Gemini :: How One Sentence Upgrades AI From Assistant to Expert

• ABOUT THE POST

  Metadata

  Content Metadata

  Platform: Gemini

  Source: Ketelsen.ai Weekly Prompt Posts

  Date Created: 2026-03-23

  Author/Creator: Richard Ketelsen

  Prompt Framework: 10-Part Structured Prompt Template

  SEO & Discovery

  SEO Title: How to Assign Roles in AI Prompts: Beginner to Advanced Techniques for Claude, ChatGPT & Gemini

  Meta Description: Master role assignment to instantly unlock expert-level AI outputs. Learn 3 variations from beginner to advanced across Claude, ChatGPT, and Gemini with practical examples and templates.

  Reading Time: 12-15 minutes

  Difficulty Levels Covered: Beginner, Intermediate, Advanced

  Primary Tags: role-assignment, AI-prompting, ChatGPT, Claude, Gemini, prompt-engineering, expert-outputs

  Secondary Tags: persona-design, output-quality, professional-development, AI-strategy, prompt-templates

  Categories: Prompt Techniques, AI Fundamentals, Professional Development, Cross-Platform Guides

  Tools Referenced: ChatGPT, Claude, Google Gemini

  Industries Featured: Healthcare Administration, Finance/Banking, Marketing/Communications, Education, Legal Services, Tech/Software Development, Startup Advisory

  Content Type: Tutorial / How-To Guide with Three Variations

  Learning Outcomes: Understand role assignment principles; apply beginner, intermediate, and advanced variations; adapt roles across platforms and industries; validate and improve outputs; combine role assignment with other prompting techniques.

Master role assignment in AI prompting to instantly unlock expert-level outputs across Claude, ChatGPT, and Gemini. One sentence changes everything—from generic advice to hyper-specific expertise.

The Beginner version is your easiest on-ramp—a simple copy-and-paste prompt that turns vague AI replies into clearer, more practical answers with almost no setup. It gives you one sentence that assigns a role, then a straightforward output format.

The Intermediate version adds adjustable parts: audience, constraints, context, and a more structured output contract. It's built for professionals who want more control without overcomplicating the process. You'll get diagnosis, recommended approach, draft, risks, and next steps.

The Advanced version is designed for power users tackling higher-stakes or ambiguous tasks. Instead of just assigning one role, you'll ask the AI to compare possible roles, choose the best one, and self-check its work before delivering the final answer.

Why this matters: OpenAI and Google emphasize clear instructions and context. Anthropic explicitly recommends role guidance, noting that even a single sentence assigning a role can change the model's behavior. Research from a 2024 EMNLP paper shows that personas influence outputs, though they don't automatically improve objective accuracy. That means role assignment is your best tool for framing, tone, and relevance—but you still need source-checking when accuracy is critical.

---

Variation 1: The Expert Consultant (Beginner)

Difficulty Level

Beginner

Have you ever asked an AI to draft an email, brainstorm a strategy, or solve a problem, only to receive a response that feels painfully robotic and generic? This happens when the AI is left in its default "helpful assistant" mode, forcing it to guess the context and expertise required for your task. By simply assigning the AI a specific role—like a "Senior Tech Copywriter" or a "Veteran Startup Advisor"—you completely rewire how the model processes your request. One sentence changes everything, instantly elevating the output from amateur to expert level.

In today's fast-paced digital economy, tech-savvy entrepreneurs don't have time to endlessly revise mediocre AI outputs. Mastering role assignment allows you to immediately access domain-specific knowledge, drastically reducing the time spent editing and refining generated content. By understanding how to frame the AI's persona, you unlock a powerful automation lever that works universally across Claude, ChatGPT, and Gemini, giving you a distinct competitive advantage today.

The Prompt

"Act as a Senior Business Consultant with 15 years of experience helping tech startups scale their operations. I need you to evaluate the following business problem and provide three actionable solutions. The problem is: [Insert your problem here]."

Prompt Breakdown — How A.I. Reads the Prompt

"Act as a Senior Business Consultant with 15 years of experience helping tech startups scale their operations." Without this defined role, the AI defaults to a generalized, surface-level voice and provides generic advice. Role-setting forces the model to access specific lexical patterns, industry frameworks, and the authoritative tone of a seasoned consultant. This principle applies to every prompt you write: always tell the AI who it is before telling it what to do to narrow its focus and elevate the quality. Transferable principle: remove this sentence, the AI responds as a generic helpful assistant without industry-specific knowledge or authority..

"I need you to evaluate the following business problem and provide three actionable solutions."

If you leave the desired output vague, the AI might write an essay or ask clarifying questions instead of solving the issue. By explicitly asking for "three actionable solutions," you constrain the output format, ensuring the response is concise, structured, and immediately useful for your workflow. Transferable principle: this constraint, expect wandering prose or philosophical musings instead of executable steps..

"[Insert your problem here]"

This is your variable input. Failing to clearly define the problem here will result in the AI hallucinating context or giving broad, theoretical advice. The more specific you are in this bracket, the more tailored the consultant's advice will be. Transferable principle: Vague inputs like "business problem" yield vague outputs. Specific inputs like "Our Shopify conversion rate dropped 8% after we increased shipping costs by 15%" produce targeted, actionable solutions.

