What is Airecting?
Airecting is the practice of directing artificial intelligence across disciplines. It begins with purpose and audience, assigns AI appropriate work, evaluates the result and keeps human intention, judgment, authorship and responsibility explicit.
Definition by Henrik Napoleon Dahnsjö · Founder of Airecting · Updated 14 August 2026
The definition.
Airecting is to AI what directing is to a film: holding the intention across many capabilities while remaining responsible for the whole.
The Airector does not need to perform every specialist task. AI may research, write, calculate, design, code, compare and operate. The Airector decides what the work is for, what standard it must meet, which constraints matter and what should happen next.
How is Airecting different from prompting?
A prompt asks for an output. Airecting holds a direction across an entire sequence.
Prompting is one technique inside Airecting. A real project also requires framing, source selection, role assignment, verification, trade-offs, iteration, integration, authority limits and accountability for the final result.
What should the model produce now?
Why should this exist and for whom?
Which capabilities and decisions come next?
Who judges and owns the outcome?
Is Airecting the same as AI consulting?
Airecting can include AI consulting, but it describes the method rather than a single service category.
Consulting may diagnose a problem and recommend action. Airecting can continue from exploration into prototyping, agent-system design, implementation and operation. It is medium-agnostic: the outcome may be a decision, product, workflow, venture, film, document or AI-operated organisation.
Who is an Airector?
An Airector is the human who holds the intention and directs AI capabilities toward a responsible outcome.
The term was developed by Henrik Napoleon Dahnsjö from more than 25 years of directing film. The central observation is simple: a director does not operate every camera, shape every light or play every role. The director coordinates specialists and remains accountable for what reaches the audience.
What can be Airected?
- An idea that needs to become a testable concept.
- An AI strategy grounded in business outcomes.
- A prototype, product or internal tool.
- An AI agent or multi-agent workflow.
- A creative work involving several AI capabilities.
- An AI-operated venture with financial and human controls.
The method is not defined by the medium. It is defined by direction, sequence and responsibility.
What does Airecting not mean?
It does not mean allowing AI to make every decision. It does not mean disguising generated work as human labour. It does not mean automating a process simply because automation is possible. And it does not remove legal, financial or ethical responsibility from the human owner.
Direction remains human.
AI widens the field. Airecting determines what deserves to become real.
See how Airecting works →