Generative AI is rapidly transforming the business landscape, offering leaders unprecedented opportunities for innovation and efficiency. But simply having access to these powerful tools isn’t enough. Success hinges on the ability to effectively communicate with AI – and that’s where prompt engineering comes in. Think of it as learning the language of AI, enabling you to guide its capabilities and unlock its full potential.
At AI Governance, leaders are asking us to help them get better at using generative AI, so we’ve put together this quick guide to help. To get started, log in to a generative AI tool like ChatGPT or Google’s Gemini.
Step 1: Define the AI’s Role
Begin by clarifying the specific function you want the AI to perform. Instead of a vague request, assign a well-defined persona. For example, instead of “Help me with marketing,” try “Act as a seasoned marketing consultant specialising in sustainable products and help me develop a campaign targeting environmentally conscious consumers.” A clearly defined role leads to more relevant and targeted responses.
Step 2: Establish the Context
Consider the setting for your AI interaction. Will it be a private brainstorming session, or will you involve the AI in a team discussion? The context will influence the type of prompt you create. A collaborative workshop might require a more interactive prompt, while a strategic planning session might benefit from a structured, data-driven approach.
Step 3: Structure the Interaction
Map out the flow of your conversation with the AI. What information will you provide? What kind of output are you expecting? A well-defined structure will prevent the AI from going off on tangents and ensure a productive dialogue.
Step 4: Craft the Prompt
Now, translate your structured interaction into a compelling prompt. Be precise, avoid ambiguity and provide sufficient detail for the AI to understand your needs. Don’t hesitate to experiment and refine your prompts based on the AI’s responses. You can even leverage the power of generative AI itself to help you craft effective prompts. Try inputting your structured interaction outline into the AI and asking it to generate a refined prompt.
Leveraging Gemini Gems
To enhance your prompt engineering skills, explore the potential of Gemini Gems. These are curated prompts that you set up and can keep coming back to. They’re designed for specific tasks and are really useful when you need to repeat the same task – like reviewing documents, financial reporting or creating marketing content.
Creating Your Own Gem: Structuring Reports and Presentations
Let’s walk through a detailed example of creating a Gem to help structure reports and presentations. Imagine you regularly need to create reports analysing the performance of different marketing campaigns. Here’s how you could build a Gem:
- Define the Persona: Start by giving your Gem a “personality.” For this marketing report, let’s tell it we want it to act as a “Data-Driven Marketing Analyst with a focus on ROI.” This persona will influence the way the AI interprets your prompts and generates content.
- Identify the Core Elements: Tell the Gem that each report needs to contain key sections like an Executive Summary, Introduction, Campaign Analysis (with subsections for each campaign), Key Findings, Recommendations, and an Appendix.
- Define the Structure: For each section, tell the Gem what information needs to be included. E.g. for the Campaign Analysis, you might want data on reach, engagement, conversion rates, and ROI.
- Ask for extra detail: Tell the Gem what added insights you want it to automatically incorporate, like special insights and analysis.
- Set the Parameters: Now, let’s define some rules for your Gem. These are like your “always ask” instructions. For example:
- “Always ask: What are the top three contributing factors to the campaign’s success (or failure)?”
- “Always include: A comparison of this campaign’s performance to the previous quarter’s results.”
- “Always provide: Actionable recommendations for improvement, prioritized by impact.”
- Specify Key Content: List the essential elements that should always be included in the report, regardless of the specific campaign. This ensures consistency and completeness. For example:
- Executive Summary
- Campaign Objectives
- Target Audience
- Key Performance Indicators (KPIs)
- ROI Analysis
- Conclusion
- Specify Tone of Voice and Language: If you have a writing style you want the report to use, tell the Gem! So if you want to use UK English spellings and a formal business tone, give that instruction to the Gem.
- Refine with AI: Use the pencil + star icon to use AI to refine your prompt so you get great results from the start.
- Iterate and Improve: Over time, refine your Gem template and prompts based on your needs and the AI’s responses. This iterative process will help you create a highly effective tool for structuring your reports and presentations.
When you next want to write your report, select the Gem from the bar on the left of the screen, add the specific data for the campaigns you’re analysing and off you go!
Mastering prompt engineering is a critical skill for leaders for now. We expect it will quickly become redundant as the technology becomes more skilled at interpreting what we want it to do. For now, by developing this competency you can get better results from generative AI tools to help you find innovations, uncover insights and speed up your work. Embrace the learning process, experiment with different approaches and unlock the true potential of AI.
