When I first heard about the rise of “prompt engineering” (the process of writing an instruction that will interpreted by generative AI and will result in the desired output) I was skeptical. I mean, seriously? We need a new job title for asking questions to AI? But I’ll admit, I was wrong. Once I started working with AI tools like ChatGPT to create more advanced and consistent outputs, I realized there’s an art and science to this process.
It’s not just about throwing a question or a request out there and hoping for the best. It’s about knowing how to guide the AI to give you exactly what you need. Whether you’re creating or building with AI, prompting is one of the biggest pieces to the equation and I truly believe it will become a table stakes skill for most professionals over the next two years.
If you’re running a small business, you don’t have a lot of time to learn complex tools. Maybe you tried AI and didn’t get something usable out of it, or maybe you’ve been waiting for someone to give you a clear and relevant use case. The good news is that there are tons of ways AI can save you time and headspace right now, and you don’t need to be a tech wizard to make it work for you.
With the right approach to prompting, AI can help streamline your operations, enhance customer engagement, and boost your marketing efforts. Let’s dive into the essentials of effective prompting and how you can start leveraging these techniques to increase your productivity.
The essentials of effective prompting
Clarity and specificity
Why it matters: Just like in any conversation, the clearer you are, the better the response. AI thrives on precise instructions.
How to do it: When crafting prompts, imagine you’re giving instructions to someone who’s completely unfamiliar with your business. They need context and direction before they are going to create anything useful.
DO NOT: “Write a blog post about our product.”
DO: “Write a blog post introducing our new eco-friendly cleaning product, highlighting three key benefits: sustainability, cost-effectiveness, and ease of use.”
Pro tip: If you have any docs, landing pages, or notes, you can feed it into the LLM (large language model) along with your request and it will pull what it needs to give you what you asked for.
Context
Why it matters: The more relevant details you provide, the better the AI can tailor its responses to your needs.
How to do it: Include background information and your objectives in the prompt.
DO NOT: “Write an email to our customers about our summer sale with 25% off outdoor furniture.”
DO: “Write a customer email announcing our summer sale, focusing on our best-selling outdoor furniture collection that is well-known for comfort and durability. Emphasize the limited-time discount and the high demand we experienced last year to drive a sense of urgency.”
Brevity and focus
Why it matters: While details are important, overloading the AI with unnecessary information can make it hard for the AI to follow all of the instructions you give it and harder to prioritize all of the information. It may leave important details out and overemphasize things that don’t really matter.
How to do it: Give it a clear goal or task to focus on before you give it the context. Define the outcome in a way that removes ambiguity.
DO NOT: “I need some social media posts that people will share. They should probably be relevant to our business, which makes and sells craft soda, but you can play on pop culture themes to get people’s attention. We want people to share them as much as possible, but we also want people to buy the soda. Can you also make some images to go with?”
DO: “Your goal is to create 10 diverse and highly shareable posts that will help drive brand awareness for our company. We make craft soda in unique and unexpected flavors like peanut butter and jelly. We’re a very colorful brand so each post should include an image.”
Types of prompting techniques
Let’s explore the different types of prompting techniques and how they can be applied to your business.
Zero-shot prompting
This is the simplest form of prompting. You ask the AI a question or give it a task, and it responds based on its pre-trained knowledge. It’s great for straightforward tasks where you don’t need to provide any examples. An example would be getting a quick answer to a common question:
“What are the most common social media channels for B2B?”
Few-shot prompting
Here, you give the AI a few examples to help it understand what you’re looking for. It’s perfect for tasks that require a bit more nuance. Let’s say you want to generate marketing copy in a specific tone, you could provide examples:
“Here are three past email subject lines we’ve used. Write a new subject line for our upcoming product launch in a similar style.”
Pro tip: You can actually create your own custom GPT and give it examples of your writing, brand guidelines, etc. to help it write in your desired style more consistently.
Chain-of-thought (CoT) prompting
This technique involves guiding the AI through a step-by-step process, much like brainstorming out loud. It’s ideal for more complex tasks like planning or troubleshooting. I’ve found that it’s important to give AI each step and clear directions for what to do with the information/context gathered in each. For example, if you wanted to generate messaging and positioning for new products or features:
“First I want you to review the attached presentation of our market and competitor analysis to identify the biggest risks and opportunities we’re currently facing. Next, I want you to analyze the product roadmap I’ve pasted below and identify which products specifically address the challenges and opportunities you identified. Finally, I want you to write three versions of high-level messaging and positioning for each with a short explanation of why each option made sense to you. “
Generated knowledge prompting
This type of prompting involves asking the AI to create new insights or suggestions based on its knowledge. It’s useful for creative brainstorming sessions or when you’re exploring new ideas. For example:
“Based on current trends in digital marketing, suggest three innovative ways we could promote our eco-friendly product line.”
