It’s called Prompt Engineering and it’s the next big skill for marketers.
You Learned SEO, Now You Need to Learn Prompt Engineering
If your’e old enough to remember the early days of the internet, one of the biggest needs and challenges was getting found online. That’s when the first search engine marketers came on to the scene; writing fun, informative, readable copy that had your content or website show up at the top of search engines. It’s was (and is) an in-demand skill.
Meet Your New Employee — the AI Writing Tool
The first AI writing tools were article spinners (blech!) they were downright awful. But when the folks at Open.AI launched ChatGPT; a chatbot that not only engages with you back and forth, it does so plain English. So much so, that it’s hard to believe that it’s just a machine.
But here’s the deal — it IS JUST A MACHINE — an amazing machine. But still a machine. And, the key to getting what you want out of this machine — is knowing what to ask and how to ask.
When you think about it – it’s like most things in life; garbage in, garbage out. And you’ll get whatever you want if you just ask nicely.
And this, my friends is why prompt engineering is your next big skill — you gotta know how to ask nicely to get what you want and avoid the garbage.
Table of Contents:
What is Prompt Engineering?
Definition of Prompt Engineering: Prompt engineering is task of finding the prompt that gets you the best output from the AI.
Prompt engineering is the skill of giving the AI context, instructions, expectations, and examples of what output you want to see.
Think of AI as an Employee and the Prompt as Your Instructions
When you start engaging with AI, it’s like you’ve added a new “person” into your life. This is nothing new If you have Siri, Google, or Alexa in your home or office, you’ve already gotten some practice with “voice prompts”. There is a “trigger word” that turns them on, you ask your question or give your command, and there is an output or outcome.
AI tools are kind of like new employees. And what do you do with new employees? You have to train them, educate them, verbalize details about what you want and why you want them a certain way so that they can give you exactly what you’re looking for.
Key Takeaway: Prompt engineering is an AI-driven process used to generate automated prompts for marketing campaigns, emails, ads and more. It provides benefits such as increased efficiency and productivity, improved customer engagement & retention, cost savings and higher ROI.
Why is Prompt Engineering Important for Marketers?
I hear you — more skills you need to learn. But hear this, if you want to up your productivity game and feel like you have a team at your fingertips, you’re going to want to lean in.
The most common complaint with small business owners is that they don’t have the time or the money to do marketing. Well, AI is here to help.
Since AI tools like ChatGPT can do anything from writing emails to video scripts to CODE (I’m serious — it writes code) you will now have time for concentrating on what YOU do and using the AI to do a lot of the tedious heavy lifting,
Saves time and money: Once you learn how to give solid directions to your AI tool, become more of an editor and customizer than a creator. If you’ve ever hired a copywriter, given them instructions, and gotten sub-par content back, but paid full price. It’s kind of like that — except that you’re not paying $300 for something you have to rewrite.
Brainstorming and idea generation: Let’s be clear about something. AI isn’t going to deliver anything new. For example, if you tell it to write like Hemingway — there has to be a Hemingway! However, AI is amazing at going into the database, finding connections and predicting what should be next. This makes it an outstanding brainstorming buddy. It’s great for generating headlines, creating outlines, writing emails, etc.
In other words, prompt engineering is as valid a skill in marketing as writing, doing presentations, webinars, and using social media. It won’t replace you — it will make you better, stronger, faster.
Key Takeaway: Prompt engineering can help marketers increase efficiency and productivity, while also improving customer engagement and retention. Benefits include: automated tasks, personalized messages, tracking of customer interactions, increased efficiency & reach.
How to Become Good at Prompt Engineering?
OK, so let’s get into the nitty gritty of prompt engineering and how to develop and use this valuable skill. In fact, if you follow these steps, you’ll become a MUCH better project manager, boss, employer, and leader.
No kidding.
Know what you want. Be very clear about what your ideal outcome looks like
This is, perhaps, the most important thing of all. Don’t expect the AI to read your mind. It’s ok to brainstorm but don’t get your expectations set impossibly high.
Imagine you are talking to a person who is just a level below you in experience (perfect – right?)
The next step is to imagine that you are talking to a person who knows how to do certain things, but need the background information.
Step 1: Provide Context
AI tools need context: The first thing you do is tell them to “Act as…”
You wouldn’t walk up to someone in the street and say something like “Come up with 5 headlines” — right? Of course not. If you hired an assistant and you were going to give them work to you, you’d say something like —
“Act as a marketing agency that works with hair salons.”
Provide as much context as a person would need to get themselves in the right headspace to deliver what you want.
Step 2: Give it some data
AI tools need some type of data set for direction and content
“I’m going to give you [an example, a blog post, etc]”
Just like you would with a person, you want to give the AI software a data set to work from. It doesn’t have to be long, but it does need to contain the vital information that the software will use to create the content output you’re after.
In this case, I gave the AI a little more direction — I also combine step 2 and step 3 — but the content needs to be there so that I can get the output I’m after.
Step 3: Define what the output looks like
AI tools need to know what the output should look like
The next step is to tell the AI what you want as a deliverable…
- “Generate 5 headline ideas for a series of blog posts about how to increase sales at your hair salon.”
- “Generate a social media campaign that features our new hair product line “NAME HERE”. Include the promo code “SAVE20″. Include relevant hashtags. End each post with this URL: 123.com/promo. Write in the first person. Include fun emojis.”
One way to do this is to do the task yourself first. Not an entire campaign, but write that social post. Or find a similar article that you’d like to emulate and make a list of the specific attributes that you like and would like to include in your content.
Key Takeaway: To become proficient in prompt engineering, marketers must develop an understanding of AI technologies and tools such as ML, NLP, and popular frameworks like TensorFlow. Additionally, they should practice writing effective prompts for different types of content marketing campaigns to hone their copywriting skills.
Start Practicing Your Prompt Engineering TODAY
I get it. No one is excited about learning yet another new marketing skill. But there are so many other benefits of getting good at prompt engineering.
You’ll be able to outsource a ton of tedious work to an AI rather than a team.
You’ll become a much better and clearer communicator
You’ll be pushed into becoming more of a thought leader, developing a clearer point of view.
Conclusion
All I can say is do NOT fight this AI thing. Embrace it! Embrace it the way you did dishwashers, cars, calculators, the internet, and GPS. Like any mechanical tool, it will make you weaker at some basic tasks but it will push you to become stronger in others.
You’ll stop wasting time with word-smithing dozens of emails or staring and a blank screen while you toss around ideas. Instead, you’ll get some content back, you’ll have an opinion about that content and you’ll make it better.
PS: A personal case study about this article — if you’re interested
I’ve been testing a long-form, SEO writing tool called Content At Scale. My experience with this tool is a big driver of the content I created above.
If you’re an agency or startup niche site, this AI content writing tool will help you produce excellent quality articles in minutes.
This is a new and layered technology that allows you to go from keyword to full keyword optimized article; that includes, keyword, title, meta description, sections, call outs and so much more.
All articles are unique and practically ready to publish.
We earn a commission if you click this link and make a purchase at no additional cost to you.
Here’s what I gave the tool as a prompt:
Content at scale requires a keyword phrase and some context. To start an article, all you need to do is fill this out.
Content At Scale uses 3 different types of AI tools; not just GPT-3, but a few others. Content at Scale also provides your post a Headline, and meta description which is really great.
However, based on the content I got back, my prompts left quite a bit to be desired.
I’m going to share as much as I can of the content the AI gave me and I’m going to share what I thought and what I learned, and how I would change my prompt.
As you can see, the AI content on the left is bereft of soul, personality, etc. This is because my prompt didn’t add important instructions that might have included even more context such as:
- You are a small business marketing expert
- Write in a friendly tone
- Write in the first person
I did give it section headings, the AI’s content was very technical and didn’t really match the exactly context I was after.
Here the AI went off on chatbots, customizing experiences, etc. And, that’s NOT what this article is about. It’s about how to develop prompts.
That brings me to another important lesson — how do YOU prefer to work.
Choose a tool that works the way you work
Like everything else, there’s a tool out there for everyone. I’ve learned that all-in-one long form content is NOT good for me. I prefer to be an iterative content creator.
Most AI tools charge based on the number of words used and when you generate a complete blog post of 1700 words and the context is off (as was the case for me), that’s a complete waste of time and money. I had to rewrite this whole article, and the points cost me about $30 per article. Hey, it’s not $300 but I wanted to show you how important getting good at Prompt Engineering is.
The better you get at prompt engineering, the better content you’ll get in less time. And this will free you up to develop NEW points of view and to become a better thought leader.