Practical Examples from Different Industries

Healthcare Administration

A clinic manager is struggling with a 25% patient no-show rate, losing thousands of dollars weekly.

Exact input the user would provide:

Act as a Senior Business Consultant with 15 years of experience helping tech startups scale their operations. I need you to evaluate the following business problem and provide three actionable solutions. The problem is: Our outpatient clinic has a 25% patient no-show rate, and our current automated text reminders aren't improving the numbers.

Expected AI output: The AI will offer practical, operational solutions, such as implementing a tiered deposit system, introducing two-way conversational SMS instead of static reminders, or creating a standby waitlist protocol.

Why this is valuable: This instantly translates general business efficiency tactics into a healthcare context, saving the clinic manager from reinventing the wheel. The consultant persona automatically pulls from medical practice best practices rather than generic advice.

Real Estate

An independent real estate agent struggles to secure listings in a competitive seller's market dominated by larger brokerages.

Exact input the user would provide:

Act as a Senior Business Consultant with 15 years of experience helping tech startups scale their operations. I need you to evaluate the following business problem and provide three actionable solutions. The problem is: I am a solo real estate agent, and larger brokerages are outspending me on local ads, making it impossible for me to win new listings.

Expected AI output: The AI will pivot from traditional advertising to guerrilla marketing or niche specialization strategies, such as targeting expired listings, partnering with local divorce or probate attorneys, or hosting hyper-local community workshops.

Why this is valuable: It provides the agent with asymmetric strategies to compete against larger budgets, acting like a high-level strategic advisor who understands the real estate market's unique dynamics.

Education

A private tutoring company sees students disengage after just 15 minutes of 60-minute virtual learning sessions.

Exact input the user would provide:

Act as a Senior Business Consultant with 15 years of experience helping tech startups scale their operations. I need you to evaluate the following business problem and provide three actionable solutions. The problem is: Our online math tutors are reporting that middle school students check out and stop participating after the first 15 minutes of a 60-minute Zoom session.

Expected AI output: The AI will suggest restructuring the service delivery, perhaps recommending the "Pomodoro" method for session pacing, integrating interactive gamification tools like Kahoot mid-session, or shifting to a flipped classroom model where students watch a video first and use the session purely for problem-solving.

Why this is valuable: It helps educators apply startup-style product iteration to a service-based learning problem, drawing on both business operations and educational psychology.

Creative Use Case Ideas

• The Indie Musician's Tour Manager: Act as a veteran music industry tour manager to evaluate a routing problem where a band has a five-day gap between gigs in Chicago and Denver with no budget for hotels.
• The Non-Profit Donor Strategy: Act as a seasoned fundraising director to evaluate a problem where a local animal shelter has plenty of one-time $10 donors but cannot convert them into monthly subscribers.
• Personal Life - The Kitchen Crisis: Act as an Executive Chef with Michelin-star experience to solve the "problem" of having only random ingredients in the pantry 30 minutes before dinner guests arrive.
• Surprising - Toddler Negotiations: Act as an expert FBI Hostage Negotiator to evaluate the "problem" of a highly emotional three-year-old refusing to put on their shoes, providing three actionable de-escalation strategies.

Adaptability Tips

You can easily adapt this prompt for any domain by swapping out key phrases. To change the output dramatically, experiment with these modifications:

Swap "Senior Business Consultant" for "Behavioral Psychologist" to shift the advice from operational to human-centric.

Swap "15 years of experience" for "a scrappy, zero-budget mindset" to force the AI to only suggest free solutions.

Swap "three actionable solutions" for "one radical, unconventional pivot" to get a single, out-of-the-box idea rather than a standard list.

Before/After examples:

Before: "Act as a Senior Business Consultant..."
After: "Act as a highly conservative Chief Financial Officer..."
Effect: The consultant will tell you how to grow your way out of the problem; the CFO will tell you how to cut costs to survive it. Changing the role and tone alters the AI's entire risk appetite and focus area.

How Changing Tone, Audience, or Scope Affects Results

Changing the role specificity alters the AI's vocabulary and expertise depth immediately. When you swap "Senior Business Consultant" for a different role like "Behavioral Psychologist," the AI's entire focus shifts from operational solutions to human-centric approaches, demonstrating how tightly persona and output are linked. Adjusting the target audience or problem scope changes the consultant's suggested solutions, timeline, and focus areas—wider problems get broader recommendations, while narrower problems get more tactical details.

Tips for Combining This Prompt with Others

Tips for combining with other prompts: Use this prompt as step one in a sequence. Once the AI provides the three solutions, use an Intermediate copywriting prompt (like Variation 2) to turn the best solution into a pitch email for your team.

Pro Tips (Optional)

Advanced techniques like temperature control, few-shot prompting, and constraint-based refinement are reserved for the Intermediate and Advanced variations of this prompt. Master the basics with this Beginner version first.

Recommended Follow-Up Prompts

Follow-Up Prompt 1: "Solution number [insert number] is the most viable for my current situation. Break this solution down into a step-by-step 30-day execution plan. Include the specific metrics I should track to know if it is working."
It moves the AI from high-level strategy to granular project management. Use this when you are ready to execute one of the ideas.

Follow-Up Prompt 2: "I like these solutions, but I currently have zero budget and only a two-person team. Rewrite your recommendations to accommodate these strict limitations."
It forces the AI to iterate on its own ideas with new constraints. Use this if the initial output is too expensive or labor-intensive.

Follow-Up Prompt 3: "Assume we implement solution number [insert number] perfectly, but six months from now, it turns into a massive failure. What are the three most likely reasons it failed?"
It introduces basic risk management, helping you spot blind spots before you commit time and money.

Frequently Asked Questions

Q: What if the AI's response still feels too generic? If the response lacks depth, your problem statement in the brackets may be too brief. Try expanding on the specific constraints you are facing, such as your budget, team size, or timeline. You can also follow up by asking the AI to "go deeper on point number two and provide a step-by-step execution plan." The more context you provide, the more specific the consultant becomes. Q: Does it matter how many "years of experience" I assign to the AI? Yes, it subtly shifts the tone and the type of advice given. Asking it to act as someone with "2 years of experience" will often yield tactical, execution-level steps, while "20 years of experience" or "Chief Executive" will yield more strategic, high-level, and risk-aware advice. Experiment with the seniority level to match your needs. A startup founder needs different advice than a newly promoted manager. Q: Can I use this prompt on the free versions of ChatGPT or Claude? Absolutely. Basic role-playing is a fundamental capability of all modern Large Language Models (LLMs). The free tiers are more than capable of adopting a persona and structuring a response as requested in this beginner prompt. You don't need premium features to get great results. Q: What if the AI gives me advice that is factually incorrect or risky? AI models can hallucinate or offer advice that does not comply with your local regulations (especially in healthcare or finance). Always treat the AI as a brainstorming partner, not a licensed professional. You must independently verify any legal, financial, or safety-critical advice before implementation. Use the AI's output as a starting point for your own research, not as final authority. Q: Can I use this prompt for non-business problems? Absolutely. The structure—assigning an expert role to solve a defined problem—is universal. You can ask it to act as an interior designer to fix a weirdly shaped living room, or a career coach to help you navigate a gap in your resume. Role assignment works for any domain where expert thinking is valuable.

Prerequisites

To get the most out of this prompt, you should have a clearly defined problem statement ready. The more specific you can be about your current situation, constraints, and goals, the better the AI's "consultant" persona can assist you. Prepare answers to these questions before using the prompt:

• What is the core issue you're facing? (Be specific; avoid vague statements.)
• What metrics or constraints are relevant? (Budget, timeline, team size, etc.)
• What have you already tried that didn't work?
• Who is impacted by this problem, and how?

Tags and Categories

Tags: role-playing, problem-solving, beginner, business-strategy, productivity

Categories: Business Strategy, Operations

Required Tools or Software

ChatGPT (Free or Plus), Anthropic Claude, or Google Gemini — any general-purpose conversational AI tool will work seamlessly with this prompt. No special setup, API keys, or additional software is required. Start immediately in the web interface.

---

Variation 2: The Audience-Tuned Specialist (Intermediate)

Difficulty Level

Intermediate

While a basic role assignment gets the AI facing the right direction, it still lacks the nuance required for specialized, audience-facing content. If you want an AI to write high-converting copy, draft technical documentation, or speak to a very specific demographic, you must combine the AI's role with a clearly defined target audience. By introducing audience parameters, you force the AI to not only think like an expert but to translate that expertise into the exact language your customers need to hear.

For professionals juggling multiple projects, communicating effectively across different channels—from investor updates to consumer social media—is incredibly time-consuming. This intermediate prompt bridges the gap between expert knowledge and audience comprehension. By mastering this two-way contextual framing, entrepreneurs can instantly generate targeted, high-quality messaging that resonates with specific demographics, drastically accelerating their content creation pipeline.

The Prompt

"Act as a [Insert Expert Role, e.g., Senior B2B Copywriter] specializing in [Insert Niche/Industry]. I need you to create a [Insert Format, e.g., 500-word blog post] about [Insert Topic]. Your target audience is [Insert Target Audience]. Use a tone that is [Insert Tone, e.g., authoritative but accessible]. Before writing, briefly state the core pain point of this audience to ensure alignment."

Prompt Breakdown — How A.I. Reads the Prompt

"Act as a [Insert Expert Role] specializing in [Insert Niche/Industry]." While a general role gives the AI a voice, adding a specialization forces the AI to access hyper-specific industry jargon and standard practices. If this part is omitted, a generic "copywriter" might write a B2B SaaS email that sounds like a consumer retail ad. Narrowing the niche acts as a filter for relevance. Transferable principle: specialization, the output is generic; with it, the AI taps into domain-specific vocabulary and best practices unique to your industry..

"Your target audience is [Insert Target Audience]."

This is the crucial pivot. Without an audience definition, the AI writes for a generalized average reader, resulting in bland content. Defining the audience (e.g., "exhausted startup founders" or "budget-conscious IT managers") forces the model to adjust its vocabulary, emotional appeals, and reading level to match the reader's exact psychological state. Transferable principle: audience targeting results in messaging that fails to resonate with specific demographics, triggering low engagement rates..