Pro tip: Use Perplexity AI to research topics and dive deep on specific subjects. I love the way it lets you thread from your first question to cover a subject end to end—and I love how good the free version is.
Tree-of-thoughts prompting
This advanced technique is about exploring multiple pathways or outcomes, giving you a comprehensive view of different scenarios. It’s great for strategic planning, generating unique content, or identifying opportunities. For instance:
“Review the top trends for the renewable energy industry and outline three potential business expansion strategies for our company, consider the risks and benefits of each, and suggest which one we should pursue.”
Practical applications for your business
Here is how you can apply prompting techniques to different aspects of your business to significantly boost productivity and efficiency.
10x marketing productivity with a custom GPT
Use some of your best marketing materials to create a custom GPT that understands your business, brand, and products.
Step 1: Gather examples of your most effective marketing content—whether it’s blog posts, social media updates, or email newsletters.Step 2: Train a custom GPT model using these materials. This can be done through platforms like OpenAI’s fine-tuning API, where you can input your data and customize the model to match your brand’s voice.Step 3: Once trained, use this custom GPT for generating new content across all your channels—social posts, blog articles, newsletters, or customer communications. This will drastically reduce the time you spend writing while ensuring consistency and quality in your messaging.
Create more engaging content without a designer, animator, or producer
Leverage AI image and video creation tools to easily produce engaging content.
Step 1: Use tools like Midjourney or Krea.ai to generate stunning visuals for your social media posts or marketing campaigns.Step 2: For video content, use Runway to create and edit videos that captivate your audience. These tools make it easy to produce high-quality animations or video content without needing a full production team.Step 3: Integrate these visuals into your marketing materials using AI-enabled design platforms like Canva, which offers templates and design tools that make it simple to create professional-looking content. Whether you’re designing social posts, sales proposals, or marketing videos, these tools streamline the process and help you maintain a consistent brand aesthetic.
Customer research
Use AI tools to conduct thorough customer research, especially if you’re in the B2B space.
Step 1: If you’re exploring new industries or need detailed insights, use Perplexity AI to research and gather in-depth information on sectors you may not be familiar with.Step 2: Input queries related to your target market or potential clients to receive comprehensive answers that help you understand industry trends, challenges, and opportunities.Step 3: Leverage these insights to refine your marketing strategies, tailor your product offerings, and engage more effectively with potential customers.
Support and product documentation
Automate the creation of support and product documentation using AI tools that convert videos or transcripts into polished, user-friendly guides.
Step 1: Record a walkthrough of your product or service using Loom or another screen recording tool.Step 2: Use a tool like Scribe to automatically generate detailed documentation from your video or transcript. Scribe can capture your browser session, convert it into step-by-step guides, and even create instructional documents with screenshots and explanations.Step 3: Share these guides with your customers or internal teams to enhance their understanding of your products and reduce the need for manual support.
Triage, troubleshoot, and prioritize support requests
Use AI to efficiently manage support requests by identifying common issues and prioritizing new or unique problems.
Step 1: Implement an AI-powered support system to handle incoming queries. This system can be trained to recognize common questions and provide relevant, automated responses.Step 2: Set up the AI to flag support requests that don’t match up with existing documentation or previous queries. This helps you prioritize issues that require immediate attention or need new solutions.Step 3: Continuously update your support AI with new data from customer interactions to improve its accuracy and efficiency, ensuring that your support team can focus on the most critical tasks.
Mastering the art of prompting
The power of AI lies in how you use it. For small businesses, learning to craft effective prompts can unlock a world of possibilities. It’s not about being an AI expert; it’s about knowing how to ask the right questions to get the answers that will help your business grow. Whether you’re writing content, generating visuals, or automating customer service, the right prompts can turn AI into your most valuable assistant.
Start with something simple and build from there. With a little practice, you’ll find that AI can handle more heavy lifting, leaving you with more time to focus on what you do best—running your business.