Choose the job, not the tool

Start where the work feels stuck.

You do not need to browse 100 resources or become an AI engineer. Pick the question closest to your real work, make one useful artifact, and follow the proof.

An ivory instrument plate showing an organized working library of blue, navy, and amber objects.

01 · Find your footing

I know AI matters. What should I actually learn and practice?

Build credible fluency from your real work without bluffing, bingeing tutorials, or comparing yourself with frontier demos.

FIRST MOVE · Start with the capability map, then choose one four-week practice plan.

PromptPersonal AI Use-Case MapFind safe, useful places for AI in your actual role.OPEN →PromptAI Fluency Practice PlannerTurn one capability gap into four weeks of real reps.OPEN →PromptAI Confidence Without BluffingSeparate demonstrated skill from exposure and anxiety.OPEN →SkillTeachTurn an explanation into a progression with practice.SOURCE ↗EssayThe AI Proficiency LadderUse a shared vocabulary for the next rung.SOURCE ↗

An ivory instrument plate with a blue context lens inspecting layered source material.

02 · Build better context

Why does the AI keep giving me generic work I have to repair?

Move the facts, examples, standards, and invisible judgment out of your head and into a reusable working system.

FIRST MOVE · Diagnose the repeated miss before adding more instructions.

PromptContext Gap AuditorTrace a weak output back to the missing context.OPEN →PromptReusable Context Pack BuilderAssemble durable context and a task intake block.OPEN →PromptInvisible Expertise InterviewerCapture cues, exceptions, thresholds, and reversal conditions.OPEN →SkillContext Pack ArchitectBuild, test, and maintain a portable context pack.OPEN →SkillConvert Documents to MarkdownMake source material portable and inspectable.SOURCE ↗TutorialTurn a Great Chat into a Reusable SkillPreserve a successful method without trapping conversation noise.OPEN →

An ivory instrument plate with alignment marks, blue forms, and an amber judgment indicator.

03 · Improve your judgment

How do I use AI without outsourcing the part of the work that makes me valuable?

Make your quality bar, assumptions, tradeoffs, and reversal conditions visible enough for AI to support judgment without replacing it.

FIRST MOVE · Use contrasting examples to make one important quality instinct legible.

PromptJudgment Calibration InterviewTurn a fuzzy instinct into criteria and paired examples.OPEN →PromptQuality Bar From ExamplesDerive what good looks like without copying surface style.OPEN →PromptAssumption Flip WorkshopTest whether a strategy survives the opposite world.OPEN →ToolAssumption Stress TestExpose the assumptions carrying a recommendation.OPEN →SkillDecideMake a choice with explicit criteria and tradeoffs.SOURCE ↗

An ivory instrument plate showing a blue agent path passing through a human review gate.

04 · Delegate safely

How do I give an agent real work without losing control of the mission?

Define the assignment, authority boundary, human gates, evidence, and stop condition before persistence turns into drift.

FIRST MOVE · Decide whether AI should do, prepare, challenge, observe, or stay out.

PromptDelegate or Keep Human?Choose the right role for AI in a specific task.OPEN →PromptAgent Assignment ContractSpecify the result, permissions, proof, and stop condition.OPEN →PromptHuman Approval Gate DesignerPut human review at the consequential edges.OPEN →SkillDelegation BriefPackage a rough request into a durable agent assignment.OPEN →SkillHuman ReviewGive reviewers a structured surface for edits and approval.SOURCE ↗TutorialAgent Reliability for NondevelopersDesign evidence, authority, failure, and recovery in plain language.OPEN →

An ivory instrument plate tracing an evidence line through blue source cards to a final amber receipt.

05 · Research with proof

How do I get to an answer I can actually defend?

Sharpen the question, govern the sources, connect claims to evidence, and show where the conclusion becomes inference.

FIRST MOVE · Write the decision the research must support before opening more tabs.

PromptResearch Question SharpenerTurn a broad topic into a decision-bearing question.OPEN →PromptSource Hierarchy DesignerDecide what governs and how conflicts are handled.OPEN →PromptClaim-Evidence MapperTrace each important claim to its strongest support.OPEN →ToolClaim MappingBuild a visible claim-to-source map.OPEN →SkillDeep ResearchRun the full research workflow with evidence boundaries.SOURCE ↗

An ivory instrument plate showing a structured repository of reusable workflow components and proof markers.

06 · Lead AI at work

How do I move a team from scattered experiments to better work?

Create shared working agreements, realistic value measures, learning loops, and reliability checks without turning adoption into tool theater.

FIRST MOVE · Choose one workflow and define what better work would look like before measuring usage.

PromptManager AI Working AgreementSet shared expectations for use, review, learning, and boundaries.OPEN →PromptAI ROI Reality CheckMeasure reclaimed capacity and quality instead of activity.OPEN →PromptCognitive Apprenticeship DesignerUse AI while preserving the reps that build expertise.OPEN →SkillWorkflow Reliability CheckTest normal, incomplete, repeated, and recovery cases.OPEN →SkillDiagram DesignMake a new operating model easier for a team to see.SOURCE ↗RecipeVerify Whether an AI Workflow Actually WorksCompare a workflow with the baseline on representative cases.OPEN →

Still not sure?

Open the whole library and search the problem in your own words.

SEARCH EVERYTHING