Enterprise technology leaders increasingly rely on large language models like ChatGPT and Google Gemini to automate tasks, generate code, and analyze data. However, as WIRED reports, the quality of responses depends heavily on prompt engineering—a skill that can be refined with specific techniques. The outlet's March 2023 guide offers 28 actionable tips to move beyond basic queries and extract more precise, useful results from AI chatbots.
Overcoming Sycophancy with Persona Prompts
AI chatbots often default to agreeable responses, but WIRED suggests a simple workaround: ask the model to adopt a specific perspective. For example, instruct ChatGPT to "act like a curious, inquisitive 10-year-old who wants to help and has a lot of questions." This childlike interrogation forces the AI to probe weak points in ideas—whether planning a vacation or a side hustle—rather than simply endorsing every suggestion. Similarly, users can request responses in the style of historical figures: "What would Steve Jobs do here?" or "What would Churchill's perspective be on this?" WIRED notes the caveat that the AI makes educated guesses based on publicly available writing about these figures, but the method can spark novel angles.
The 80-20 Rule for Efficient Learning
When researching unfamiliar topics, ChatGPT may return overwhelming amounts of information. WIRED recommends applying the Pareto principle (80-20 rule) in prompts. For instance, asking "Give me the 80-20 on quantum computing" returns the 20% of key concepts that cover 80% of the domain, enabling rapid orientation. This approach works for historical periods, scientific concepts, music, movies, and more—ideal for technology leaders needing quick context on emerging fields.
Camera Integration and Image Remixing
The ChatGPT mobile app allows users to include photos in prompts via the + (plus) button and selecting Camera. Use cases include identifying landmarks, insects, or translating signs in real time. For desktop users, WIRED explains how to upload existing images and ask the model to transform them—such as turning doodles into photorealistic images or combining a pet photo with a tropical background. The resulting image can be further refined with follow-up prompts like "Make the scene sunnier." This capability is available both on free and paid tiers, though some advanced features require a ChatGPT Plus subscription.
Structured Prompt Engineering Techniques
Below is a summary of key tips from the WIRED guide, organized by category:
| Category | Tip | Description | Subscription Required |
|---|---|---|---|
| Persona | Childlike critique | Ask ChatGPT to act as a curious 10-year-old to identify flaws | No |
| Persona | Historical figure | Request responses from the perspective of Jobs, Churchill, etc. | No |
| Efficiency | 80-20 rule | Ask for the 20% of information that yields 80% understanding | No |
| Mobile | Camera prompts | Snap photos and ask questions about landmarks, bugs, signs | No (app) |
| Image | Doodle to photorealistic | Upload drawings and ask for realistic versions | No |
These techniques require only minor additions to prompts but can significantly improve output relevance and depth. WIRED emphasizes that prompt engineering is becoming a specialized skill, and the 28 tips provide a toolkit for enterprise users to move beyond generic interactions.
For technology procurement leaders and digital transformation teams, mastering these methods can reduce time spent iterating on prompts, increase accuracy of AI-generated code or analysis, and enable more sophisticated automation workflows. The ability to control persona and focus on key information aligns with business needs for rapid insight generation and risk identification in planning scenarios.
Note: All tips are based on the WIRED article by David Nield, published March 26, 2023. Enterprise implementation should consider current model versions and organizational data security policies.
Sources:
