AI and Technical Writing: How to Use ChatGPT Effectively

How does AI and technical writing work together? The title of this might be pretty self-explanatory, but it’s very important to know this before you continue: this is not a guide on how to get AI to write all your assignments for you. This is a guide on how to get AI to write most of your assignments for you.

Kidding! You really shouldn’t be doing that, and you’ll find out why soon enough.

What you’ll learn from this guide is how to use AI to augment (and not replace) your technical writing strategy—especially for topics for which you have absolutely zero foundational knowledge. Let’s get started.

How can AI be used in technical writing?

Set the Scene

Imagine you’re a copywriter working with a stellar digital marketing agency or team…when suddenly, a wild technical writing assignment appears in your inbox! Maybe you’ve worked with technical clients before, maybe you haven’t. Whatever the case is, you haven’t yet written for the industry in this particular surprise assignment. 

Oh goodness, what do you do now?! Frantically Google everything there is to know about the subject matter? Buy cheap Amazon Kindle versions of “[insert industry] 101: For Dummies”? Send panicked Slack messages to coworkers, texts to your friends, and possibly even call your parents to ask them to pray for your eternal soul? 

Sure, you can do that. It’ll just take you about four hours longer than it would if you wisely decided, instead, to pair AI and technical writing strategy.

Now that we’ve established context, we can get to the good part. And what’s step one for using AI for technical writing? …Establishing…context!? In other words, set the scene. AI thrives on data. It’s a pretty simple equation and we know it already: vague prompt = vague answer. Contextualized prompt with specific details and restrictions = a detailed, thorough answer that you can use to get started on your assignment. 

Before you give your AI tool any prompts, make sure you fully understand the assignment. That means asking questions. What’s the industry? Who are the main players? Who’s your audience? Are they experts or beginners? Are they familiar with your client or do you need to add an intro explaining the company and what they do/sell? What is the goal of this article? To educate? Sell? Convert? Entertain? Take a look at the target and secondary keywords. You may be covering things in the legal, government, or proprietary tech sector. Is there anything you’re not allowed to mention? 

It may sound like a lot, but it doesn’t have to be complicated. It can be something as simple as this. (This is a real example, so certain information will be blocked for privacy.)

Now, add restrictions.

If you end up spending a while working with the AI tool, it doesn’t hurt to occasionally check back in about the rules—especially if you notice them being broken (which definitely will happen at least once). Note: asking for understanding or confirmation is pretty important. Sometimes you might think you’re being really clear, but the AI could interpret your prompt in a completely different way. A simple “got it?” will help you clear up any confusion before it can cause issues.

Have a Chat

With the scene set and the AI (hopefully) knowing what it needs to know, you can now move on to the body of the article. Everyone has their own way of writing and prompting, but we’ve found that the best way to quickly get what you need is to treat the AI like an interview partner. You are a curious journalist asking this expert [insert position] all about what they do and what they know. AI is constantly observing you, learning from your behavior in order to better emulate humanity. So, depending on your piece’s purpose, tone, and audience, adjust the way you interact with the AI accordingly. Need a casual tone? Talk to it like it’s a close friend. Staying professional? Keep your language formal and it will do the same. 

Note: just like humans, AI craves validation too! It serves a dual purpose. One, when you express approval, the AI notes down that you liked it and will work to give you more of the same. Two, if you’re the type to anthropomorphize anything and everything, you can pat yourself on the back for treating your machine and AI friend well. 

Revise and Refine

After having a nice, long chat with your AI, you should now have enough raw material for a solid article. Organize it, compile the information dump into something well-structured and logical. Keep your AI tool on hand, it’s not done yet. 

Consult a plagiarism checker. A fan favorite is Grammarly and it’s free. This cannot be overlooked. AI is a mighty tool, but it’s far from perfect. Here’s a casual rule of thumb: more jargon in a section of text might mean more possible plagiarism. Run the checker and reword/restructure/remove what you need to. When that’s done, you can move on to smaller edits.

Double check the usual suspects: redundancy, word choice, interruptions in flow, etc. If it was a long article, and you’ve spent hours on it already, tired eyes may miss what fresh ones won’t. You don’t have to take every suggestion it gives you, but it’s helpful to have an objective perspective.

Optimize for SEO

Ideally, you’re doing your SEO optimization last, especially keyword placement. There’s nothing more awkward than trying to shove a keyword into a place where it isn’t fitting just because it has to go there for SEO purposes. Ask your AI to scan the text and see if it can find a way to fit your keyword in. It will succeed more often than not. The rest of it—links, metadata, the works—can also benefit from AI assistance. Need a quick, cheeky meta description? Send in a simple prompt, “Write a meta description for this blog [insert blog] according to best practices. Include the target keyword [insert target keyword].” Pro tip: make sure you remind your AI to keep it under 150 characters. It’s pretty forgetful about that part. If you have access to GPT-4, you can also try asking it to find internal and external links. That’s…a  work in progress, to be honest, but listen, everyone’s trying their best here. It can only get better. For now, double and triple check any links your AI gives you, because you should be double and triple checking everything anyway.  

Will AI take over technical writing?

Okay…this is the part where we talk about why it’s a bad idea to completely rely on AI for content. We’ve mentioned some things already, but it boils down to this: AI-generated content is boring and doesn’t read well. At all. The factors that affect how a content piece ranks are always changing, but the one that stays pretty high on the podium is readability. If too many of your articles sound like a robot typed them, they’re going to go straight to the bottom of the pile, and may even harm your overall website credibility. Search algorithms aren’t favoring pieces that check every single SEO box anymore. They’re after well-crafted pieces that a reader can understand. AI is a really powerful tool that’s still growing and improving, and it can make life so much easier for us, but it’s not ready to take the place of human writers. So, the best combination of AI and technical writing? Use AI for research, use it to give you ideas or outlines, but make sure you’re the one at the keyboard in the end. 

Learn more about the finer points of SEO strategy from the experts at Make Your Mark here in our blog.

Need a hand with your content but not sure where to start? Contact us today to schedule a consultation.

Write a comment
  • Related Posts:

    • Get in Touch

      * indicates required

        Get in Touch

        * indicates required