Your 30-day plan
to become the person AI can't replace.

Audit Your Real Week

Task

Reconstruct where your time actually went this week from raw calendar and task exports, then classify every block by what it demands of you. The goal is an honest map of your real work, not the job description version of it.

1. Export last week from your calendar and pull your closed and touched Jira items
2. Paste both into the model with your role and typical responsibilities
3. Ask it to list every distinct activity with an estimated hours-per-week figure
4. Have it tag each activity as thinking, coordinating, producing, or reacting
5. Copy the table into a sheet and correct any hours that feel wrong

Tool

ChatGPTWeb

Artifact

Weekly Task Audit sheet

Ship

A single sheet listing every recurring activity in your week with hours and a work-type tag.

Concept lockYou can only automate work you have named and measured.

WhyYou cannot automate work you have never named, and most of your week is invisible to you until it is written down. This audit is the baseline every later day measures against, so an honest one now pays off for the whole month.

Reflect before moving onWhich activity surprised you with how many hours it takes?  ·  How much of your week is reacting versus thinking?  ·  Which tasks would you never miss if they vanished?

Score Your Task Map

Task

Take yesterday's audit and score each activity on two axes: how exposed it is to AI and how much it drains you. The output is a ranked triage list that tells you exactly what to automate first.

1. Feed the model your Weekly Task Audit sheet from day one
2. Ask it to score each activity from one to five on AI-exposure and on personal drain
3. Have it multiply the two scores into a single priority number
4. Sort the list so the highest combined score sits on top
5. Mark the top three as your automation targets for this month

Tool

ClaudeWeb

Artifact

Scored Task Map sheet

Ship

A ranked sheet where the top three targets for automation are explicitly circled.

Concept lockAutomate where high AI-exposure meets high personal drain.

WhyNot all tasks are worth automating, and chasing the wrong one wastes a week of effort for little return. Scoring forces you to spend your energy where exposure and drain overlap, which is where the fastest, most durable wins live.

Reflect before moving onDid the highest-priority task match your gut, or surprise you?  ·  Is anything high-drain but low-exposure that you should hand off instead?  ·  What would this week feel like without the top three tasks?

Automate Your Weekly Status Report

Task

Reverse-engineer the prompt that reproduces your weekly status report from raw inputs, then run it against this week. The goal is a one-pass draft you trust enough to edit rather than rewrite.

1. Collect this week raw inputs: Jira export, key Slack decisions, and metric deltas
2. Paste them with a description of your report format, audience, and tone
3. Ask the model to produce the full report in your structure
4. Diff its draft against what you would have written and note every gap
5. Fold the gaps into the prompt and save the final version

Tool

ChatGPTWeb

Artifact

Status Report Prompt prompt

Ship

A saved prompt that produces an 80-percent-usable status report in one pass.

Concept lockA good prompt encodes your format once so you never re-explain it.

WhyStatus reporting is your single highest-exposure task, so automating it first frees the most time immediately. It also proves the whole method works on day one, which builds the momentum the rest of the month depends on.

Reflect before moving onWhere did the draft still need your judgment?  ·  What context did you have to add by hand?  ·  Which other weekly report could reuse this pattern?

Meeting Prep Automation

Task

Build a reusable template that turns a meeting title and a few links into a full prep pack: objective, key questions, and the decision you need out of the room. Run it on your three biggest meetings this week.

1. List the recurring meetings you walk into underprepared
2. Give the model the agenda, attendees, and any relevant docs or threads
3. Ask it to produce an objective, three sharp questions, and the single decision to land
4. Have it flag what could derail the meeting and how to steer back
5. Save the structure as a template you fill in before every meeting

Tool

ChatGPTWeb

Artifact

Meeting Prep Template prompt

Ship

A reusable prep template that turns any meeting into a one-page brief in under five minutes.

Concept lockPreparation is a template, not a personality trait.

WhyWalking into meetings unprepared is where PMs quietly lose influence and time, meeting after meeting. A prep pack that writes itself means you are the sharpest person in every room without the hour of manual reading.

Competitor Scan System

Task

Build a repeatable competitor scan that pulls recent product, pricing, and positioning changes across your top rivals into a single brief. Run it on three competitors and save the query so next month is one click.

1. Name your three closest competitors and the signals you care about
2. Ask the tool to surface changes from the last quarter with sources for each claim
3. Have it separate confirmed changes from rumours and speculation
4. Request a short so-what line on how each change affects your roadmap
5. Save the query and set a recurring reminder to rerun it monthly

Tool

PerplexityWeb

Artifact

Competitor Brief doc

Ship

A sourced competitor brief covering three rivals with a so-what line on each change.

Concept lockA standing scan beats a frantic search every time.

WhyCompetitive blind spots get discovered in the worst possible moment, usually mid-review in front of leadership. A standing scan means you always know what moved and why it matters before anyone thinks to ask you.

PRD First-Draft Generator

Task

Turn a rough problem statement and a handful of inputs into a structured PRD first draft you can react to instead of stare at. The point is to kill the blank page, not to ship the model's words unedited.

1. Gather the problem, target user, known constraints, and any research links
2. Feed them to the model with your standard PRD section headings
3. Ask for a full first draft with explicit open questions flagged inline
4. Challenge every assumption it made and rewrite the ones that are wrong
5. Save your section headings and instructions as a reusable PRD scaffold

Tool

ClaudeWeb

Artifact

PRD Draft doc

Ship

A structured PRD first draft with open questions flagged, ready for your edits.

Concept lockA draft you can argue with beats a blank page you can only fear.

WhyThe blank page is where product work stalls, and a first draft you disagree with is far more useful than no draft at all. Reacting is faster than authoring, so you get to the sharp thinking sooner and spend your judgment where it counts.

Reflect before moving onWhich assumption did the model get wrong, and why?  ·  Where did reacting to a draft sharpen your own thinking?  ·  What belongs in your permanent PRD scaffold?

Customer Feedback Synthesis

Task

Feed a month of raw feedback from support tickets, reviews, and calls into a single pass that clusters it into themes with frequency and representative quotes. This replaces the afternoon you usually lose skimming and half-remembering.

1. Gather raw feedback: support tickets, app reviews, sales notes, and call snippets
2. Paste it in with instructions to cluster into themes, not summarise
3. Ask for each theme with a frequency count and two verbatim quotes
4. Have it separate feature requests from pain points from praise
5. Rank themes by impact and drop the top three into your roadmap inputs

Tool

ClaudeWeb

Artifact

Feedback Themes doc

Ship

A themed feedback summary with frequency counts and real quotes for the top three issues.

Concept lockThemes with counts and quotes beat opinions in every roadmap debate.

WhyCustomer signal is your strongest source of authority in any roadmap fight, but only if you can produce it fast and back it with real quotes. Turning a messy inbox into ranked themes in one pass makes you the person who actually knows what users think, not just claims to.

Reflect before moving onWhich theme was bigger than you assumed it was?  ·  Did any quote change how you frame the problem?  ·  Which theme do you now have the evidence to push on?

Roadmap Update Automation

Task

Generate a clean roadmap update from your current Jira state and the decisions made this week, framed for a non-technical audience. The goal is a narrative of progress and change, not a raw ticket dump.

1. Pull your epics and their current status from Jira
2. Give the model the status plus any scope or priority changes from this week
3. Ask for a themed update grouped by outcome, not by ticket
4. Have it call out what changed since last update and why
5. Save the format so each roadmap update takes minutes, not hours

Tool

ChatGPTJira

Artifact

Roadmap Update doc

Ship

A stakeholder-ready roadmap update grouped by outcome with changes explained.

Concept lockReport outcomes, not tickets, and you sound like an owner.

WhyStakeholders do not want a list of tickets, they want to know if the bet is on track and what shifted. An update that speaks in outcomes makes you look in command of the plan rather than buried in it.

Stakeholder Digest

Task

Turn a week of scattered Slack decisions and updates into a tight digest that each stakeholder group actually reads. One input, several framings, so nobody has to dig through channels.

1. Collect the key threads, decisions, and updates from the week across channels
2. Paste them in and name your stakeholder groups and what each one cares about
3. Ask for one digest per group, reframed for their priorities
4. Have it keep each digest to five lines with a clear so-what
5. Save the prompt and send the digest every Friday

Tool

ChatGPTSlack

Artifact

Decision Digest email

Ship

A per-audience decision digest, five lines each, ready to send Friday.

Concept lockThe person who writes the record controls the memory.

WhyDecisions that live buried in Slack threads get forgotten and relitigated, which costs you the same argument twice. A weekly digest makes you the source of record and quietly puts you at the centre of what the team knows.

Chain Two Tools

Task

Wire two tools into one flow where the output of a research pass becomes the input to a synthesis pass, with no manual reshaping between them. This is your first real system rather than a single clever prompt.

1. Pick a recurring job that needs research then synthesis, such as a market brief
2. Design a research prompt whose output is clean structured notes
3. Design a second prompt that takes those notes and produces the final artifact
4. Run the chain end to end and fix the handoff format until it flows
5. Document the two prompts and the handoff shape as one named chain

Tool

PerplexityWeb

Artifact

Research-to-Synthesis Chain prompt

Ship

A two-step chain where research output feeds synthesis with no manual cleanup between.

Concept lockThe value is in the handoff, not in either prompt alone.

WhySingle prompts save minutes, but chained prompts save whole workflows and are far harder for anyone to replicate. Learning to design the handoff between steps is the core skill that separates an AI user from an AI orchestrator.

Reflect before moving onWhere did the handoff format break, and how did you fix it?  ·  What other two-step jobs could become a chain?  ·  How much manual reshaping did the chain remove?

Reusable Prompt Library

Task

Consolidate every working prompt you have built so far into one organised library with a clear format, so nothing lives only in a chat history. This turns scattered wins into an asset you compound on.

1. Collect the prompts from days three through ten into one place
2. Define a standard entry: name, purpose, inputs, the prompt, and known limits
3. Rewrite each prompt to that standard and remove one-off specifics
4. Tag each by task type so you can find the right one in seconds
5. Store it in Notion and add a rule to log every new prompt here

Tool

ClaudeNotion

Artifact

Prompt Library doc

Ship

A Notion library holding every prompt so far in one consistent, searchable format.

Concept lockA prompt you cannot find again is a prompt you have to rebuild.

WhyPrompts trapped in old chat windows are prompts you will rebuild from scratch, wasting the work you already did. A living library means each new prompt makes you faster forever, which is how the compounding actually happens.

Discovery Interview Analyzer

Task

Build a system that turns raw discovery interview transcripts into a structured insight table: the need, the underlying job, the pain, and a supporting quote. Run it on your last three interviews.

1. Gather transcripts or detailed notes from recent discovery calls
2. Paste one in and ask for a table of need, job-to-be-done, pain, and quote
3. Have it flag where the user said one thing but implied another
4. Run all three and ask the model to merge them into shared patterns
5. Save the format as your standard post-interview analysis step

Tool

ClaudeWeb

Artifact

Insight Table sheet

Ship

A merged insight table across three interviews with patterns and supporting quotes.

Concept lockThe insight is in the pattern across interviews, not any single one.

WhyDiscovery loses most of its value in the gap between the conversation and the write-up, where insight quietly evaporates. A consistent analyzer captures the signal every time and lets you spot patterns across interviews that no single call reveals.

Metrics Narrative Generator

Task

Turn a raw metrics dump into a narrative that explains what moved, why it likely moved, and what you would do about it. Numbers on their own do not persuade anyone.

1. Export the current period metrics with prior-period comparisons
2. Paste them in with context on any launches or events that could explain shifts
3. Ask for a narrative: what moved, the likely driver, and a recommended action
4. Have it separate real signal from noise and normal variance
5. Save the format for your monthly business review prep

Tool

ChatGPTWeb

Artifact

Metrics Narrative doc

Ship

A metrics narrative that explains movements and recommends an action for each.

Concept lockNumbers report the past, narrative drives the decision.

WhyA dashboard tells people what happened, but a narrative tells them what it means and what to do, which is where your value sits. Being the one who explains the numbers rather than just reports them is what gets you into the strategy conversation.

Prioritization Copilot

Task

Build a copilot that scores a backlog against your real prioritization framework and explains each score, so you can defend the order in any room. Run it on your current backlog and stress-test the output.

1. Write out the prioritization framework you actually use, criteria and weights
2. Give the model the framework plus your current backlog items with context
3. Ask it to score each item, show the maths, and rank the list
4. Argue with three of its scores and see if it holds or folds correctly
5. Save the framework and prompt so every backlog runs through it

Tool

ClaudeWeb

Artifact

Prioritization Copilot sheet

Ship

A scored, ranked backlog with a transparent reason behind every score.

Concept lockA ranking you can explain is a ranking you can defend.

WhyPrioritization is where PMs are most often overruled, because a ranking without a rationale is just an opinion. A copilot that shows its working turns your backlog order into an argument you can win in front of anyone.

Reflect before moving onWhere did arguing with a score reveal a flaw in your framework?  ·  Which item jumped or dropped once scored honestly?  ·  Could you defend this order to your toughest stakeholder now?

Spec-to-Ticket Pipeline

Task

Build a pipeline that turns an approved spec into a clean set of engineering-ready tickets with acceptance criteria, ready to drop into Jira. This kills the tedious breakdown work that eats your afternoons.

1. Take one approved spec and paste it in with your ticket conventions
2. Ask the model to break it into tickets sized for a single engineer
3. Require acceptance criteria and dependencies on every ticket
4. Have it flag anything underspecified rather than inventing detail
5. Review, adjust sizing, and paste the set straight into Jira

Tool

ChatGPTJira

Artifact

Ticket Set checklist

Ship

A spec broken into engineering-ready tickets with acceptance criteria and dependencies.

Concept lockAutomate the breakdown, keep the judgment.

WhyBreaking specs into tickets is high-volume, low-judgment work that quietly consumes your week. Automating the breakdown while keeping the judgment calls to yourself frees hours and gets engineers cleaner tickets than a rushed manual split ever would.

Release Notes Autowriter

Task

Generate customer-facing release notes straight from merged pull requests and commit messages, translated out of engineering language into user benefit. Run it on this release and save the flow.

1. Pull the merged PRs and commit messages for the current release
2. Paste them in with a note on who reads your release notes
3. Ask for notes grouped by user-facing benefit, not by internal component
4. Have it drop purely internal changes and flag anything needing a heads-up
5. Save the prompt so every release turns notes around in minutes

Tool

ClaudeGitHub

Artifact

Release Notes doc

Ship

Customer-ready release notes generated from PRs, grouped by user benefit.

Concept lockShip the benefit, not the changelog.

WhyRelease notes are usually written last, in a rush, and read like a changelog nobody wanted. Turning raw commits into clear benefits fast means your releases land with users instead of landing with a thud.

Ship an AI-Native Workflow

Task

Package one of your automations into a shareable kit your team can run without you: the prompt, the inputs, and a short how-to. This is the moment your personal win becomes a team capability.

1. Pick your most reliable automation from the past two weeks
2. Write a plain-language how-to a teammate could follow cold
3. Bundle the prompt, an example input, and an example output together
4. Have a colleague run it once and note every point they got stuck
5. Fix the friction and post the finished kit to a shared team channel

Tool

ClaudeSlack

Artifact

Team Workflow Kit doc

Ship

A self-serve workflow kit a teammate ran successfully without your help.

Concept lockA workflow others can run is leverage, one only you can run is a hobby.

WhyAn automation only you can run is a personal trick, but one your team can run is leverage that spreads your name with it. Shipping the first kit marks your shift from doing the work faster to changing how the team works.

Quantify Hours Reclaimed

Task

Build a simple model that estimates the hours you and your team have reclaimed from the automations shipped so far, with defensible assumptions. Numbers turn a good story into a case leadership can act on.

1. List every automation you have shipped and who now uses it
2. Estimate the manual time each task took before and takes now
3. Multiply by frequency and headcount to get monthly hours saved
4. Keep the assumptions visible and deliberately conservative
5. Convert the hours into a rough cost figure using a blended rate

Tool

ChatGPTWeb

Artifact

Time-Savings Model sheet

Ship

A conservative model showing monthly hours and cost reclaimed, with assumptions shown.

Concept lockUnmeasured impact is unrewarded impact.

WhyImpact that is not measured is impact that is not rewarded, and vague claims of saving time convince nobody. A defensible model turns your month of work into a number leadership can put in a plan, which is what gets you the next mandate.

Present the ROI to Leadership

Task

Turn your time-savings model into a short, sharp deck that makes the case for expanding this work, aimed at the people who control mandates and budget. This is where you convert effort into permission.

1. Start from your quantified hours and cost reclaimed
2. Structure the deck as problem, what you did, result, and the ask
3. Ask the model to draft speaker notes that anticipate the skeptical question
4. Cut every slide that does not move the decision forward
5. Rehearse once against the notes and tighten the ask

Tool

ClaudeWeb

Artifact

ROI Deck slide

Ship

A tight leadership deck ending in a specific ask backed by your numbers.

Concept lockInvisible results change nothing, presented results change your mandate.

WhyA result nobody with power sees is a result that changes nothing about your trajectory. Presenting the ROI well is how a month of quiet automation becomes visible leverage and an explicit mandate to do more.

Train One Peer

Task

Teach one colleague to run the method on their own work, and capture what you taught as a reusable training note. Teaching forces you to make your own process explicit and multiplies its reach.

1. Choose a peer whose work has an obvious automatable task
2. Sit with them and map that task the way you mapped yours on day one
3. Build one automation together, letting them drive the prompting
4. Write up what you covered as a short training note others can follow
5. Ask them to teach the next person, and note where they struggled

Tool

ChatGPTWeb

Artifact

Peer Training Note doc

Ship

One trained peer with a working automation, plus a reusable training note.

Concept lockYou lead the shift by creating others who can run it.

WhyThe fastest way to be seen as a leader in this shift is to visibly create other people who can do it. A training note turns a single mentoring session into something that scales past the two of you.

Propose an AI-Product Initiative

Task

Draft a concrete proposal for an AI-powered feature or internal capability that only someone who has lived this month could credibly pitch. This moves you from automating your job to shaping the product.

1. Look across your themed feedback and metrics for an AI-shaped opportunity
2. Frame the problem, the proposed capability, and who it serves
3. Ask the model to pressure-test the idea for feasibility and obvious risks
4. Add a small, cheap first experiment to prove or kill it fast
5. Write it as a one-page proposal aimed at a decision-maker

Tool

ClaudeWeb

Artifact

Initiative Proposal doc

Ship

A one-page proposal for an AI initiative with a cheap first experiment defined.

Concept lockAutomating your job makes you efficient, directing AI in the product makes you strategic.

WhyAutomating your own tasks makes you efficient, but proposing where AI should go in the product makes you strategic. This is the day you stop being someone AI could replace and become someone deciding how AI gets used.

Reflect before moving onCould only someone who lived this month have written this proposal?  ·  What is the cheapest way to prove the idea is wrong?  ·  Who needs to see this, and what is your ask of them?

Draft the Business Case

Task

Expand your initiative proposal into a business case with costs, expected value, and a phased plan a decision-maker can fund. This is the document that separates an idea from a mandate.

1. Start from your one-page initiative proposal
2. Estimate build cost, ongoing cost, and the value in the same units
3. Lay out a phased plan where phase one is small and reversible
4. Ask the model to stress-test the numbers and flag the weakest assumption
5. Tighten it into a fundable case with a clear recommendation

Tool

ChatGPTWeb

Artifact

Business Case doc

Ship

A phased business case with costs, value, and a clear funding recommendation.

Concept lockIdeas get admired, business cases get funded.

WhyIdeas do not get funded, business cases do, and the ability to write one is what pulls you toward the decision table. Framing your initiative in the language of cost and value is how you get taken seriously by the people who allocate resources.

Claim the Next Seat

Task

Write a sharp repositioning brief that redefines your role around AI orchestration, backed by the evidence you have built this month. This is you naming the seat you intend to take before anyone offers it.

1. Gather the proof: automations shipped, hours saved, peers trained, the initiative
2. Draft how your role should be described a year from now
3. Ask the model to contrast your current scope with the target scope
4. Have it surface the gap you need to close and how to close it
5. Turn it into a brief you can use in a career conversation

Tool

ChatGPTWeb

Artifact

Repositioning Brief doc

Ship

A repositioning brief that redefines your role around AI orchestration, backed by evidence.

Concept lockThe seat goes to whoever describes it first.

WhyIn a shift this fast, roles get redefined by the people who define them first, not the ones who wait to be told. A clear brief means you walk into your next career conversation with the new seat already described and half-earned.

Document Your Playbook

Task

Consolidate everything you built this month into a single AI playbook that captures your method, your prompts, and your standards in one place. This is the asset that outlives any single automation.

1. Gather your prompt library, your chains, and your kits into one outline
2. Write the method as principles, not just a list of tools
3. Add a standards section: when to automate, when not to, quality bars
4. Have the model tighten the structure and flag any gaps in your logic
5. Publish it in Notion as the living reference for you and your team

Tool

ClaudeNotion

Artifact

AI Playbook doc

Ship

A single published playbook holding your method, prompts, and standards.

Concept lockA documented method is a system, an undocumented one is a streak.

WhyScattered wins are fragile, but a documented playbook is a durable asset that keeps paying out after the novelty fades. Writing down your method is also what makes it teachable, which is the difference between a skill and a system.

Delegate a Workflow

Task

Fully hand off one of your automations to someone else, so it runs without you at all. Real leverage is not doing the task fast, it is no longer being in the loop.

1. Pick a stable automation that does not need your judgment to run
2. Write a handover that covers inputs, the run, the checks, and failure modes
3. Walk the new owner through one full run start to finish
4. Have them run the next one solo while you only watch
5. Remove yourself from the loop and note where they still need backup

Tool

ChatGPTWeb

Artifact

Delegation Handover doc

Ship

One automation running under a new owner with you fully out of the loop.

Concept lockLeverage is not doing it fast, it is no longer being in the loop.

WhyIf every automation still routes through you, you have built a bottleneck instead of leverage. Delegating cleanly is what frees your attention for the higher work and proves your method survives without you.

Weekly AI Review Ritual

Task

Design a lightweight weekly ritual that keeps your automations healthy and surfaces the next thing to automate. A system without maintenance quietly rots, so you build the upkeep in.

1. List the checks that keep your automations trustworthy over time
2. Draft a 15-minute weekly agenda: what to review, retire, and add
3. Include a slot to capture one new automation candidate each week
4. Add a quality check for any prompt whose output has drifted
5. Put a recurring block on your calendar and commit to the first one

Tool

ClaudeCalendar

Artifact

Review Ritual checklist

Ship

A 15-minute weekly review ritual booked as a recurring calendar block.

Concept lockAn unmaintained system is a system quietly failing.

WhyAutomations degrade as tools, data, and needs change, and an unmaintained system slowly loses the trust you built. A short standing ritual keeps the whole method alive and ensures your automation backlog never runs dry.

Team Prompt Registry

Task

Turn your personal prompt library into a shared team registry with contribution rules, so the whole team compounds instead of just you. This is where your individual method becomes team infrastructure.

1. Adapt your prompt library structure for shared team use
2. Define contribution rules: format, review, and who owns quality
3. Seed it with your best prompts, generalised past your own context
4. Add a simple tagging system so people find prompts by task
5. Announce it and ask each teammate to contribute one prompt this week

Tool

ClaudeNotion

Artifact

Team Prompt Registry doc

Ship

A shared prompt registry seeded with your best prompts and open for contribution.

Concept lockPersonal tools help you, shared tools make you the standard.

WhyA personal library helps one person, but a team registry makes every colleague faster and quietly positions you as the person who built the standard. Shared infrastructure is how a private habit becomes something the whole team runs on.

Measure Team Adoption

Task

Build a snapshot of how widely the team is actually using the tools and prompts you introduced, so you can see what stuck and what needs a push. Adoption is the real measure of whether you changed anything.

1. List the tools, kits, and prompts you have put in front of the team
2. For each, gather a rough measure of who uses it and how often
3. Ask the model to turn it into an adoption snapshot with obvious gaps
4. Flag the one high-value tool with the lowest uptake
5. Draft a single targeted nudge to lift that one tool next week

Tool

ChatGPTWeb

Artifact

Adoption Snapshot sheet

Ship

An adoption snapshot showing usage per tool and the one gap worth closing.

Concept lockIntroducing a tool is easy, adoption is the only proof it mattered.

WhyIntroducing tools is easy, but only adoption proves you shifted how the team works rather than just adding noise. Measuring it tells you where to push and gives you the evidence that your influence is real and spreading.

Write Your New Role Description

Task

Write the honest, forward-looking description of the role you now occupy, one that would not have existed thirty days ago. This is you defining your value in a language the org has not caught up to yet.

1. Pull together the playbook, the registry, the initiative, and the adoption data
2. Draft a role description centred on orchestration, systems, and judgment
3. Ask the model to strip out anything AI could do and keep only what needs you
4. Contrast it plainly with the role description you started the month with
5. Refine it into something you would put your name to in a review

Tool

ClaudeWeb

Artifact

New Role Description doc

Ship

A role description built on orchestration and judgment that did not exist 30 days ago.

Concept lockDefine your value before the org defines a lesser version of it.

WhyThe safest position in this shift is a role defined by what only a human orchestrator can do, and no one is going to write that description for you. Naming it yourself is how you make your value legible before the org invents a lesser version of it.

The 30-Day Review and Next 60

Task

Run a full review of the thirty days, honestly assessing what changed in your work and your standing, then chart the next sixty. The month was the proof of concept, the next quarter is where it compounds.

1. Gather every artifact and metric from the month in one place
2. Ask the model to assess your move from AI user to orchestrator against the evidence
3. Name what genuinely changed and what was only motion, honestly
4. Set three concrete goals for the next sixty days that build on this
5. Write the review as a document you would revisit in two months

Tool

ClaudeWeb

Artifact

30-Day Review doc

Ship

An honest 30-day review with three concrete goals for the next sixty days.

Concept lockA month only compounds if you name what it earned and where it points.

WhyA month of work only compounds if you stop to name what it earned you and decide where it points next. This review turns thirty days of scattered wins into a clear trajectory, so the momentum outlives the challenge that created it.

Reflect before moving onWhat genuinely changed in your work versus what was only motion?  ·  Which single artifact from this month will still matter in a year?  ·  What is the one thing you must not let slide over the next sixty days?