Content marketers have been some of the first to go from "just playing around with AI" to "using AI in my daily work." To find out what parts of AI content creation have had the most impact, I asked other marketers how they're using AI tools to develop content. Here's what they had to say.
1. Use AI to generate content ideas
Ideation is the first step in the content process, but many marketers struggle to consistently develop interesting ideas. When creative block sets in, you can use AI tools to generate new ideas—whether that's for blog articles, social media posts, newsletters, or any other type of content you're creating.
Vic, customer success and community manager at Remotive, loves AI's brainstorming power. Here's what she told me:
"I use ChatGPT to generate ideas for social media and community engagement. After three years of managing our community, it's sometimes difficult to think of new ideas. But ChatGPT helps me come up with topics for our daily watercooler threads and LinkedIn polls."
On the visual side, Shane McEvoy, managing director at Flycast Media, says he combines the AI image generator Midjourney with ChatGPT to create attention-grabbing social media posts:
"Chat GPT's ability to condense several paragraphs of text into well-written tweets and hashtags helps me create engaging content that resonates with my target audience. Midjourney, on the other hand, provides me with unique, high-quality concept images that are not available in stock image libraries. This helps my social media posts stand out, as they have a unique visual appeal that captures the viewer's attention."
When it comes to social media content, AI tools can brainstorm post ideas, hashtag suggestions, and image captions in seconds. But apart from generating these ideas from scratch, they can also analyze your existing content, extract key messages, and then use all that to write engaging social media posts.
And it works for blog content too. Use an AI chatbot to brainstorm a list of topic ideas to draw inspiration from. All you'll need to provide is a broad topic or keyword and some context about your blog and company. Then ask it: "Brainstorm 10 unique blog topics related to [topic] for the [company] blog." The more specific you can be, the better, but sometimes letting AI do its thing will help you come up with ideas you'd never have thought of otherwise.
2. Create content outlines
Once you have your content ideas, AI tools can develop a potential structure with headings and talking points to get you in the writing zone. Certain AI writing tools offer this as a template, but you can get similar results from ChatGPT or another chatbot if your prompt is detailed enough.
Michael Maximoff, co-founder and managing partner at Belkins, says his team is primarily using ChatGPT in this capacity.
"Our team primarily uses AI tools to generate ideas and content outlines. Specifically, we use ChatGPT to create content ideas and potential structures for our articles. We then tweak these structures until we are satisfied with them. All of this, of course, can be done manually, but AI tools really help to reduce the workload and speed up the process."
When I asked ChatGPT to create an outline for this article, here's what it came up with.
This looks decent. It organizes the ideas into a logical structure that can get my creative juices flowing. And if you write thousands of words every week, as I do, these AI outlines will help you create content at scale.
But there's a major caveat: the outline ChatGPT generates will basically be a copycat of other articles that already exist. Using it as it is can remove originality from your work and throw it into the sea of sameness. So, like Belkins, spend time reviewing and tweaking the output to find new angles that will make your piece stand out.
3. Write first drafts
With your content ideas and outline in hand, you can use ChatGPT or other AI text generators, like Jasper or Writesonic, to generate your first draft. Depending on the tool and your process, our little robot friends can write a complete long-form article at once—or they can do it in smaller segments, with more input from you along the way.
The team at Verblio has found that AI works best when they use it to build content in segments. Head of AI projects, Megan Skalbeck, tells me:
"At our agency, we've done a lot of iterating on the best process for using AI to produce blog posts and web content for clients. Using GPT-4 through OpenAI's API, we've found that it works best to use the tool to build content sequentially—first the outline, then the intro, then the full article—and give the human writer the ability to course correct at each step."
The process for creating content with AI will differ for every company and content marketer. But it's essential that your final outputs read and sound like human-written content. Because even though AI content tools are developing fast, they still lack the emotional intelligence, creativity, and understanding of nuance that make content shine.
4. Fast-track content research
Whether you're writing technical papers or gathering information for industry reports, scouring the internet to find information is hectic. AI tools are great at this: they can collect information, analyze it, and deliver consolidated data in minutes.
Christian, the content lead at AquaSwitch, told me:
"Our company writes interesting takes on energy, sustainability, and water. However, to find the data we need for content, we have to sort through long reports from large organizations, which can take several hours. But since using GPT-4, we now get the bulk narrative and relevant data in minutes. And with the right prompts, GPT-4 can analyze and organize this data into easily digestible formats like this."
Content agency founder Rebekah Edwards agrees that AI streamlines the research process. After experimenting with GPT-4 for a new site, she told me the tool significantly cuts down her content creation time:
"As a freelance health writer, I could look for scientific research and write an article in ~4-6 hours. That content still needs to be edited by another person and uploaded to a website. But with ChatGPT, I've been able to write, edit, upload, and internally link to my new page in between 1.75-2.5 hours (max)."
While AI tools can do the heavy lifting for you, be careful not to pass off the entire content research process to them. AI is known to hallucinate, so while the data it returns might look accurate, it's very possibly made up. Especially when you're using it for data purposes, you always want to confirm at the source. Christian told me, "We collect 90% of our facts and data from GPT-4, but we don't use it to find sources. After it gathers the data, we look on Google for the sources as most of the links the tool returns don't exist or are out of date."
5. Do keyword research and clustering
Usually, keyword research starts with brainstorming a list of phrases relevant to your niche or topic. It's largely a manual and time-consuming process. AI marketing tools like ChatGPT can expedite the process by generating a large list of relevant keywords and their variations in seconds. It can then categorize these keywords into relevant topic clusters for your content strategy plan.
Ben Poulton, founder of Intellar, has been using GPT-4 to research keywords and cluster them into topics that form the basis of his SEO content targets.
Here's what his process looks like:
"I get ChatGPT to give me a list of keywords on a topic. I sometimes also feed in keywords I research using SEO tools such as Semrush, or manually on Google.
With this extensive list, I ask ChatGPT to cluster these keywords into relevant topics.
To bulk it out even more, I ask ChatGPT to list a bunch of subtopics for each main topic it's given me. Then, I repeat step 1."
He explained further:
"This process of keyword research and clustering saves me both time and money. Time because clustering keywords manually is a tedious task that can take hours. And money because many other tools offering keyword clustering processes have these gated by a higher subscription level or by credit usage. For a typical keyword research process involving 100-200 keywords, my AI process would save me 1-2 hours of work and the money/credits I'd otherwise have to use in other SEO tools."
And Ben's certainly not alone. In their 2023 content marketing and generative AI outlook report, Beantown Media Ventures found keyword identification to be the most common use case for AI among marketers in B2B and B2C brands.
Some AI tools will do this better than others. For example, ChatGPT doesn't have access to the SERPs and can't check for keyword overlap or cannibalization. Bard, on the other hand, is connected to Google Keyword Planner. And then there are tools built specifically for this purpose, like Surfer, which will give you much more granular insights.
6. Streamline the creative review workflow
Ziflow, the creative workflow platform powered by ChatGPT, provides suggestions on the best ways to improve creative ideas, convey your message, and approach creative problems. Ultimately, this helps remove friction from the feedback process.
Liza Kirsh, chief marketing officer at Dymapak, can testify to this. "Ziflow allows our team to never skip a step during the creative process," she said. "After creating content, each team member can provide feedback instantly, and activating ZiflowAI makes the process easier."
She continued:
"Using ZiflowAI in our workflow eliminates a lot of the back-and-forth teams are used to. It completely streamlines the process, so feedback is easy to understand and implement. With ZiflowAI, our creative and marketing teams have reduced project delivery time by 50%."
If you use Ziflow for creative review, you can also connect it to all the other apps your team uses with Ziflow's Zapier integration.
Zapier is a no-code automation tool that lets you connect your apps into automated workflows, so that every person and every business can move forward at growth speed. Learn more about how it works.
7. Edit and proofread your drafts
AI content editing tools like Grammarly and Notion AI serve as valuable assistants. They'll help you fix grammar mistakes, typos, and incoherent phrasing. But more than that, they can help ensure your content follows the appropriate brand tone and style.
On Grammarly, for example, you can set goals that reflect your brand's style guide, so the tool can offer tailored editing suggestions.
Other tools do the same thing. Jasper has a brand voice feature, where you can teach it what your brand sounds like using examples, and Notion AI allows you to change the tone of anything you write.
While these tools streamline the editing process, they're not going to replace human editors. Some of their suggestions can be out of place or even change the meaning of the content, and they don't have emotional intelligence because…they're robots. So think of them more as editing assistants, and use your judgment as you follow their suggestions.
8. Check for plagiarism and AI content
Generative AI tools like ChatGPT generate content based on the original datasets they're trained with. So, as you use them for content creation, keep in mind that some of the outputs may be uncomfortably close to content that already exists. Translation: it's plagiarized.
In good news, you can now use AI to check if your AI-generated content is plagiarizing other web pages. (What a world.)
Grammarly has a plagiarism checker that compares your content to billions of web pages and databases. If it detects duplicate content in your text, it'll highlight them for your review.
Other tools, like Copyleaks, can catch plagiarized content and AI-generated content. Even if you wrote the content yourself, if an AI detector tells you it sounds like a robot, that's not great.
Use AI content tools for support, not replacement
AI tools can support every step of the content creation process, but remember, they're assistants, not replacements. I'll pass the mic back to Megan:
"The most difficult part of working with AI is its inconsistency. A lot of what it produces will be perfectly fine, but in other places, it will say completely false things—and there's no way to know where those problems are without having a human review it. So please, for the sake of everyone on the internet, don't publish pure AI content without human oversight."
Related reading: