The Method End to End

This is Unit Nine of The M-CaReM-AI Method, the method end to end. Units Four to Eight took the phases one at a time. This unit runs all five on one real-world case, then hands the method to you in a four-part capstone on your own work. Carry one question through this unit: what does the whole method look like when it is reduced to its smallest form? The answer arrives with five questions to ask at every step. By the end you will be able to follow the hospital case through all five phases. To state the lessons its team drew. To apply the method to any field. And to carry out the capstone on your own work.

Eight entries on the book's hospital case, its team, and the lessons it drew.

The case's hospital.

MCM 201 The Case Study's Hospital

The book's end-to-end case follows Greenfield Medical Center, a composite hospital, replacing its incident reporting system with the method: a twelve-year-old application, manual regulatory reports, and leaders keen to try AI drafting. Course One covered this case's AI lessons; this unit follows the method.

The team.

MCM 202 The Team and the Goal

The case team spans product, analysis, quality, safety, nursing, data, testing and AI engineering. The goal is a system that is easier to use, compliant, and ready for safe AI help. The team keeps its stories and backlog, with things, cards, Packs and matrices underneath.

Names matter.

MCM 203 Names Shape the Work

In the first workshop, the team names a zone Action, until the quality manager insists on Improvement and Follow-Up, because names shape how people think about the work. The words on the map become the words in meetings.

The Pack reveals a card.

MCM 204 Writing the Pack Reveals Cards

Drafting the Incident Report's Pack, two clinicians add a state no tool could have known, returned for more information, and with it a new card. The book's general lesson: writing the Pack reveals missing cards.

The catch that mattered most.

MCM 205 Usually True Somewhere

In validation, the safety officer stops at a drafted rule, all severe incidents are reportable. It depends on the type and policy version; the rule came from generic documentation. Over-general rules are dangerous because they are usually true somewhere; verify them against their source.

What the team learned.

MCM 206 Lessons: Language, Packs, Cards

The retrospective found that the language of things stuck, Packs reduced re-explaining, and cards sharpened stories and tests. Validation also changed the tone: disagreements became about definitions and states, more concrete and less personal.

More lessons.

MCM 207 Lessons: Matrix, AI, Gold

The retrospective also found that the matrix proved its worth in the first policy change, not everything needed Gold, and AI sped up the method's mechanics but was no substitute for domain judgment. The book says both halves of that sentence are the lesson.

Prompts by phase.

MCM 208 The Method's Prompts, by Phase

The book's Toolbox gives a prompt pattern for each phase, from discovering things to scouting impacts. Every pattern supplies the thing first, asks for one job, and forbids inventing fields, rules or states. Course One teaches the prompting technique.

Seven entries that walk the hospital case through the five phases, then generalize it. The opening question is answered here.

Phase one in the case.

MCM 209 The Case: Mapping

The case's mapping workshop runs ninety minutes: a noun storm of things, five zones, a rough map from the Incident Report through investigation to action plans and tasks, and five core things. An AI narrative of the map adds an invented risk-score field, which the analyst deletes. The safety officer calls the map better than the last forty-page specification he saw.

Phase two in the case.

MCM 210 The Case: Cards

The case drafts the create card with AI, then three people correct it: the hospital's own harm scale, a rule flagging suspected malpractice, and permanent created metadata. The other critical cards follow. The change-status card gathers rules once spread across a specification, a diagram and an old email.

Phase three in the case.

MCM 211 The Case: The Pack

The case's Pack is drafted with AI help, then clinicians add the returned-for-more-information state and its card. The team makes it Gold, because it is central to safety, sensitive for regulators and likely to be touched by AI, while shift schedules stay Bronze or Silver.

Phase four in the case.

MCM 212 The Case: Validation

The case's review session widens the definition to cover trend reports, lets harm level start as unknown with a card for updating it, adds a reportable flag, and rewords closed to mean the safety review is complete. And it catches the over-general reporting rule, rebuilt from type, harm level and policy version.

Phase five in the case.

MCM 213 The Case: Tracing a New Rule

The case's matrix links policies, nine cards, data, screens, reports, tests, procedures and the AI drafting prototype. A month later, the state makes a new medication error reportable even at no harm, and the team walks the matrix row by row, through to the drafting prompt. It still had to do the work, but now it had a map.

The pattern for any field.

MCM 214 Translating the Case to Your World

The book's eight steps for any field: start with one domain; identify three to seven core things; map them lightly; write cards for one anchor thing; build its Silver or Gold Pack; validate it with real stakeholders; build one matrix for a high-risk thing; and use AI to draft and speed up while people judge what is true.

Five questions at every step.

MCM 215 Five Questions at Every Step

The book gives five questions for every step: what is this thing in our world; what change are we making to it; where else does it show up; what happens when policy or technology changes; and how can AI help without defining our world for us. That answers this unit's question: in its smallest form, the whole method is these five questions, asked in order whenever a request, a design or a change arrives.

Five entries: the four capstone tasks on your own work, and the close.

Capstone one.

MCM 216 Capstone One: Your Inventory and Map

The first capstone task, constructed for this course, takes one area of your own work through phase one. Interview yourself or colleagues with the four questions, including the auditor question. Build an inventory of ten to thirty rows, name zones in your own words, draw arrows with verbs, and highlight core things. It is finished when someone who does the work says that is basically our world.

Capstone two.

MCM 217 Capstone Two: Three Cards

The second task writes three full cards for your anchor thing, at least one that is not a create card. Run the six checks on each, record anything an AI draft invented or missed, and show one card to someone who does the work. The stakeholder check turns a well-formed card into a true one.

Capstone three.

MCM 218 Capstone Three: Your Pack

The third task builds a Pack for your anchor thing, Silver unless a mistake would cost money, safety or rights, then Gold. Build it in the book's order, use AI only on your own pieces, check its change log both ways, and name a steward. Most of the Pack should come from your first two tasks; one you invent from scratch is guessing.

Capstone four.

MCM 219 Capstone Four: Review and Matrix

The fourth task validates your Pack with real people, using at least one technique for surfacing disagreement, then builds a matrix with each row's owner and walks one realistic change through it into a plan. If nothing is triggered, check the matrix. Write two sentences: what the review changed, and what the matrix found you would have missed.

The whole method.

MCM 220 The Whole Method

The closing entry puts the method back together: an inventory and map for what the world is made of, cards for what is done to it, Packs for where each thing is kept, people for whether it matches reality, and matrices for what else must change. At every step AI drafted and suggested, and at every step the frames caught its errors. The method's smallest form is still the reflex from Unit One: which thing is this about, and what is that thing made of?

Look at what you can now do. You can follow the hospital case through all five phases, from the noun storm to a traced regulatory change. You can state the lessons its team drew: the language of things sticks, and AI speeds the mechanics but never replaces judgment. You can apply the eight steps and the five questions to any field. And you can carry out the capstone on your own work. Start with one thing, one card and one Pack, and grow the method from there.