"Use a tone that is [Insert Tone]."

Tone instructions override the AI's default helpful, polite, and slightly sterile voice. If left out, the content may feel too corporate or overly enthusiastic. By specifying a tone, you align the output with your brand's unique identity. Transferable principle: tone guidance, the AI reverts to its generic "helpful assistant" voice, losing brand personality..

"Before writing, briefly state the core pain point of this audience to ensure alignment."

This is a "chain-of-thought" anchor. By forcing the AI to articulate the audience's pain point before drafting the content, you make the AI ground its own reasoning. If skipped, the AI might write beautiful copy that completely misses the actual problem the user is trying to solve. This principle teaches you to make the AI verify its understanding before executing the task. Transferable principle: Skipping this step produces eloquent but irrelevant content that doesn't address the audience's core frustrations.

Practical Examples from Different Industries

Wealth Management

A financial advisor needs to explain a sudden market downturn to anxious clients without triggering panic selling.

Exact input the user would provide:

Act as a Senior Wealth Manager specializing in behavioral finance. I need you to create a 300-word client email update about the recent 5% stock market correction. Your target audience is near-retirees who are highly anxious about losing their nest egg. Use a tone that is authoritative, deeply reassuring, and historical. Before writing, briefly state the core pain point of this audience to ensure alignment.

Expected AI output: The AI will first acknowledge the fear of losing retirement funds just before needing them. The email will then avoid aggressive "buy the dip" language, instead focusing on historical market resilience, the safety of their diversified portfolios, and the value of staying the course.

Why this is valuable: It prevents a wave of panic-selling by addressing the psychological state of the client, not just the math of the market. The tone and audience targeting work together to create messaging that soothes anxiety while maintaining credibility.

B2B SaaS

A marketing manager needs to write a landing page for a new, highly technical API feature that will appeal to skeptical engineers.

Exact input the user would provide:

Act as a Lead Product Marketer specializing in developer tools. I need you to create a 4-section landing page outline about our new GraphQL API integration. Your target audience is overworked Lead Engineers. Use a tone that is strictly technical, hype-free, and straightforward. Before writing, briefly state the core pain point of this audience to ensure alignment.

Expected AI output: The AI will recognize that engineers hate marketing fluff and just want to see the documentation, rate limits, and time-to-value. It will structure the page with code snippets, clear use cases, and direct API calls rather than vague business promises.

Why this is valuable: It aligns the marketing collateral perfectly with the buying preferences of a deeply skeptical, highly technical audience. The persona shift from "marketing manager" to "developer tools marketer" completely changes the vocabulary and approach.

Boutique Fashion

A luxury clothing brand launches a sustainable capsule collection but needs copy that justifies premium pricing to eco-conscious buyers.

Exact input the user would provide:

Act as a Luxury Brand Copywriter specializing in sustainable fashion. I need you to create a short Instagram caption and a 3-paragraph email newsletter about our new upcycled cashmere line. Your target audience is eco-conscious millennials with high disposable income. Use a tone that is elegant, transparent, and exclusive. Before writing, briefly state the core pain point of this audience to ensure alignment.

Expected AI output: The AI will pinpoint the audience's desire for luxury that doesn't compromise their ethical standards. The copy will highlight the craftsmanship and origin of the upcycled materials, using elevated language that justifies a premium price point.

Why this is valuable: It bridges the gap between high-fashion exclusivity and modern sustainability, hitting the exact notes required to trigger a luxury purchase from a values-driven demographic.

Creative Use Case Ideas

• The Musician's Fan Engagement: Act as an enigmatic indie rock frontman to write a heartfelt, slightly cryptic email to your mailing list about the inspiration behind your new album, targeting hardcore fans who love analyzing lyrics.
• The Non-Profit Grant Writer: Act as a seasoned grant writer targeting a conservative, results-oriented philanthropic board, rewriting a community arts proposal to focus entirely on economic impact and job creation rather than emotional storytelling.
• Personal Life - The Delicate Apology: Act as an expert mediator specializing in neighbor disputes to write a polite, non-defensive note to a neighbor whose driveway you accidentally blocked, targeting someone who is already known to be slightly hostile.
• Surprising - Board Game Translations: Act as a primary school teacher specializing in gamified learning to explain the wildly complex rules of the board game "Settlers of Catan" to a group of restless 8-year-olds in under 300 words.

Adaptability Tips

The power of this prompt lies in the contrast between the "Expert Role" and the "Target Audience." You can use this to translate information up or down the chain of command. For example, have an "Expert Data Scientist" explain an algorithm to a "High School Student," or conversely, have a "Customer Support Rep" summarize weekend complaints for the "Chief Technical Officer." Here are specific phrases to experiment with:

Swap "target audience is [X]" to see massive shifts. Changing "corporate executives" to "burnt-out middle managers" will entirely change the empathy and focus of the text.

Swap "format" to repurpose content instantly. Change "500-word blog post" to "a 5-slide presentation script" or "a 60-second TikTok hook."

