Subtitle: The specific mental models and prompting strategies that separate AI users who get mediocre results from the ones who generate outsized competitive advantage — and how to shift from one to the other starting today.
Here is a question worth sitting with.
The last ten times you opened an AI tool, what did you ask it to do?
If the honest answer is mostly “write this email,” “generate five social media ideas,” or “summarize this article,” you are using one of the most powerful thinking tools in human history as a slightly faster clipboard.
That is not a criticism. That is where most people start. The problem is that most people never move past it.
The marketers and entrepreneurs generating genuinely outsized results from AI are using the exact same tools you have access to. They have not found a secret model or a hidden feature. They are asking fundamentally different questions.
And the difference in what they get back — the quality of insight, the strategic advantage, the compounding leverage — is not incremental. It is categorical.
Let me show you what is actually possible when you change the question.
Key Takeaways
- Most entrepreneurs use AI to produce outputs. High-performing marketers use AI to improve the quality of their thinking before producing anything.
- There are three advanced AI use cases that drive the most strategic value: devil’s advocate stress-testing, research acceleration, and systems extraction.
- The devil’s advocate use case — asking AI to argue against your best plan — consistently produces more defensible strategies than asking AI to validate them.
- Research acceleration means using AI to surface evidence you would never have found through standard search, not to retrieve what you already know.
- Systems extraction turns every high-quality AI-assisted output into a documented, replicable process — transforming one-time results into operational leverage.
The Problem: You Are Using the Advisor as a Typist
When AI tools became widely accessible, the first use case everyone reached for was production. Generate the content. Write the email. Draft the proposal. Complete the task faster.
That is valuable. I am not dismissing it. The time savings from AI-assisted production are real and material.
But here is the economic reality of production-focused AI use: it is increasingly commoditized. Your competitors have the same tools. They can produce the same volume at the same quality floor. The production advantage erodes quickly when everyone has the same capability.
A 2024 survey by Forrester Research found that 60% of businesses using AI reported productivity improvements in content creation, but only 22% reported measurable competitive advantage from their AI implementations. The gap is instructive. Productivity improvement is easy to get. Competitive advantage requires using AI differently.
The businesses in that 22% are not the ones with the best tools or the biggest budgets. They are the ones who have moved beyond production-focused AI use into something more strategically valuable: using AI to think better.
When you use AI to think better, every downstream decision is made from a higher-quality starting point. Every strategy is stress-tested before you commit resources. Every piece of research is more complete. Every process you build is documented and replicable rather than tacit and fragile. These advantages compound. Production speed does not.
The Evidence: Three Use Cases That Drive Disproportionate Value
There are three specific ways high-performing marketers use AI that produce the most outsized returns. They are not complicated. They are just not the first use case most people reach for.
1. Devil’s Advocate Stress-Testing
The devil’s advocate use case works like this: you write out your best plan, strategy, pitch, or assumption, and you ask AI to argue against it as vigorously as possible. Not to validate it. Not to improve it. To find every weakness it can find.
The psychological barrier to this is real. Nobody loves handing their best idea to something designed to find its flaws. But the entrepreneurs who do this consistently produce measurably better plans than those who do not.
Research on adversarial thinking in strategic planning found that teams who engaged in structured devil’s advocate processes identified 47% more plan weaknesses before implementation than teams that did not. Failures caught before implementation cost dramatically less than failures discovered after resources are committed.
The AI version of this is particularly valuable because AI has no ego stake in your plan being correct. It will find the weaknesses without the social friction that comes with asking a colleague to tell you what is wrong with your idea.
2. Research Acceleration
Research acceleration is different from search. When you use AI as a search tool, you are asking it to retrieve what you already know exists. When you use it as a research accelerator, you are asking it to surface what you do not know you do not know.
The practical version: before you write any important piece of content, create a strategy, or make a significant decision, ask AI to map the landscape of evidence. Who are the strongest voices on both sides of this question? What is the most counterintuitive finding in this area? What data have most people in my industry not encountered? What case studies are underrepresented in mainstream content on this topic?
Content produced from this kind of research is differentiated before you write a single word. You are working from a more complete information set than your competitors who are writing based on what they already knew. A Content Marketing Institute study found that data-backed content generates 70% more organic traffic than non-data-backed content, with the differentiation increasing as specificity of data increases.
3. Systems Extraction
This is the use case with the longest compounding runway, and it is the most underused of the three.
Here is the concept. Every time you produce a high-quality output with AI assistance — a great email, a strong proposal, an effective research process — there is an underlying process that generated it. Most people produce the great output and move on. They do not capture the process.
High performers spend an extra ten minutes after any productive AI session asking AI to extract the process: “What was the underlying framework we just followed?” “How would you describe this to someone who needed to repeat it?” “What are the key decision points in what we just did?”
The result is a growing library of documented, replicable processes — built from your actual work, reflecting your actual standards — that makes every subsequent project faster, more consistent, and more delegatable. According to research from the Association for Talent Development, teams with documented processes for their core workflows perform 23% better on both quality and efficiency metrics than teams without.
The Solution: The Four-Level AI Use Framework
The shift from production-focused to thinking-focused AI use is a practice change, not a tool change. Here is a framework for understanding where you are and where to move.
Level 1 — Production: AI writes, drafts, generates, summarizes, reformats. You review and edit. This is where most people operate.
Level 2 — Research: AI surfaces information, data, and perspectives. You evaluate and synthesize. This is where people start to pull ahead.
Level 3 — Strategic Challenge: AI stress-tests your plans, challenges your assumptions, identifies weaknesses, and presents the strongest counter-arguments. You listen, revise, and make better decisions.
Level 4 — Systems Building: AI extracts the frameworks and processes from your best work, converting one-time outputs into replicable systems. You accumulate operational leverage over time.
Most people spend 90% of their AI time at Level 1. The competitive advantage is concentrated at Levels 3 and 4. The move from Level 1 to Level 3 does not require a new tool. It requires a different question.
Practical Steps
1. Run a prompt audit on your last two weeks of AI use.
Go through the prompts you have used in your last 10 to 15 AI sessions. Classify each one as production, research, strategic challenge, or systems building. If the ratio is heavily weighted toward production, you now know exactly where your opportunity is.
2. Use the devil’s advocate prompt before your next major decision.
Before you commit resources to your next marketing campaign, hiring decision, pricing change, or strategic initiative, write out the plan and ask AI: “Argue against this plan as vigorously as possible. Find the weakest assumptions, the most likely failure mode, and the objections I have not considered.” Take the response seriously. Revise accordingly.
3. Do research-first before your next major content piece.
Before you write anything important, ask AI: “What is the most counterintuitive research finding on this topic? What data are most people in my industry not citing? What case study is underrepresented in mainstream content here?” Let that research inform the angle, the evidence, and the argument before you write a single word of the piece.
4. Build a 10-minute systems extraction habit.
After any AI session that produces something you are genuinely happy with, spend 10 minutes asking AI to extract the framework. “What was the underlying process we just followed? What would I call this? How would I describe it to someone who needed to repeat it?” Capture the output in a running document. After 30 days, you will have a library of replicable processes built from your own best work.
5. Write your high-value question library.
For each major area of your business — marketing, sales, operations, content, hiring — build a list of three to five AI prompts designed to surface your blind spots and challenge your assumptions. These are not production prompts. They are thinking prompts. Use them at the start of any significant project or decision in that area.
6. Apply the “what am I missing?” question before any major commitment.
Develop the habit of asking AI “what am I missing?” before you commit to any significant course of action. Not “how do I execute this better?” Not “help me write the plan.” Just “what am I not seeing?” The answers are often the most valuable output of any AI session.
Frequently Asked Questions
Does using AI for strategic thinking take longer than using it for production tasks?
Initially, yes — because you are learning a new prompting practice. Within a few weeks of consistent use, the devil’s advocate and research acceleration sessions typically take 15 to 30 minutes and replace hours of either incomplete analysis or expensive mistakes made on the basis of insufficient thinking. The ROI shifts dramatically in your favor once the practice is established.
What if AI gives me devil’s advocate arguments that are not very strong?
Push back explicitly. Say: “I need you to find stronger objections than that. Assume I am a skeptical investor with deep knowledge of my industry. What is the most compelling argument against this approach?” The quality of devil’s advocate responses scales directly with how rigorously you demand rigor.
How is AI research acceleration different from just searching the web?
Standard search retrieves what you already know to look for. Research acceleration surfaces what you did not know to look for — the counterintuitive finding, the overlooked data set, the expert consensus you were not aware of. The difference is the question: “Tell me what I know about X” versus “Tell me what I don’t know I don’t know about X.”
Can I use AI to extract systems from work that happened before I started using AI?
Yes. Describe the best version of a workflow or output you have produced, from memory or from a sample of past work. Ask AI to infer and document the underlying process based on the description and examples you provide. This works especially well for workflows you have been running for years and have never written down.
How do I know if I am actually using AI for strategic thinking or just producing longer, more complicated prompts?
The test is the output. A production prompt produces a document or piece of content. A strategic thinking prompt produces insight — a perspective, a challenge, a discovery — that changes what you decide to do next. If the AI output is making you reconsider your assumptions or revise your plan, you are operating at the strategic level. If it is producing content you review and accept without changing your direction, you are producing.
The Close
The gap between how most people use AI and how the best AI users use it is not about tools. It is not about access. It is not even about how much time you spend with the technology.
It is about the questions you ask.
Here is the thing about asking harder questions: it requires a willingness to be challenged. To hear that your plan has weaknesses. To discover that you did not have all the evidence. To find out that the framework you were working from was incomplete.
That willingness is not comfortable. But it is the specific practice that separates the entrepreneurs who are building genuinely durable competitive advantages from the ones who are just producing more content a little faster than they used to.
Stop asking AI to type for you. Start asking AI to help you think.
The difference in what comes back is not incremental. It is categorical.
And it is available to you with the exact tools you already have.
About Jonathan Mast
Jonathan Mast is the founder of White Beard Strategies and teaches entrepreneurs how to build AI-powered businesses that actually produce competitive advantage — not just production efficiency. His community at White Beard Strategies is one of the most active AI implementation networks for entrepreneurs in the country, and his approach is consistently the same: frameworks first, tools second, thinking always. He has never met an AI tool that was more valuable than the quality of the question you asked it.
