A few tools serve more than one step (for example the Control Tower KPI sheet and the Say vs. do review), so they show up under each step they help.
Personal SIPOC
A one-page map of Suppliers, Inputs, Process, Outputs and Customers, kept alive like a will or a living trust: reviewed on a set schedule and updated as life changes.
- In my system
- Phase 1 · A living map. Our life's work on one page; the process column runs capture → triage → decide / route → do → file & improve. The quality column holds each output's CTQ, and quality and delivery measures become the requirements behind our notifications and the daily newspaper.
- What AI adds
- Drafts a first SIPOC from our real email and calendar, then checks that every output serves a named customer.
- Guardrail
- We review it quarterly and update it as life changes.
Stakeholder analysis
Once you know who you serve, map what each one needs from us, how often, and how we keep them in the loop.
- In my system
- Phase 1 · the grid for family, spouse, work team, close friends, guests and myself. It doubles as a spec for the helpers: the “how we keep them in the loop” column is a list of jobs a specialist can draft.
- What AI adds
- Drafts the first version from our real calendar and inbox, and turns the loop-in column into drafts we can approve.
- Guardrail
- The people are ones we love. AI suggests; we decide what each one needs.
Voice of the Customer → CTQ tree
Listen first: write what people would actually say (their feedback and what we notice). Then write the CTQ, critical to quality: a checkable standard for “good” that lives in the SIPOC's quality column.
- In my system
- Phase 1 · the stakeholder grid (family, spouse, work team, close friends, guests, myself) and a CTQ for each. Voice of the Customer and CTQ are related, but they're two different steps.
- What AI adds
- Turns “be more present” into a CTQ we can check, and keeps the list in context so every draft knows who it's for.
- Guardrail
- The customers are people we love. The AI suggests; we decide what “good” means.
P · PlanLean / TPS · strategy deployment
Macro plan (Hoshin kanri)
Cascade a few big goals down into yearly themes, quarterly rocks and weekly actions, checking each level with the one above (the catchball).
- In my system
- Phase 0 · Future self → macro plan (3–4 year-level themes) → 90-day plan → the week.
- What AI adds
- Reads my future-self lines before it helps with anything, and flags a rock that can't name the theme it serves (the catchball).
- Guardrail
- Values stay private; the AI reads a short, shareable version.
Baseline & data collection plan
Decide exactly what you'll count and how (operational definitions), then measure a normal stretch before changing anything.
- In my system
- Phase 3 · Capture, plus two normal weeks of calendar and inbox data (Callie's time audit) as the baseline.
- What AI adds
- Measures quietly in the background (calendar tags, inbox counts, routine heartbeats), so the baseline builds itself.
- Guardrail
- Write the definition before you look at the number.
Also: the first Control Tower reading and a first say-vs-do look are baselines too.
A3 brief & mind map
Put a problem or project on one page (background, current state, goal, plan, check), after mapping it out.
- In my system
- Phase 8 · Deep-work lane: map a one-off, then hand Claude the brief, not a blank prompt.
- What AI adds
- Turns the map into a working brief and does the deep work, and the session ends with a finish line.
- Guardrail
- The “keep out” branch: private details stay out of the brief.
Note: an A3 can stand in for a whole plan-do-check-act cycle on one page, since it's both a planning and a root-cause-and-fix tool.
D · DoLean · workplace organization
Virtual 5S
Sort, set in order, shine, standardize and sustain, so everything has one home.
- In my system
- Phase 2 · Information homes: one cloud source of truth, a few lanes, a naming standard and the Downloads rule.
- What AI adds
- Runs the weekly shine nudge, suggests where stray files go, and can finally find things because they live in one place.
- Guardrail
- Nothing is deleted without my yes.
Standard work
Write down the current best way to do a job: when, what, and how you know it's done.
- In my system
- The daily loop as a standard work card. Once a step stops changing, a routine can run it.
- What AI adds
- Runs the steps that have a clear “done when” on a schedule, every time, without being asked.
- Guardrail
- Only quiet, reversible, pre-approved steps run on their own.
Poka-yoke & jidoka
Poka-yoke designs the process so the mistake can't happen. Jidoka stops and signals when something's off, so a person can step in right away.
- In my system
- Nothing sends, spends, books or deletes without my yes; specialists produce drafts, not sends. When a specialist is unsure, it stops and puts the item in “Needs you”, and when a routine breaks it fails loud.
- What AI adds
- Makes draft-and-hold the default in every lane, handles the normal case, and stops at the edge case instead of guessing.
- Guardrail
- If it leaves the house, it waits for a yes. When unsure, it goes in “Needs you.”
Kanban board & WIP limits
A visible board of work by status, with a limit on how much can be in progress at once.
- In my system
- “Waiting on you” is a kanban column; a small limit keeps drafts from piling up on me.
- What AI adds
- Keeps the board current on its own, and holds new drafts back when the limit is hit.
Same idea, different level: the Level-week planner. Cross-reference: Visual management & andon (Act), since a kanban board is a visual management board.
Level-week planner (Heijunka)
Spread work evenly instead of in spikes, so the system never overloads.
- In my system
- A level week: deep work, admin and people time spread across the days, with one day kept free.
- What AI adds
- Callie suggests where things fit, spots overloaded days, and protects the free day.
- Guardrail
- Proposes moves; never reschedules people without my yes.
Same idea as the Kanban board and other visual boards, at a different level: the week instead of the day-to-day items.
D · DoProject & process management
RACI chart
For each task, name who's Responsible, Accountable, Consulted and Informed.
- In my system
- Phase 5 · every lane has a row; I'm always Accountable for anything that leaves the house.
- What AI adds
- Specialists can be Responsible for the work, and route updates to whoever is Consulted or Informed.
- Guardrail
- Specialists own the work, never the yes.
FMEA-lite
For each step, ask what could go wrong, how bad it would be, how likely it is, and what guardrail prevents it.
- In my system
- A short risk list for every agent: wrong tone, a missed reminder, a stalled routine. Each one has a guardrail.
- What AI adds
- Helps list the failure modes, and the guardrails (poka-yoke, jidoka, heartbeat) run automatically.
- Guardrail
- Re-check it when a new bot joins the team.
Control Tower KPI sheet
One page that tells us whether the system is working: three to six measures, each tied to a theme, reviewed weekly.
- In my system
- Phase 6 · the weekly Control Tower, read in the Saturday week-ahead. We can see where we are and take action on anything that isn't performing the way we expect.
- What AI adds
- Checks every day so we only look weekly, and turns a tile red only when a decision is truly needed.
- Guardrail
- Red means a decision, not a guilt trip.
Also in Plan: the first reading, taken before we change anything, is our baseline.
Flow metrics
Measure how smoothly work moves: cycle time, process time and lead time, plus how much is stuck in progress.
- The three clocks
- Cycle time: how long one task takes from start to done; a long one can point to a bottleneck. Process time: the time we actually spend working on it. Lead time: from the ask to delivery; a long one can show where delays hide.
- In my system
- A simple read on the household to-do list: how long items take to close, and how many roll into next week. They roll up to the Control Tower.
- What AI adds
- Timestamps every action automatically, so the clocks are just there, no stopwatch.
Say vs. do review
Compare what you say matters with what actually happened, then pick one small change.
- In my system
- The monthly coach review: my stated priorities (Faith first) against where the calendar time went.
- What AI adds
- Brings the facts, one observation and one experiment, with no judgment and no grade.
- Guardrail
- Facts, not judgment. One experiment at a time.
Also in Plan: a first look is our baseline. Also in Act: it closes the loop on the experiment we picked.
Gemba / waste walk (8 wastes)
Go look at the real work, map each step, and tag the waste: Defects, Overproduction, Waiting, Non-utilized talent, Transportation, Inventory, Motion, Extra processing (DOWNTIME). It has its own printable sheet.
- In my system
- The daily loop as a value stream (capture → triage → decide → do → file), walked with real screenshots, not assumptions.
- What AI adds
- Pulls the facts (a real month of email, calendar, open loops) so the walk is about what happened, not what I remember.
- Guardrail
- Look at redacted real data; keep private details out of anything shared.
Pareto tally (80/20)
Tally causes and sort them biggest first; usually a vital few make most of the pile.
- In my system
- Which few senders fill most of the inbox, and which few tasks eat most of the week.
- What AI adds
- Does the tally in seconds across a month of email or calendar, and proposes the rule that handles the top few.
- Guardrail
- Proposes filters and unsubscribes; I approve them.
5 Whys
Ask “why” until you reach a cause you can fix, and stop when the answer is a process, not a person.
- In my system
- When something slips (a reminder, a guest reply), we find the root cause instead of trying harder.
- What AI adds
- Asks the next why, gathers the evidence for each answer, and drafts the countermeasure.
- Guardrail
- The AI can suggest causes; I choose the root cause and the fix.
RCA: pairs with Fishbone / CEDAC to dive deeper into causes.
Fishbone / CEDAC
Sort the possible causes by category on a fishbone. CEDAC takes the fishbone to the next level.
- In my system
- When causes pile up, we sort them by method, tools, inputs, people, timing and environment, then circle the likeliest and test it with 5 Whys.
- What AI adds
- Suggests possible causes and gathers the evidence for each.
- Guardrail
- The AI can suggest causes; I choose the root cause and the fix.
RCA: CEDAC extends the fishbone and pairs with 5 Whys to dive deeper into causes.
Control plan & signal vs. noise
For each key measure: the normal range, who checks it, and what to do when it's off. Plus the control-chart idea that one odd week is noise and a run is a signal.
- In my system
- The Control Tower measures, each with a target, an owner (a specialist) and a reaction plan.
- What AI adds
- Checks every day so I only look weekly, and flags real signals instead of every wiggle.
- Guardrail
- Reactions that leave the house still wait for a yes.
Visual management & andon
Make status visible at a glance, and give problems an obvious signal (the andon light).
- In my system
- Phase 6 · Visual home: the “Waiting on you” card, green / amber / red, and the Control Tower.
- What AI adds
- Keeps the board updated and turns a tile red only when a decision is truly needed.
- Guardrail
- Red means a decision, not a guilt trip.
Cross-reference: the Kanban board in Do is a visual management board too.
Leader standard work & tiered rhythms
A fixed rhythm of daily, weekly and monthly check-ins that keeps the system honest.
- In my system
- Phase 7 · Rhythms: morning brief (daily), Saturday week-ahead (weekly), coach review (monthly).
- What AI adds
- Runs the briefs on the clock and brings the right numbers to each tier.
- Guardrail
- Sundays are for being present, away from screens. The plan waits.
A · ActDeming · continuous improvement
PDCA feedback log
Every time we edit or turn down a draft, we jot what changed and what the system should keep, and carry that lesson into the next Plan. PDCA is kaizen at its core.
- In my system
- The draft-and-hold loop: my edits and no's become style notes, rules and memory; new routines get a two-week pilot first.
- What AI adds
- Captures what I changed and learns from it, so the next draft needs less from me.
- Guardrail
- Learns preferences; it never learns to skip the yes.