The 30 Plays

Don't Ask “How Can I Use AI?”

Stop collecting AI tricks. Start with work that is difficult, expensive, slow, impossible, or not good enough.

You win when you can identify worthwhile problems without needing an influencer to give you a use case.

Dream Bigger

AI is not only for making small tasks faster. Reconsider what becomes possible when the cost of time, labor and expertise collapses.

You win when one desired outcome would have been impractical before AI.

Begin With the End

Describe the finished state before prescribing the steps.

You win when you can define success without describing the existing process.

Make AI Interview You

Stop struggling for the perfect prompt. Give AI the rough intention and make it interrogate the ambiguity.

You win when AI helps you articulate the outcome more clearly than when you began.

Show AI What Good Looks Like

Examples, screenshots, prior work, inspiration, and good and bad outputs communicate the target better than adjectives do.

You win when you give AI evidence of “good” instead of repeatedly correcting its interpretation.

Define Done

Define audience, quality, constraints, timing, authority, evidence and acceptance criteria.

You win when two people could independently judge whether the outcome succeeded.

Preserve the Value. Eliminate the Mechanism.

Decompose a mechanism into the value it provides, then rebuild around the value.

You win when you radically change or remove a mechanism without losing the value it was meant to provide.

Your Spreadsheet Is Trying to Tell You Something

Before automating spreadsheet updates, ask why the spreadsheet exists, who consumes it, where the data originates, and what decision it supports.

You win when a recurring spreadsheet becomes smaller, automatic, unnecessary, or disappears.

Stop Being the Human API

Hunt for copy → paste → download → upload → reformat → forward → reconcile.

You win when you remove yourself from an information handoff that never required human judgment.

Stop Doing Work That Doesn't Require You

Audit attention, not hourly wage. Preserve judgment, taste, relationships, creativity and accountability; question mechanical cognition.

You win when more of your attention goes to work where you uniquely add value.

Stop Sending Everyone Everything

Stop making every recipient perform relevance classification.

You win when information increasingly finds the people for whom it matters.

Automate Last

Question, delete and simplify before automating.

You win when you kill or simplify at least one automation candidate instead of automating it.

Never Read Everything Again

Give AI the corpus; interrogate, compare, challenge and synthesize, then verify what matters.

You win when your instinct changes from “I need to read all this” to “I need to interrogate this body of information.”

Stop Making Humans Find the Signal

Find where humans scan, skim, monitor, watch, attend, check, refresh, review or search because something important might be there. Put intelligence between the stream and the human.

You win when you no longer consume an information stream to discover what matters. You enter the process when something matters.

Never Summarize When You Can Filter

A shorter pile is still a pile. Determine what deserves attention, and permit “nothing requires your attention.”

You win when AI can tell you nothing instead of manufacturing something to consume.

Don't Ask AI for Data. Ask for Intelligence.

Data → Information → Analysis → Knowledge → Decision. Ask what changed, why, what's unusual, what contradicts expectations, and what question is missing.

You win when AI regularly tells you something useful that was not explicitly present in one row, metric or document.

State, Not Messages

Extract commitments, decisions, requests, owners, deadlines, dependencies and state changes from communication.

You win when you stop asking “where was that discussed?” and start asking “what's currently true?”

Build Memory That Gets Smarter

Source → Intelligence → Memory. Preserve the evidence while accumulating structured understanding and chronology.

You win when AI can explain why something is true today and show the history and evidence that got there.

Give AI a Job, Not a Prompt

Define mission, responsibility, context, judgment, examples, tools, memory, output contract and done.

You win when you stop re-explaining a recurring task from scratch.

Teach AI Your Judgment

Show what you elevated, ignored, accepted and rejected — and why.

You win when AI makes routine distinctions similarly to you without hundreds of brittle rules.

Define the Source of Truth Before Connecting Everything

Decide which source wins for which question before agents roam across systems.

You win when AI knows not only where it can retrieve an answer but which source has authority.

Your Best AI Builder Might Be in Sales

Harness bottom-up experimentation without letting personal experiments silently become corporate infrastructure.

You win when employees experiment freely while shared capabilities have ownership and standards.

Stop Building AI Features Nobody Uses

AI makes creation cheap, so adoption becomes the bottleneck. Measure changed behavior, not feature count.

You win when success is repeated usage and outcome change rather than capabilities shipped.

Don't Teach Every Tool

Just-in-time capability beats just-in-case training. Surface the right tool when the problem occurs.

You win when people discover AI capability while solving a real problem, not because they attended another training session.

Never Automate Work You Haven't Learned to Delegate

Prove AI can perform the job when explicitly assigned before removing the human from initiation.

You win when removing yourself from initiation feels boring rather than scary.

Manage by Exception

Routine state transitions proceed under policy; humans receive exceptions, disagreements, low-confidence cases and decisions.

You win when normal work completes without demanding attention.

Give AI Initiative

Responsibility becomes initiative. AI acts because state requires it, not because you remembered to ask.

You win when useful work begins because the situation requires it.

Give Your Systems a Night Shift

While you sleep, AI reviews errors, anomalies, stale commitments, complaints and performance; performs safe work; and prepares the exceptions.

You win when you wake up to useful work that happened because AI recognized it needed doing.

Write an Autonomy Contract

Autonomy requires an explicit trigger, authority, boundaries, approvals, exceptions, escalations, observability and failure behavior.

You win when you can explain precisely what the system may do without you.

RE+SHAPE

Review what you stopped, kept human, augmented, delegated and made autonomous. Then identify the next things worth revealing and examining.

You win when your instinctive question becomes: how should this get done now that AI exists?