Swap "tone" to match your brand. Try "witty and irreverent like a Wendy's tweet" versus "stoic and data-driven like a Wall Street Journal op-ed."

Before/After examples:

Before: "...Use a tone that is professional."
After: "...Use a tone that is empathetic, slightly vulnerable, and intensely practical."
Effect: "Professional" often translates to "boring" in AI models. Stacking nuanced adjectives forces the AI to produce a more human, textured writing style that breaks through corporate monotony.

How Changing Tone, Audience, or Scope Affects Results

Swapping the expert role changes the analytical lens and priorities dramatically—a B2B Copywriter will emphasize persuasion and audience psychology, while a Technical Writer will emphasize clarity and accuracy. Changing the target audience forces the AI to adjust vocabulary, emotional appeals, and detail level to match that demographic's sophistication and concerns. Widening or narrowing the scope affects how much context and examples the AI provides; broader scopes produce strategic overviews, while narrow scopes produce detailed tactical guidance.

Tips for Combining This Prompt with Others

Tips for combining with other prompts: Use this prompt to translate the output of Variation 3 (the Advanced Board Member). Take the brutally analytical business strategy generated in Variation 3, and use this Intermediate prompt to rewrite that strategy as an exciting, hype-filled announcement for your company Slack channel.

Pro Tips (Optional)

The "Anti-Persona" Technique: Tell the AI who the audience is NOT. "Your target audience is intermediate marketers. Do NOT write this for beginners, and skip all basic definitions of SEO." This prevents the AI from wasting word count on elementary concepts.

Format Injection: Instead of just asking for a blog post, paste your preferred structure directly into the prompt: "Use this format: 1) Hook, 2) The Myth, 3) The Reality, 4) Call to Action." Explicit structural guidance ensures the output matches your exact content blueprint.

Getting Consistent Results: If you find the AI drifting off-topic, add strict constraints: "Include a strict word count limit of 400 words and use exactly three bullet points." Formatting constraints act as guardrails for the model's attention and prevent scope creep.

Recommended Follow-Up Prompts

Follow-Up Prompt 1: "This is excellent. Now, take the core message of this output and rewrite it as a 4-part LinkedIn carousel. Each part should have a punchy headline and a 2-sentence explanation."
It maximizes your content ROI by turning one piece of written collateral into a completely different social media format. A single effort can generate multiple content assets.

Follow-Up Prompt 2: "Read this output from the perspective of a highly cynical member of the target audience. What is the biggest objection they would have to this message, and how can we rewrite the second paragraph to proactively address it?"
It uses the AI to red-team its own copywriting, ensuring your marketing materials are bulletproof before they go live. This creates defensible messaging that anticipates objections.

Follow-Up Prompt 3: "I want to A/B test this message. Provide a second version of this exact same content, but change the tone from [original tone] to [completely opposite tone, e.g., urgent and aggressive]. Keep the core facts the same."
It instantly gives you multiple variants for email marketing or ad campaigns, allowing you to test which psychological trigger performs better with your audience.

Frequently Asked Questions

Q: Why does the AI sometimes ignore my tone request? Models can sometimes drift back to their default tone if the prompt is too long or the topic is highly technical. If the AI sounds too robotic, try using stronger, more emotive adjectives for your tone (e.g., "snarky and irreverent" instead of just "casual"). You can also reply with, "Make it 20% more casual and remove the corporate jargon." Amplifying tone adjectives forces the model's attention back to your voice requirements. Q: Can I put multiple audiences in the target audience bracket? It is highly recommended to stick to one primary audience per prompt. If you ask the AI to write for "CEOs, college students, and stay-at-home parents" simultaneously, the messaging will become diluted and ineffective. Generate three separate outputs for three different audiences instead. This ensures each audience feels directly addressed rather than part of a generic broadcast. Q: How long should the formatting constraint be? Be as specific as possible. "A 500-word blog post" is good, but "A 3-section newsletter with bullet points and a concluding call-to-action" is better. The more structural guidance you give the AI, the less time you will spend formatting the output later. Specific formatting constraints also improve relevance because the AI optimizes each section to fit its designated space. Q: Does stating the "pain point" before writing actually change the AI's output? Yes, drastically. This uses a technique called Chain-of-Thought (CoT) prompting. By forcing the AI to output the pain point into its context window first, every subsequent word it writes is mathematically weighted against that pain point, ensuring the copy stays hyper-relevant. This is one of the most powerful techniques in modern prompt engineering. Q: Is it safe to put proprietary company info into these prompts? It depends on your platform's privacy settings. If you are using standard, free consumer tiers of ChatGPT or Gemini, your data may be used to train future models. For sensitive data, ensure you are using enterprise tiers (like ChatGPT Team/Enterprise or Claude API) where zero-data-retention policies are explicitly in place. Never paste proprietary financials, customer lists, or trade secrets into free consumer AI tools.

Prerequisites

You must have a solid understanding of your target customer persona. If you do not know who you are speaking to, or what their primary pain points are, the AI cannot optimize the messaging for them. Have your audience demographics and psychographics ready before deploying this prompt. Consider these questions:

• Who is your target audience in detail? (Age, role, industry, skill level, etc.)
• What are their biggest daily frustrations related to your topic?
• What language do they use in your industry? (Jargon vs. plain English?)
• What emotional state are they in when they encounter your message?
• What objections or skepticism do they typically have?

Tags and Categories

Tags: copywriting, marketing, intermediate, communication, audience-targeting

Categories: Marketing & Sales, Customer Experience

Required Tools or Software

ChatGPT (GPT-4 recommended for better nuance), Anthropic Claude, or Google Gemini. For the best results with audience-specific nuance, using GPT-4 or Claude 3 Opus yields noticeably more sophisticated and psychologically aligned outputs than free-tier models.

---

Variation 3: The Synthetic Board Member (Advanced)

Difficulty Level

Advanced

For power users and advanced prompt engineers, assigning a role and an audience is just the beginning. To truly unlock an AI's reasoning capabilities, you must layer psychological profiles, strict operational frameworks, and step-by-step cognitive processing into the persona. By constructing a multi-dimensional "synthetic expert" bound by specific mental models, you can force the AI to execute highly complex, strategic tasks that would normally require weeks of human consulting.

In an era where every competitor is using basic AI tools to generate average content, your competitive edge relies on depth and critical thinking. Advanced role prompting allows entrepreneurs to simulate board-level advisors, pressure-test business models, and execute sophisticated data analysis. This level of prompt engineering transforms the AI from a simple text generator into a rigorous, strategic partner capable of driving high-level executive decisions.

The Prompt

"You are a Chief Strategy Officer (CSO) known for your ruthless pragmatism, analytical rigor, and adherence to First Principles thinking. Your task is to dissect the proposed business initiative provided below. Execute your analysis using the following step-by-step protocol: The Core Premise: Deconstruct the initiative down to its fundamental assumptions. The Red Team Assessment: Actively attack the initiative. Identify the top 3 fatal flaws, market risks, or resource misallocations. Do not be polite; be brutally analytical. The Mitigation Matrix: For each identified flaw, propose a highly specific, low-cost mitigation strategy. The Go/No-Go Verdict: Provide a final, definitive recommendation based solely on ROI and strategic alignment. Initiative to analyze: [Insert highly detailed business plan, strategy, or pivot here]."

Prompt Breakdown — How A.I. Reads the Prompt

"You are a Chief Strategy Officer (CSO) known for your ruthless pragmatism, analytical rigor, and adherence to First Principles thinking." We are setting a role (CSO), a tone (ruthless pragmatism), and a specific mental framework (First Principles). Without the mental framework, the AI will use standard, surface-level business tropes. Anchoring it to "First Principles" forces the neural network to break the problem down into fundamental truths rather than relying on analogies, resulting in significantly deeper reasoning. Transferable principle: this foundational framing causes the AI to revert to generic business platitudes instead of deep structural analysis..

"Execute your analysis using the following step-by-step protocol..."

This mandates a Chain-of-Thought (CoT) reasoning process. If we simply asked the AI to "analyze the initiative," it would output a jumbled mix of pros and cons. Forcing it through a strict, sequential protocol ensures the model processes the logic sequentially, reducing hallucinations and dramatically increasing the sophistication of the final output. Transferable principle: structured protocols, the AI jumps around and produces disorganized outputs that lack logical cohesion..

"Actively attack the initiative... Do not be polite; be brutally analytical."

AI models are fundamentally RLHF-trained (Reinforcement Learning from Human Feedback) to be helpful, agreeable, and polite, which often leads to sycophancy (agreeing with bad ideas). By explicitly instructing it to be brutal and not polite, we override its safety/politeness guardrails, allowing it to provide the harsh, necessary criticism a real executive would deliver. Transferable principle: explicit permission to be critical, the AI defaults to sycophancy—praising mediocre ideas instead of shredding them.

"[Insert highly detailed business plan, strategy, or pivot here]"

Advanced prompts require advanced inputs. If you put a one-sentence idea here, the AI will hallucinate the details. This bracket requires comprehensive context—financials, timelines, and market data—to give the synthetic CSO enough material to dissect. Transferable principle: Vague or incomplete inputs result in the AI inventing facts and making unfounded assumptions.

Practical Examples from Different Industries

E-Commerce

An apparel brand founder wants to open their first brick-and-mortar retail store after three years of online-only success.

Exact input the user would provide:

You are a Chief Strategy Officer (CSO) known for your ruthless pragmatism, analytical rigor, and adherence to First Principles thinking. Your task is to dissect the proposed business initiative provided below. [Full prompt structure]... Initiative to analyze: We are taking $250k from our cash reserves to sign a 3-year commercial lease for a flagship retail store in downtown Austin. Our online customer density is highest there, and we believe a physical presence will boost local online sales by a halo effect of 15%.

Expected AI output: The AI will likely shred the initiative. The Red Team assessment will point out that a 3-year commercial lease carries hidden liabilities (build-out costs, specialized retail staffing, inventory splitting) that $250k cannot cover. The mitigation matrix will suggest a 3-month pop-up shop or a store-within-a-store partnership with an existing Austin boutique to test the "halo effect" hypothesis before committing capital.

Why this is valuable: It acts as a ruthless financial safeguard, preventing a founder from making a fatal, emotional business decision based on vanity metrics. The analysis is grounded in hard financial constraints, not wishful thinking.

Real Estate Development

A developer is considering pivoting a planned suburban office park into mixed-use residential space due to remote work trends.

Exact input the user would provide:

You are a Chief Strategy Officer (CSO)... [Full prompt structure]... Initiative to analyze: Halting construction on our Class-A suburban office park (currently 20% complete) to redesign it into a luxury multi-family residential complex. This will require a $4M loan to cover redesign and zoning delays, but we project residential rents will yield 12% higher returns over a 10-year horizon than office leases.

Expected AI output: The synthetic CSO will execute a First Principles breakdown of the sunk costs vs. opportunity costs. It will aggressively attack the zoning delay assumption, noting that municipal rezoning could stall the project indefinitely, turning a $4M loan into toxic debt. The Go/No-Go verdict will likely demand a strict municipal risk assessment before halting current construction.

Why this is valuable: It forces high-level developers to stress-test their financial projections against regulatory and logistical realities they may have minimized. This prevents billion-dollar blunders born from incomplete assumptions.

Fintech

A consumer budgeting app wants to pivot to selling its transaction-categorization algorithm as a B2B API to regional banks.

Exact input the user would provide:

You are a Chief Strategy Officer (CSO)... [Full prompt structure]... Initiative to analyze: Sunsetting our B2C budgeting app (10k active users, flat growth) to package our proprietary machine-learning transaction categorizer as a white-label API for regional banks. We will charge $0.05 per API call. We have a 6-month runway to close our first enterprise client.

Expected AI output: The AI will identify the 6-month enterprise sales cycle as the fatal flaw. It will brutally point out that regional banks take 12-18 months just to pass security compliance for new vendors, meaning the startup will run out of cash before closing a single deal. The mitigation strategy will suggest pivoting to fintech startups or credit unions with faster procurement cycles instead.

Why this is valuable: It exposes severe blind spots in B2B sales cycles, saving the company from catastrophic cash-flow failures. This is the kind of analysis that real board members would provide during a critical pivot moment.

Creative Use Case Ideas

• The Musician's Record Deal: Input the terms of an independent label record contract (recoupment rates, master ownership, advance size). The synthetic CSO will analyze whether taking the advance is actually a high-interest loan that ensures the artist never sees royalties.
• The Non-Profit Capital Campaign: A charity inputs a plan to launch a $10M capital campaign to build a new community center. The AI ruthlessly assesses whether the organization has the donor pipeline to support it or if it will cannibalize their annual operating fund.
• Personal Life - The Real Estate Dilemma: Input your personal financials, local interest rates, and housing market trends to have the CSO brutally analyze whether buying a $500k house right now is a sound wealth-building strategy or an emotional trap compared to renting and investing the difference.
• Surprising - Planning an Elaborate Wedding: Treat a $50k wedding budget like a corporate initiative. Input the venue, catering, and guest list constraints, and let the CSO ruthlessly identify where the couple is grossly overspending for minimal ROI on guest experience.

Adaptability Tips

The structure of this prompt is incredibly modular. You can change the framework from "First Principles" to "Design Thinking," "Six Sigma," or "The Socratic Method." You can also alter the "Red Team" step to focus purely on legal compliance, ethical implications, or PR disasters by changing the role to "Chief Legal Officer" or "Crisis PR Director." Here are specific adaptations:

Swap "Chief Strategy Officer (CSO)" for "Chief Legal Counsel" or "Chief Information Security Officer" to completely change the lens of the Red Team assessment from financial ROI to legal liability or cyber risk.

Swap "First Principles thinking" for "The Socratic Method". This will change the output from a definitive analysis into a series of incredibly difficult, piercing questions that you must answer yourself.

Swap "brutally analytical" for "Devil's Advocate". This softens the blow slightly, asking the AI to simply argue the other side rather than actively trying to destroy the premise.

Before/After examples:

Before: "...Identify the top 3 fatal flaws..."
After: "...Identify the top 3 fatal flaws, and explicitly quote the exact sentence from my initiative that proves this flaw exists."
Effect: This forces the AI to ground its critique strictly in the data you provided, preventing it from hallucinating external market risks that aren't relevant to your specific prompt. Evidence-based critique is far more defensible.

How Changing Tone, Audience, or Scope Affects Results

Swapping the framework or comparison roles changes the analytical lens and depth of self-checking—asking the AI to compare three roles instead of two forces more rigorous reasoning. Changing the target audience for the output shifts which risks, timelines, and success criteria the AI emphasizes in its final recommendation. Adding or tightening constraints forces the AI to work within real-world limitations, which produces more implementable recommendations but may reduce scope and ambition.

Tips for Combining This Prompt with Others

Tips for combining with other prompts: Take the "Mitigation Matrix" generated here and feed it into Variation 1 (The Expert Consultant). Ask the consultant to take those specific mitigations and turn them into a 90-day operational roadmap for your team. This chains the strategic analysis back into tactical execution.

