Training and workshops

The workshop is three days, and each ends in one signed document. On day one the leaders who can change the operating model decide on three teams and three use cases. On day two those teams build something small and release it before 16:00. On day three what the first two days found becomes a plan for the next increment, with governance around it. We sell the days a firm can complete: day one alone, days one and three, or all three.

Before the workshop

The workshop does its best work behind a diagnostic and a selection. In the diagnostic, each discipline owner says, practice by practice, whether what the practice describes happens nowhere, somewhere or everywhere in the firm, and every “somewhere” carries the team where it was seen. The selection then pairs teams with use cases and orders the pairs into a pipeline the sponsor agrees. With both done, day one argues with a list rather than building one.

Before day two, the firm loads what it already holds into the record: its structure, systems, business processes, capabilities, policies and finance model. A readiness gate then reports, area by area, whether teams may start. It blocks nothing; where an area is incomplete, the sponsor decides whether to collect it, defer it to a date, or start anyway, and the decision is recorded.

The three days

Each day runs from 09:00 to 16:00, and its last half hour is a line-by-line read-back of the document, corrected live and signed before anybody leaves.

Day one Choose and commit

From awareness of AI to a signed decision to proceed with three real teams.

Objective

Confirm three pairings of team and use case, each with an accountable owner and a business expert released to it, and record the decision each leader will make to let the work proceed.

Who attends

The leaders who can change the operating model, among them the sponsor and the owner of each pairing; and, where they are already known, the people who will carry each pairing.

Needed beforehand

Ideally the diagnostic and an agreed selection, though neither is required. At least one leader present must be able to release a person by name.

Produces

The Mandate and Priorities Record, one page: why value is not landing here, the three pairings and their owners, each leader’s decision and its date, signed by the senior leader present.

The day assumes little knowledge of AI. It opens by showing a model complete a task and then fail on a near-identical one, then turns to why money spent on AI is not producing value. After lunch, where a diagnostic and selection have run, each proposed pairing is put with its evidence, and its owner confirms that the use case is real, painful and measurable, or amends it. Where neither has run, the leaders learn the pairing method and choose three under a forced ranking, and the record states that the three are untested. Each pairing leaves with a business expert to be released from their previous job rather than lent, because a lent expert is the first thing reclaimed when the day job calls. Each leader then writes down the one decision they must make in the next fortnight, such as releasing a person, funding or granting access, and its date.

Day one needs nothing after it. A firm that buys only this day holds a decision with owners and dates that any delivery partner can work from.

Day two Build and release

Three teams, each with the framework installed and one thin slice released by 16:00.

Objective

Install the scaffold for each confirmed team and have it ship one small slice of its use case as a demonstration release.

Who attends

The three teams, each with its five roles filled (product, domain, integration, quality and engineering) and its domain role held by the business expert released to it.

Needed beforehand

Day one, with its pairings settled. The readiness gate read and put to the sponsor. A route by which work can reach an environment within a day.

Produces

A Scaffold and First Iteration Record per team: its members, the slice that shipped, the test that passed and where its evidence sits, its working agreement for the increment and the one thing it learned, signed by its lead.

Two of the six working hours are spent building, and that proportion is the day. Each team watches the framework installed and its first checks fail against an empty record. It cuts its slice to the thinnest thing that runs end to end on the real system and writes the test before the build: the domain role says what correct looks like and the quality role turns that into a test. Each team then builds at one screen with an AI coding assistant, and the slice is done when its tests pass and the evidence is recorded, not when it looks done. Finally each shows the other two its running slice and says what would have to be true to ship every week.

Day two cannot run without day one, because a team cannot form around a use case nobody has confirmed, and it cannot run where every change waits for a weekly approval board.

Day three Plan and govern

Three teams become a programme with a roster, decision rights and release gates.

Objective

Agree the increment and its roadmap, write the decision rights and release gates, and give every supporting process an owner.

Who attends

The engagement lead and the leaders’ direct reports who will be the discipline owners.

Needed beforehand

Day one. Day two strengthens the day but is not required.

Produces

The Increment Plan and Governance Agreement: the increment’s length and weekly cadence, the sequenced plan, the roster with owners, decision rights with one decider each, the release gates and their owners, the four metrics and who reports each, what the change ends and by when, and the confirmed request for money, roles and decision rights, signed by the accountable leader.

The firm chooses the increment’s length, with a quarter suggested: three teams, one consultant, weekly iterations. The plan is sequenced so the increment tests the operating model where the firm is weakest, rather than having three teams repeat the same easy thing. The selection runs again, now with building evidence behind it, and the roadmap it produces may list further teams and further workshops, which is the ordinary outcome. Risk and finance are placed inside the teams’ cadence and the release path rather than at a late gate. Before the signature, every process, gate and metric must carry a person’s name, and none may be the consultant’s.

Day three runs without day two. Its governance then draws on the firm as it already runs rather than on the teams’ working agreements, and we say on the day that this is a weaker input.

Combinations

Day two needs day one, and day three needs day one but not day two. Within those rules the days can be run together or apart.

Three days in sequence

The fullest form. Day three plans against two days of evidence, including each team’s own statement of what it would take to ship weekly. The condition is that the leaders’ decisions from day one, above all releasing the business experts, are made before day two rather than within the fortnight day one allows.

Day one, then the other two

The gap gives the leaders their fortnight to make the decisions they signed for, and the firm time to complete setup so the readiness gate can be read before any team starts. The signed record keeps the pairings, owners and dates from being reopened in the interval.

Days one and three, without day two

A firm that cannot yet ship still gets a decision and a governed plan. It loses the evidence: day three designs governance against what the leaders believe about their firm rather than against what three teams found by building.

Day one alone

The smallest purchase. Without a diagnostic behind it, the pairings are recorded as untested, and the diagnostic should follow rather than be skipped.

Day two is never sold on its own, only as part of a full programme that uses the discipline skills in teams and takes as its input the findings of a diagnostic and a use case selection already run.

The nine disciplines the workshops teach →