Today, let’s talk about how you can harness AI as a tool to uncover meaningful insights, refine processes, and empower your teams. Human Factor Digital is not just about digital services for your website – we can also help you unlock the power of AI through smart, effective prompt engineering. We’ve observed excellent results from implementing these strategies in 2024 with ChatGPT, Google Gemini, Adobe AI Assistant, and Microsoft Copilot; we’ll update this guidance periodically as the technologies evolve in 2025 and beyond.
1. Set Clear Goals for Every Prompt
It sounds obvious, but the AI doesn’t know what’s in your head. Be explicit about your objective in every prompt. This clarity helps the AI zero in on what matters and skip the filler.
- Bad Prompt: “Summarize this data.”
- Good Prompt: “Summarize this data to highlight key performance indicators.”
- Better Prompt: “Summarize this data to show which KPIs improved compared to last quarter.”
- Best Prompt: “Summarize the data to show improvements in KPIs from last quarter, identify specific factors contributing to growth, and suggest one or two areas for additional focus next quarter.”
2. Specify the Output Format You Want
Save time on reformatting by telling the AI up front how you’d like the response structured.
- Bad Prompt: “What are some recent industry trends?”
- Good Prompt: “List five industry trends in digital marketing.”
- Better Prompt: “List five industry trends in digital marketing in bullet points, with a one-sentence summary for each.”
- Best Prompt: “List five digital marketing trends relevant to Q2 2023, with a short description, any known benefits, and examples of companies leveraging these trends. Format in bullet points for quick reference.”
Note: as of the date of publication of this article, ChatGPT’s most current knowledge dates to October 2023.
3. Break Down Complex Requests
Rather than asking for a big-picture analysis in one go, split it up. This keeps the AI focused and ensures each aspect of your question gets proper attention.
- Bad Prompt: “Analyze customer complaints in these reviews and suggest how to improve.”
- Good Prompt: “Identify the top three customer complaints in these reviews.”
- Better Prompt: “Identify the top three customer complaints in these reviews, then list common themes.”
- Best Prompt: “First, identify the top three customer complaints in these reviews. Next, analyze any recurring themes among these complaints, and finally, suggest specific, actionable improvements to address each theme.”
4. Use Open-Ended Questions to Spark Ideas
When exploring new territory or brainstorming, invite the AI to think creatively by asking open-ended questions. This flexibility allows for innovative suggestions that might surprise you.
- Bad Prompt: “List some potential issues with switching to video content.”
- Good Prompt: “What issues might arise if we switch to video content?”
- Better Prompt: “What are three potential issues we might face if we switch to video content? List any major risks.”
- Best Prompt: “Identify potential issues with switching to a video-focused content strategy. Include any challenges with production cost, audience engagement, and distribution, and suggest strategies to address each.”
5. Be Specific with Scope and Detail
For larger projects or complex data sets, let the AI know just how deep to dive.
- Bad Prompt: “Review our sales data.”
- Good Prompt: “Review Q3 sales data and identify trends.”
- Better Prompt: “Review Q3 sales data and identify any underperforming products.”
- Best Prompt: “Review Q3 sales data and provide a high-level summary of underperforming products, focusing on those that showed a drop in revenue by 10% or more compared to Q2. Offer possible reasons for this decline.”
6. Ask for Examples or Evidence
Making analysis concrete helps you spot actionable insights faster. Asking for examples or evidence lets you see how the AI reaches its conclusions, which is critical in data-heavy scenarios.
- Bad Prompt: “Tell me how customer complaints changed over time.”
- Good Prompt: “Identify examples of recurring customer complaints.”
- Better Prompt: “Identify three recurring customer complaints with examples from reviews.”
- Best Prompt: “Identify three recurring customer complaints, give examples from the most recent reviews, and note any changes in frequency or severity over the last three months.”
7. Build Continuity in Conversations
Referencing prior prompts can keep your analysis consistent and cumulative. Just like any conversation, AI can “build” on previous responses for a more robust, ongoing analysis.
- Bad Prompt: “What can you tell me about customer satisfaction?”
- Good Prompt: “Analyze customer satisfaction based on recent data.”
- Better Prompt: “Analyze customer satisfaction trends in our latest quarterly data, focusing on Q3 feedback.”
- Best Prompt: “Following our last analysis on customer satisfaction, evaluate the Q3 feedback to identify any new patterns, especially around service speed and product quality, and note whether prior issues have improved.”
8. Use Neutral Language for Objective Analysis
Neutral language helps the AI deliver responses that are fact-based and objective – essential for any data analysis or strategic review.
- Bad Prompt: “Tell me what’s wrong with our customer service reviews.”
- Good Prompt: “Analyze common themes in customer service reviews.”
- Better Prompt: “Analyze common themes in customer service reviews, focusing on areas of improvement.”
- Best Prompt: “Analyze common themes in customer service reviews, highlighting recurring issues and any specific patterns related to response time, communication style, or problem resolution. Avoid subjective language and provide evidence-based insights.”
9. Experiment with Hypothetical Scenarios
To get a sense of “what if,” try hypotheticals. This is especially handy for forecasting, risk management, or testing different strategic approaches.
- Bad Prompt: “How would focusing on Product B affect sales?”
- Good Prompt: “How could shifting focus to Product B impact overall revenue?”
- Better Prompt: “Analyze how shifting our primary focus to Product B could affect our quarterly sales.”
- Best Prompt: “If we shift focus to Product B next quarter, estimate its impact on total revenue, considering customer demographics, market trends, and any potential trade-offs with other products in the lineup.”
10. Tailor the Tone and Language for Complex Analysis
When it’s a high-stakes question, use language that encourages the AI to engage with more “thoughtful” analysis.
- Bad Prompt: “Tell me if we should change vendors.”
- Good Prompt: “Analyze the pros and cons of changing vendors.”
- Better Prompt: “Assess the advantages and disadvantages of switching vendors, with a focus on customer satisfaction and cost.”
- Best Prompt: “Considering our priorities of cost savings and customer satisfaction, analyze the advantages and disadvantages of switching vendors. Include potential impacts on service quality, timelines, and customer loyalty.”
Conclusion
These techniques are just the beginning. With prompt engineering, Human Factor Digital helps companies turn AI from a generic tool into a customized powerhouse for solving your unique business challenges. If this strikes a chord with your goals for 2025, reach out. We’re here to bridge the gap between AI’s potential and practical outcomes that move your business forward.