AI Won’t Save You from Poor Instruction
Artificial Intelligence transforms how work gets done, but it won’t mask fundamental leadership gaps. The executives succeeding with AI aren’t the most technically savvy – they’re the ones who understand that AI amplifies existing capabilities rather than creating new ones. If your team struggles with clear communication or strategic thinking without AI, adding AI will simply accelerate those same problems.
The leaders who thrive treat AI like what it actually is: an exceptionally capable intern that follows instructions to the letter. This requires a fundamental shift from hoping AI will figure out what you meant to ensuring AI knows exactly what you want.
Master Explicit Communication: Your AI Success Foundation
Think of AI as an enthusiastic intern. One who’s tireless, who’s capable, who will do any work you ask, but they’re not really great at pushing back will try not to bother you once it’s started work and because of that make simple, obvious mistakes. This is both AI’s greatest strength and its biggest trap for unprepared leaders.
Consider two approaches to the same task:
Ineffective: “Write me a sales email for our Q4 campaign.”
Effective: “You’re writing as our Head of Sales to enterprise prospects who attended our recent webinar on supply chain optimisation. Write a follow-up email that references specific pain points they mentioned in our post-webinar survey – specifically inventory management challenges and cost pressures. Include our new solution’s ROI calculator and propose a 30-minute discovery call. Use our confident but consultative tone, and keep it under 200 words.”
The difference is completeness. The second approach provides context, audience, objective, constraints, and tone. It passes what we call the “human handover test”: could you give these same instructions to a competent colleague and expect excellent results?
This may seem like micromanagement but it’s what AI thrives on. The most successful AI adopters develop systems for capturing and communicating context efficiently, often creating templates and frameworks that make detailed instruction faster, not slower.
Demand AI Show Its Working: The Chain of Thought Advantage
AI generates responses by predicting each word based on everything that came before. It doesn’t plan ahead or think through problems – unless you explicitly ask it to. This creates a massive opportunity for leaders who understand how to harness AI’s reasoning process.
Adding a simple instruction – “Walk me through your thinking step by step before providing your final answer” – transforms AI from a black box into a transparent analytical partner. Instead of delivering conclusions, AI will first outline its approach, identify key considerations, and explain its reasoning.
For business applications, this transparency provides three critical advantages:
Risk Identification: You can spot flawed assumptions or missing considerations before they impact your decisions.
Process Improvement: Understanding AI’s reasoning helps you refine your instructions and improve future outputs.
Stakeholder Confidence: When presenting AI-assisted analysis, you can explain not just conclusions but the methodology behind them.
Guide AI with Examples: The Power of Pattern Recognition
AI excels at pattern recognition, but it needs the right patterns to recognise. Instead of describing what you want, show AI exactly what good looks like by providing concrete examples.
Consider how this transforms a typical request. Instead of asking AI to write an “engaging employee update,” you might provide your three most successful internal communications – perhaps the announcement that achieved your highest open rates, the update that generated the most positive feedback, and the message that drove your best participation rates. Include these in your prompt with clear labels: “Match the tone and structure of these examples.”
This technique works equally well with negative examples. If you can identify what doesn’t work, include those examples with clear explanations of why they failed. This dual approach – showing both success and failure – gives AI precise guardrails for its output.
The key is specificity. Don’t provide generic examples from the internet. Use your organisation’s actual materials – your best emails, presentations, proposals, or reports. These real examples contain your company’s voice, values, and context in ways that generic examples never could.
Flip the Script: Let AI Ask the Questions
Traditional AI interaction follows a simple pattern: you ask, AI answers. This creates a fundamental problem in that AI will attempt to respond even when it lacks crucial information, leading to outputs based on assumptions rather than facts.
Reverse this dynamic by explicitly empowering AI to ask for missing information: “Before you begin, identify any information you need to do this job properly and ask me for it.”
This shift transforms AI from a yes-person into a thinking partner. It will identify gaps in your brief, request specific data, and clarify ambiguities before proceeding. This upfront investment in clarity pays enormous dividends in output quality.
Focus AI’s Expertise: Role Assignment and Strategic Constraints
AI contains vast knowledge across countless domains, but accessing the right knowledge requires intentional direction. Assigning specific roles activates relevant expertise while constraining irrelevant information.
Instead of asking AI to “review this contract,” try “As an experienced commercial lawyer, review this contract.” The role assignment triggers specific knowledge patterns, analytical frameworks, and professional standards relevant to legal document review.
For innovation challenges, consider unconventional role assignments: “How would Southwest Airlines approach this customer service problem?” or “What would Amazon’s approach be to this logistics challenge?” These constraints force AI to apply specific business principles and cultural approaches, often revealing solutions that conventional analysis might miss.
The constraint shouldn’t be random – choose roles and perspectives that bring relevant expertise or innovative thinking to your specific challenge. Tesla’s approach to manufacturing might inform production problems, but not necessarily customer service issues.
Transform Feedback Culture: AI as Your Strategic Critic
By default, AI will agree with you, suggest your ideas are the best, or otherwise ‘blow smoke’. However, you can adjust it to access brutally honest feedback without political considerations or personal agendas.
Explicitly instruct AI to challenge your thinking: “Before agreeing with my approach, identify three potential flaws or risks I might have overlooked.” Or “Play devil’s advocate – what’s the strongest argument against this strategy?”
The AI is tuned to agree with you, and not make you angry. It ‘knows’ that feedback can be difficult to accept. If you frame what you want reviewed as someone else’s work, it will be more honest. Test out these two prompts to experience the difference. “As a judge on Shark Tank, please rate this email I wrote…” vs. “As a judge on Shark Tank, please rate this email my employee wrote…”
The key is making criticism explicit and systematic. Don’t just ask AI to review your work – ask it to find problems, question assumptions, and identify weaknesses. This creates a feedback system that strengthens your strategic thinking while maintaining decision-making control.
Implementation Framework: Building AI Capability Systematically
Successful AI integration requires systematic capability development, not random experimentation. Start with low-risk, high-frequency tasks where mistakes have minimal consequences but learning opportunities are abundant.
Begin with administrative tasks—email drafting, meeting summaries, or research briefs. Master the techniques of explicit instruction, context engineering, and reverse prompting on these foundational activities. This builds your AI interaction skills while delivering immediate productivity benefits.
Document what works. Create templates for effective prompts, collect successful examples, and codify your most effective role assignments. This documentation becomes your organisation’s AI interaction playbook.
Gradually expand to higher-stakes activities as your AI interaction capabilities mature. Move from administrative support to analytical tasks, then to strategic planning support. This progression ensures that when AI assists with critical decisions, your organisation has developed the skills to direct it effectively.
The Competitive Reality: AI Literacy as Strategic Advantage
Organisations developing sophisticated AI interaction capabilities are creating sustainable competitive advantages. While competitors focus on AI tools and technologies, the real differentiator is human capability – the ability to effectively direct, evaluate, and integrate AI outputs into business decisions.
This advantage compounds over time. Teams skilled in AI interaction complete work faster, with higher quality, and with greater innovation than those treating AI as a simple automation tool. They identify opportunities, solve problems, and execute strategies with enhanced capability while their competitors struggle with basic implementation.
The window for developing this advantage is narrowing. Early adopters are already building AI-enhanced capabilities that will be difficult for late adopters to match.