See the whole set.
A director knows a little about every department and stays curious about all of them. This module shows you how to build that wide view fast, and how to use it to point AI in the right direction and catch it when it is wrong.
- Production
- The wide view
- Director
- You
- Crew
- AI
- Take
- 01
- Scene
- 03
- Part
- I
By the end of this module you will be able to:
- Describe the five departments a director needs to know enough about.
- Explain the wide-across, deep-in-one shape and why it matters with AI.
- Build a working map of a new field in 15 minutes.
- Apply the "good enough to steer" test and know when to bring in an expert.
This module follows the guide section by section. Keep it open as a one-page reference while you work.
Know a little about every department
A director does not light the set, record the sound or cut the film. They know enough about each job to ask for the right thing and to notice when something is off. That wide view is what lets them run the whole production.
Directing AI works the same way. You do not need to master every field it touches. A general map of the work and the habit of asking how things work are enough to point AI in the right direction. Then you go deeper where it matters.
| Department | Know enough to | Ask AI |
|---|---|---|
| WritingScript | Tell whether a draft makes one clear point for one reader. | "What is the one message here, and who is it for?" |
| VisualsCamera | Say what the viewer should notice first, and what should fade back. | "What would a designer change in this layout, and why?" |
| StructureEditing | Spot what is missing, repeated or in the wrong order. | "Show me the outline before you write the full draft." |
| NumbersProduction | Ask where a figure came from, and whether it is plausible. | "Show your working, and list the source for each figure." |
| Rules and riskClearances | Recognise private data, and claims that need someone's approval. | "What in this should a person check before it goes out?" |
| The whole setThe director | How the departments fit together, and who signs off each part. | "What does the whole production need next?" |
Wide across, deep in one
Picture a capital T. The bar is the director's view: enough of each field to brief it and to check it. The stem is your own craft, where you can judge quality in depth.
You already have the stem. The bar is what lets you direct AI outside your own lane, and it is also what protects you when AI is confidently wrong.
In a Harvard Business School and BCG field experiment with 758 consultants, AI helped on tasks it handled well. On a task outside its abilities, consultants using AI were 19 percentage points less likely to reach the correct answer. Knowing enough to spot a wrong answer is part of the job.
Dell'Acqua et al., "Navigating the Jagged Technological Frontier", 2023Build a working map in 15 minutes
New field? Get the general picture first, with AI as your teacher. Five steps, about three minutes each.
- Map"Explain how this field works to a smart newcomer."
- Words"Give me the ten terms a professional here uses."
- Checks"What do experts check first? What do beginners miss?"
- ExplainSummarise it in your own words. Ask AI to correct you.
- BriefUse the new words to set the vision, limits and checks.
Here is the method used by a marketer asked to write a one-page explainer on sign-in security.
- Map
- Asked AI how multi-factor sign-in works, and where each method can fail.
- Words
- Learned "push fatigue" and "phishing-resistant".
- Checks
- Asked which methods experts trust, and which they avoid.
- Brief
- "For office managers. Compare methods by strength. One page."
- Review
- The security lead checked it before it went out.
The Explain step matters most. Saying it back in your own words is how you find the parts you have not understood yet.
Five curiosity habits
Curiosity is a set of questions you can ask on purpose. Count the ones you used this week. 0 to 1: try one today. 2 to 3: map your next field first. 4 to 5: teach your team.
- Ask how it works. Knowing the steps shows where AI can help.Try it: "Walk me through how this is done."
- Learn the basics. A short lesson first makes the brief better.Try it: "Teach me the basics before I brief you."
- Ask for reasons. The reasons behind an answer show what to check.Try it: "Why this? What did you assume?"
- Find blind spots. The best question is often one you did not ask.Try it: "What would an expert ask me that I have not?"
- Ask the people. Colleagues know how things work here. AI does not.Try it: ask a colleague what they check first.
The "good enough to steer" test
You do not need to know how to do every job. You need enough to direct it and to check it. Before you brief, confirm four things.
- I know what good looks like.
- I know the key terms.
- I would spot an error.
- I know who checks it.
Go deeper yourself when
- You will do this work again. A deeper map pays off.
- The answers keep missing. Learn one level more.
- You cannot tell good from bad. Ask AI for a strong and a weak example.
- It is your own craft. This is your stem. Keep it deep.
Bring in a real expert when
- Mistakes are costly. Legal, financial, safety and security calls go to the people accountable.
- You cannot check the answer. Find someone who would notice an error.
- The facts are yours. Your clients, systems and history live with your colleagues.
Curiosity is no licence to skip review. Someone qualified approves the final cut.
Seek a different view, then decide.
The fifth curiosity habit is the leadership one: ask the people. AI will give you a fluent answer from its general knowledge. Your colleagues will tell you what happens here, and they will often disagree with you. That disagreement is valuable.
Good leaders go looking for the view that differs from their own, listen to it properly, and then make the call. Seeking dissent is different from handing over the decision. You still decide, and you decide better.
- Before a big brief, ask one colleague who sees the work differently what they would check first.
- Ask AI to argue against your plan, then weigh both sides.
- When you decide, say what you heard and why you chose as you did.
A physicist's way of seeing work as a system
I hold a BSc in Physics and a Diploma in Engineering. I am not a software engineer. That training taught me to look at complicated work as a system that can be understood, redesigned and improved, which is the wide view this module describes. On a roughly 300-page proposal that won a 20-year contract, my job was to coordinate content and assembly across a 15-person team of executives and subject experts. I needed enough of each field to brief it and check it, and I relied on the experts for depth.
- Education
- BSc in Physics, Diploma in Engineering, Mount Allison University
- Proposal
- About 300 pages, 15-person team, 20-year contract won
Map a field you need next
Choose a field you will have to brief AI on soon. Spend 15 minutes on the five steps and record what you find.
Check your understanding.
Module 3 in four lines.
- A director knows enough of each department to brief it and to check it.
- Wide across, deep in one. Your craft is the stem, the wide view is the bar.
- Fifteen minutes and five steps give you a working map of any new field.
- Curiosity is no licence to skip review. Someone qualified approves the final cut.
Before your next unfamiliar brief, spend 15 minutes on the map. Then ask one colleague what they check first.
Further reading on my site
Pass the knowledge check to complete this module.