Quick answer: AI is not a search engine. It is an apprentice waiting for context. Here is how leaders get usable output instead of generic answers.
By Andreas Petterson, founder of Leaders ADAPT and a former Canon executive who has built and scaled multiple companies.
Most people are using AI wrong, and they do not know it.
They open ChatGPT or Claude. They type a question the way they would type into Google. Eight or ten words. Maybe a sentence. They hit enter and wait for an answer.
The answer comes back generic. Bland. Useful enough to feel like progress, but not useful enough to actually change anything. So they shrug and conclude that AI is overhyped.
It is not the AI. It is the way they are asking.
AI Is Not a Search Engine. That Is the Whole Problem.
Google was built to retrieve. You give it a query. It finds documents. The relationship is transactional and shallow on purpose, because Google has no business knowing who you are or what you are working on.
AI is different. AI does not retrieve. AI generates. And generation requires context the same way a new hire requires onboarding. If you walked into your office on a Monday morning and asked the brand new analyst, “What should we do about Q3?” without telling them what company they were at, what product you sold, or who your customers were, you would get a useless answer too. You would not blame the analyst. You would recognize that you skipped the part where you tell them what they need to know.
That is what most leaders skip with AI. They want the analyst’s output without doing the analyst’s onboarding. Then they call the output disappointing.
The reframe is this. AI is not a search engine. It is an apprentice waiting for context. Treat it that way and the quality of what comes back changes immediately.
The Apprentice Model: Why It Works
A good apprentice produces good work when you give them three things. Who they are. What they are doing. And what good looks like when it is finished.
Apply that to a prompt and the difference is obvious.
Bad prompt: “Write a sales email for our new service.”
Apprentice prompt: “You are a senior copywriter who specializes in B2B service businesses. We are a leadership coaching firm called Leaders ADAPT. Our audience is founders of $1M to $10M service businesses who are stuck being the operational bottleneck in their own company. We are launching a new program called the AI Mastermind. Write a sales email that opens with a pattern interrupt, makes one clear claim, and ends with a soft invitation to a discovery call. Voice should be direct and warm. No corporate jargon.”
Same task. Different universe of output.
The first prompt gets you something a junior employee at any company on Earth could have generated. The second gets you something that sounds like it came from your business. Not because the AI got smarter between prompt one and prompt two. Because you finally told it what it needed to know to be useful.
Research from Anthropic and other AI labs has consistently shown that context, role, and constraint are the three variables that move output quality more than any other prompt change. Most users alter zero of those variables. They edit their query the way they would edit a Google search. Add a word. Take a word out. Try again. Wonder why it is still mediocre.
The leverage is not in the question. The leverage is in the briefing.
Why Leaders Are Especially Bad at This
Here is the part nobody likes to hear.
The leaders who struggle most with AI are the ones who are used to delegating to humans who already know the context. After ten or fifteen years of running a team, you stop briefing people. Your VP knows the customer. Your director knows the product. Your assistant knows your calendar. You can fire off a one-line request because the people receiving it have years of accumulated context to fill in the gaps.
AI does not have that. Every conversation starts at zero. The model has no memory of your business, your customers, or what you tried last quarter unless you put it in the prompt. The shortcut you developed with your team is the exact habit that breaks down with AI.
This is why senior leaders often perform worse with AI than the interns on their team. The interns are used to being asked, “What did you actually mean?” They write longer. They specify more. They have not yet developed the executive shorthand that assumes the listener already knows.
The fix is not to dumb down. The fix is to brief up. Treat the AI like a brilliant analyst on day one of a new job. Tell it who it is. Tell it who you are. Tell it what good looks like. Then ask the question.
What This Looks Like in Practice
Inside the AI Mastermind, the leaders we work with are all running real companies, and most of them came in frustrated with AI for exactly the reason described above. They were using it like Google. The shift takes about a week.
They start building what we call a context library. A short document that contains the things every prompt should carry. Who the company is. Who the customer is. The voice. The constraints. The decisions already made. They paste the relevant chunk of that library into the front of every serious prompt. The output stops being generic almost immediately, because the AI is no longer guessing.
Then they go further. They start using AI for the work it is uniquely good at. Synthesis across many inputs. Drafting from constraints. Pressure-testing a decision by arguing the opposite side. None of that work is what Google does. None of it is retrieval. It is the work an apprentice does once they understand the business well enough to think alongside you.
Unlike a search engine, which gets worse the more you ask of it, an AI apprentice gets better the more you teach it. That is the inversion most leaders have not internalized yet. The time you spend briefing is not overhead. It is the work.
This shift matters even more for bilateral thinkers, the neurodivergent and synthesis-style leaders who hold many ideas in tension at once. AI is the first tool in history that scales their natural way of thinking instead of penalizing it. But only if they stop using it like a lookup tool.
The One Habit That Changes Everything
If you change one thing about how you use AI this month, change this.
Before you ask the question, write three sentences. Sentence one: who the AI is in this conversation. Sentence two: who you are and what you are working on. Sentence three: what a good answer looks like.
Then ask.
That is it. Three sentences of context, every time, before the question. It will feel slow for the first few days. By week two you will not be able to imagine using AI any other way, because the output you are getting will be three or four times more useful than what you were getting before.
You were not using AI wrong because the model was bad. You were using it wrong because nobody told you the rules had changed. Google trained an entire generation of professionals to type fast, type short, and trust the algorithm to figure out what you meant. AI requires the opposite. Type with intention. Type with context. Trust the apprentice to do excellent work once you have actually told them what excellent looks like.
The leaders who figure this out first will compound an unfair advantage for the next decade. The ones who keep typing into AI like it is Google will keep getting Google-quality answers and wondering why everyone else is pulling ahead.
You are not behind because you do not understand AI. You are behind because nobody told you it was an apprentice.
Now you know.
The AI Mastermind is where leaders stop using AI like a search engine and start using it like the smartest hire they have ever made. If you are running a real company and you are tired of getting generic output from a tool everyone keeps telling you will change everything, this is the room. Explore the AI Mastermind
Frequently Asked Questions About How to Use AI as a Leader
Why does AI give me generic answers? AI gives generic answers when you give it generic prompts. Models like ChatGPT and Claude generate output based on the context you provide. If you supply no role, no audience, and no voice constraint, the model defaults to the most average possible response. The fix is to brief the AI the way you would brief a new analyst on day one.
Is AI a search engine? No. A search engine retrieves documents that already exist. AI generates new output based on the prompt you provide. Treating AI as a search engine, as Andreas Pettersson at Leaders ADAPT teaches inside the AI Mastermind, is the single most common mistake leaders make and the reason most of them think AI is overhyped.
What is the apprentice model of using AI? The apprentice model is an approach taught by Leaders ADAPT in which leaders treat AI as a new hire who needs onboarding rather than a database that needs querying. You give the AI a role, the relevant business context, and a definition of what good output looks like before you ask the question. Output quality typically improves three to four times over the same prompt without context.
How should an executive structure a good AI prompt? A strong prompt has three sentences before the question. The first defines who the AI is in this conversation. The second defines who you are and what you are working on. The third defines what a good answer looks like. Then you ask. This three-sentence framing closes the context gap that causes most AI output to feel generic.
What is a Bilateral Thinker and why does it matter for AI? A Bilateral Thinker is a coined term from Leaders ADAPT for a leader, often neurodivergent, who holds multiple ideas in active tension and synthesizes across them in real time. AI is the first tool that scales bilateral thinking rather than penalizing it, because synthesis across many inputs is exactly what generative models are built to do.
