scale

Tool 15 — AI Leadership Operating Rhythm

A monthly dashboard for active use cases, risks, adoption, and value — so AI readiness stays a rhythm, not a one-time project.

Area	Status (Green/Yellow/Red)	Decision Needed
Active Use Cases
Risks / Issues
Adoption / Trust
Value / Measurement

Tiny Output: This month, the most important AI readiness decision is ____________.

Sustain & Scale

Tool 16 — Final AI Readiness Action Plan

The synthesis page. Pull everything you've built into one clear plan you can share, revisit, and execute.

Area	Decision
Our AI Readiness Zone & Primary Pattern Flag
Our strongest area / biggest gap
Our first AI use case & biggest risk
Our first guardrail priority
Our 30 / 60 / 90-day actions
Owner / Sponsor / Next Review Date

Our AI Readiness Assessment showed we are in the ____ Zone, strongest in ____, with our biggest gap in ____. Our primary pattern is ____. Over the next 90 days, we will focus on ____, ____, and ____.

Governance Appendices

Appendix A — AI Governance Charter

Without visible governance, AI decisions drift. With too much governance, AI freezes. The goal is right-sized structure.

Charter Area	Decision
Governance Group Name / Purpose
Executive Sponsor / Chair
Core Members
Decision Rights
Meeting Cadence / Escalation Path

Governance Appendices

Appendix B — AI Governance RACI

AI work gets messy when everyone is interested but no one owns the hard parts. R = Responsible, A = Accountable, C = Consulted, I = Informed.

Decision / Activity	Executive Sponsor	Business Owner	IT/Data	Security/Legal
Define AI priorities
Identify use cases
Approve pilots

Governance Appendices

Appendix C — AI Tool / Vendor Intake Checklist

Use before approving a new AI tool, vendor, or embedded AI feature inside a platform you already use.

What tool or vendor is being reviewed, and what AI capability does it provide?

What data will it access — sensitive, regulated, customer, employee, financial?

Does the vendor train on our data, and can that be disabled? Where is data stored and retained?

What controls exist — SSO, role-based access, audit logs, admin controls?

Final decision: Approve / Approve with Conditions / Pilot Only / More Review Needed / Do Not Approve?

Governance Appendices

Appendix D — LLM + Agent Security Checklist

The practical question: could the AI system be manipulated, leak sensitive information, act beyond its authority, or produce output people trust too quickly?

Risk Area	Question	Finding
Prompt Injection	Could a user or external content manipulate the AI's instructions?
Sensitive Info Disclosure	Could sensitive information be exposed in prompts, outputs, or logs?
Excessive Agency	Could the AI take actions beyond what it should be allowed?
Auditability	Can we see what the AI did, changed, or recommended?

Security Review Decision: Low concern / Proceed with controls / Security review required / Do not proceed?

Capability & Value

Appendix E — AI Capability + Training Plan

Different audiences need different levels of AI understanding — executives, managers, employees, and technical teams all need something different.

Audience	Capability Needed	Format & Owner
Executives
Managers
Employees
IT / Data / Security

Capability & Value

Appendix F — AI Adoption Communications Pack

Ready language for the questions people actually ask — and aren't sure it's safe to ask out loud.

"I'm worried AI will just mean more work."That's a fair concern. One of the things we need to look at is whether AI actually removes, simplifies, or supports work — not just increases output expectations.

"I don't trust the output."You shouldn't trust AI output automatically. We need verification habits and clear rules for when human review is required.

"People are already using it anyway."That is exactly why we need to understand current use and create guidance that is practical enough for real work.

How do people raise a concern in our organization — who do they contact?

Capability & Value

Appendix G — AI Value + Measurement Scorecard

AI pilots shouldn't be evaluated only by enthusiasm. They should be evaluated by evidence.

Measure	Baseline	Target	Actual	Evidence
Time saved
Quality improved
Trust improved

AI doesn't improve work on its own. People do. Systems do. Decisions do. Trust does. Readiness is what allows all of it to work together.