Skip to main content
Thierry Gilgen

Who owns the work?

Read ownership as an argument you can test. A named owner needs actual authority; a well-orchestrated workflow can still optimise the wrong objective. The following sequence places the practical handover beside that objection.

  1. Thesis

    AI strategy without operational ownership is theatre

    Slide decks do not run inference, govern data, or absorb incident load.

    7 · Current edition 1

  2. Extension into handover

    Who owns the AI system after the pilot?

    An ownership map for decision rights, monitoring, escalation and exit once real work depends on the system.

    45 · Current edition 1

  3. Operating model

    AI-First Is an Operating Model

    Seats, handoffs, and how to read the agent bill

    40 · Current edition 1

  4. Counterargument about objectives

    The optimisation trap

    The biggest lesson from autonomous AI isn't that it becomes conscious. It's that sufficiently capable systems optimise objectives—not intentions. Why Goodhart's Law is becoming one of the defining engineering challenges of the AI era.

    5 · Current edition 1

  5. Practical application

    Stop Building an AI Org Chart. Build a Workflow.

    The practical Grok Bot playbook: five role prompts, a first assignment, a handoff contract, and twelve mistakes to avoid.

    41 · Current edition 1

Published explanations

These directed connections retain the editions used in their explanation. Current articles can change; follow the cited edition for the earlier argument.

Who owns the AI system after the pilot? · extends · AI strategy without operational ownership is theatre

The pilot essay extends the strategy argument into the handover: named ownership must survive the departure of the demonstration team, with operating authority and incident duties assigned.

Source edition 1Target edition 1

The optimisation trap · challenges · AI-First Is an Operating Model

The optimisation argument challenges an AI-first workflow to define who sets its objectives and checks its incentives. Smooth orchestration can scale the wrong goal as efficiently as the right one.

Source edition 1Target edition 1

Stop Building an AI Org Chart. Build a Workflow. · applies · AI-First Is an Operating Model

The workflow playbook applies the operating-model argument through a concrete brief, evidence artefact, handoff and review. Its warning that prompts do not create access controls qualifies the seating plan: assigned remits still require actual permissions and approval boundaries.

Source edition 1Target edition 1

Decision to take away

For one workflow, record its owner, deputy, stop authority, acceptance check and unresolved objective. Keep the competing argument in your packet rather than selecting only passages that agree with you.

Selected context packet / decision dossier

Choose up to five volumes. The packet contains selected extracts, edition identities, recorded sources and objections; follow links for complete articles. Preview exactly what you take away.

These fields stay on this page. Nothing is sent to an assistant. Private material is excluded unless you select the option above.