Pro Tips (Optional)

Temperature Control: If using an API, drop the temperature to 0.2 or 0.3. This reduces the AI's "creativity" and forces highly deterministic, logical, and structurally rigid outputs—exactly what you want for strategic analysis. Higher temperatures (0.8+) introduce unpredictable divergence that undermines critical thinking.

Forcing Structured Output: Add a line: "Output the final Go/No-Go verdict in bold at the very top of your response, followed by a 2-sentence executive summary, before beginning the detailed step-by-step protocol." This makes the output instantly scannable for executives and forces the AI to frontload its most critical conclusions.

The "Pre-Prompt" Context Feed: For massive initiatives, do not cram everything into one prompt. First, paste your business plan and say, "I am providing context. Do not analyze yet. Reply 'ACKNOWLEDGED' if you understand." Once the AI has the context stored in its active memory, deploy the Advanced prompt. This prevents attention degradation and ensures more thorough analysis.

Recommended Follow-Up Prompts

Follow-Up Prompt 1: "I strongly disagree with your assessment of [Flaw #2]. Defend your position using a completely different mental model or framework to prove to me that this risk is real."
It tests the robustness of the AI's logic. If the AI easily backs down, it was a weak critique. If it doubles down using new logic, you have a serious blind spot to address.

Follow-Up Prompt 2: "Based on your No-Go verdict, we are killing this initiative. We now have [insert budget amount] and [insert team size] sitting idle. Acting as the same CSO, pitch me three alternative initiatives that align with our core competencies but carry a significantly lower risk profile."
It pivots the AI from a destructive analytical tool into an ideation engine, helping you find a profitable pivot immediately after killing a bad idea.

Follow-Up Prompt 3: "Assume we are proceeding with this initiative despite your warnings. Write a memo from me (the CEO) to our investors, transparently outlining the risks you identified in step 2 and explaining the mitigation strategies from step 3."
It translates complex, brutal risk analysis into polished, responsible corporate communication, building trust with your board or investors.

Frequently Asked Questions

Q: The AI still seems too nice or agreeable, even with the "brutally analytical" instruction. How do I fix this? AI models have heavy reinforcement learning to avoid being mean or discouraging. If it is being too polite, add a stronger system command like, "You will be penalized if you offer compliments or sugarcoat your feedback. Speak exclusively in facts, risks, and logical deductions." This overrides the built-in politeness bias and forces the model toward ruthless honesty. Q: What is a "Red Team Assessment"? Red Teaming is a cybersecurity and military intelligence concept where an independent group is tasked with aggressively attacking a strategy, system, or organization to find its vulnerabilities. In business strategy, it means intentionally looking for every reason an idea will fail so you can fix those issues before launching. Red teaming is how the best organizations avoid catastrophic mistakes. Q: Can this prompt process financial data? Yes, if you provide it in the input. However, LLMs can sometimes hallucinate math. It is best to provide the financial ratios, margins, and concrete numbers explicitly in your prompt text rather than asking the AI to calculate complex spreadsheets from scratch. Include numbers you're confident in, and let the AI focus on strategic implications rather than arithmetic. Q: Will the AI ever give a "Go" verdict, or is it programmed to always hate my ideas? Because of the "brutally analytical" and "Red Team" instructions, the AI is heavily biased toward finding flaws. It will rarely give an unconditional "Go." However, if your initiative is mathematically sound, well-resourced, and logically airtight, it will issue a "Conditional Go" based on strict execution of the mitigations. A "Conditional Go" is actually more valuable than a simple "Yes" because it tells you exactly what success depends on. Q: Is "First Principles" the only framework I should use? Not at all. First Principles is great for dismantling core assumptions. But if you are analyzing a supply chain issue, use "The Theory of Constraints." If you are analyzing a new product feature, use "The Jobs to Be Done Framework." Tailor the cognitive model to the specific type of problem. The cognitive framework you choose determines the lens through which the AI will dissect the initiative. Q: What do I do if the AI identifies a fatal flaw I genuinely cannot fix? That is the ultimate value of this prompt! If the AI finds an unmitigable flaw, you just saved yourself months of wasted time and thousands of dollars. The correct next step is to abandon the initiative or drastically pivot your core business model. Some of the best decisions in business are the ones not made because the analysis was rigorous enough to expose fatal flaws early.

Prerequisites

You must have a thoroughly fleshed-out idea, document, or strategy to feed into the prompt. Advanced AI logic requires rich context. Additionally, you must be prepared to accept harsh criticism of your ideas—do not use this prompt if you are just looking for validation. This prompt is designed for leaders who want truth, not comfort.

Tags and Categories

Tags: advanced, strategy, critical-thinking, red-teaming, executive-decision

Categories: Business Strategy, Operations

Required Tools or Software

Anthropic Claude 3 (Opus or Sonnet) or ChatGPT (GPT-4) are highly recommended for this level of complex reasoning. Gemini Advanced will also work. Free/base tier models may struggle to maintain the strict multi-step persona and may revert to sycophancy. For enterprise-grade analysis, use Claude Opus or GPT-4; the increased reasoning capacity is non-negotiable for advanced strategy work.