Start an assessment
Start an assessment

Chapter VI · Begin

dagg.ai · 07

The first engagement is a decision, not a project.

A torn edge of paper crosses a bone ground; one small red fragment rests on the edge in raking light.
Plate · The first pathOne path, taken from the whole. One mark where the decision will be made.

An assessment answers one question. In one consequential path through your company, what should change, and what, if anything, should be built. We learn enough to agree the path. We map it from the traces the work already leaves. We decide, one verdict per step. We recommend the smallest justified change, written by the same people who would build it, including, when that is the honest answer, keeping the work as it is. You leave with a Decision Pack. It is yours, with us or without.

Plate VIII

The Decision Pack.

What you leave with. Six parts, the same for every assessment. Yours, with us or without.

  1. 01LearnEnough to understand the operating question and agree the path. Data and access boundaries first.
  2. 02MapThe visible and hidden work, the handoffs, the waiting, the exceptions, the decisions and their owners.
  3. 03DecideOne of five verdicts for every part of the path, with the reasoning attached.
  4. 04RecommendThe smallest justified change, written by the same people who would build it. Including keeping the work as it is.
  5. 05Decision PackDelivered, walked through, yours.
Decision PackSix parts. Yours, with us or without.Delivered, walked through
  1. 01 Operating map One path, as it is: the visible steps and the hidden ones, the handoffs, the waiting, the exceptions, the decisions and their owners.
  2. 02 Recommendation One of five verdicts for every step, with the reasoning attached.
  3. 03 Intervention brief The smallest justified change: what would be built, if anything, and its boundary.
  4. 04 Accountable owner Who owns the outcome, by name.
  5. 05 Evidence boundary What the map rests on, and what it does not.
  6. 06 What would change the decision The evidence that would send it the other way.

Fig. 8 · Six parts. The same for every assessment.

A good starting point

A good starting point has consequence, an owner and evidence.

Evidence comes from the work, not from a workshop: the traces the path already leaves, and a small number of conversations with the people who carry it.

A good fit

  • The path matters to a customer, a margin or a regulator.
  • Someone owns the outcome and can decide.
  • The work leaves traces: tickets, records, messages, system events.
  • You are open to the answer being “keep this as it is” for part of the path.

Not yet

  • The question is “what should we do with AI” in general.
  • No one can decide on the path.
  • The work exists only in conversations and nothing can be measured.
  • A build has already been decided and needs a justification.

Begin

Begin with one path.

One consequential path is enough to begin. Describe it in a few sentences. We reply to propose a scoping conversation. There is no commitment beyond that conversation.

hello@dagg.ai

An order, a claim, a release, a hire — the path you would look at first, and why.

Dagg uses this information only to respond to your inquiry. Privacy

The end of the book. The beginning of the work. Dagg AI AB · Stockholm