quarta-feira, 13 de setembro de 2023

How to Make AI Your Writing Sidekick for Content Marketing

Artificial Intelligence (AI) has blown up like a supernova. Conversations about AI are happening everywhere. As you browse LinkedIn, read your favorite SEO blog, or tune into a marketing podcast, you will see discussions about AI.

Some marketers see AI as a threat. But that doesn’t have to be the case. You can learn how to work with AI. Building AI into your processes will future-proof your marketing skills if AI takes over the world. As a freelance copywriter and content marketer, I’ve seen ‘I, Robot,’ and I’m not taking any chances.

AI can be your marketing sidekick. You can use AI to support your writing process, be a fresh pair of eyes for proofreading, or act as a second brain. This makes it a powerful companion for content marketing.

AI won’t replace content marketers (yet)

AI opens up a universe of opportunities. As new AI tools emerge every day, so do new ways of using AI in your business and everyday life. You can use AI to do almost anything.

But that doesn’t mean AI will replace content marketers, SEOs, or writers — or, at least, won’t replace them yet.

AI has a lot of potential. But it also has limitations. Before using AI for content marketing, it’s helpful to get to grips with the constraints of AI.

It isn’t sentient

AI doesn’t have a brain like you or I do. It can’t think or feel.

AI’s knowledge is based solely on logic. In contrast, our knowledge is based on a mix of reason and emotion. Our experiences and feelings shape how we see things. But AI doesn’t have that ability.

AI might feign human emotion, but it’s merely an illusion. AI’s emotion is curated. AI can only perceive things based on the information it is fed.

You can’t rely on it to be accurate

Sure, AI is powerful. But it isn’t always accurate. Just like humans, AI sometimes gets things wrong.

When using ChatGPT, you will see a disclaimer stating, “ChatGPT may produce inaccurate information about people, places, or facts.” While AI tools do their best to provide factually accurate information, they still need a human touch.

If you’re using AI, fact-check the information it gives you. This is particularly important when writing YMYL (Your Money or Your Life) content that has real-world effects on the readers. In these cases, a human touch is crucial for ensuring E-E-A-T (Experience, Expertise, Authoritativeness, and Trustworthiness) runs through your content. AI cannot replicate this, running the risk of your content being inaccurate and low quality.

The output is limited by the input

Getting the best possible results from AI depends on feeding it the best possible information.

It’s like asking someone to design you a logo without giving them the exact details about what you want the logo to look like. They might get it right. But it is more likely that they will get it wrong — and you will end up in a back-and-forth ping-pong match of reiterating the design.

You need to tell AI exactly what you want. Otherwise, it’s going to hazard a guess (and likely get it wrong). Be specific in your request — like really specific. Want AI to give you ten content ideas for a blog post on programmatic SEO? Tell it! Want to make sure those results are all about the SaaS industry? Tell it! The more specific your generative AI prompt is, the better the output will be.

The responses can feel canned

You can get AI to write your landing pages, blog articles, and ad copy. But you might not want to.

Remember how I said AI isn’t sentient? And that the output is limited by the input? Well, both of those factors also contribute to the responses feeling somewhat meh.

AI responses can feel canned. They lack the originality and personality you get from a human writer who knows your brand inside and out.

With that said, you can train some AI tools, such as ChatGPT, to produce content that mirrors your brand’s voice, personality, and tone. You can do this by sharing examples of your writing style, providing background information on your company, and feeding it your tone of voice guidelines. But I’ll share more on that later in the article.

Another attribute that limits the creativity of AI outputs is that it doesn’t automatically consider the user journey stage, their current motivations or desires, or the exact words that drive action for them. AI doesn’t know your deeper brand origins or future goals — unless you train it on them.

All that adds up to creating generic content anyone could write. Use AI enough times, and you will notice familiar patterns in the responses. It sounds like AI.

But AI can be a powerful content marketing sidekick

It pays to be aware of the shortfalls of AI before you add it to your content marketing process. As we’ve established, AI isn’t perfect. But it can still be useful.

You can use AI to optimize the research, writing, and editing phases of content marketing.

You might not be able to rely on AI alone. But you can pair it with a human touch to turn it into a powerful content marketing sidekick.

Here are some ways you can use AI to optimize and streamline various aspects of your content marketing process.

How to use AI in the research phase of content marketing

The research phase of a content marketing strategy can be time-consuming. While every company’s strategy is different, the core elements will follow a similar pattern.

At some point during on-site content planning, you will need to find content ideas, conduct keyword research, and create outlines for blog posts. There are a few ways you can use AI to simplify these content marketing research and planning steps.

Use AI to find content ideas

In its most basic form, the content you create for your brand should focus on two things:

  • What you do

  • Who you do it for

We can get more nuanced than that by layering in factors like goals, campaigns, and user intent. But most content will cover topics that are related to your business (what you do) and of interest to your audience (who you do it for).

You can use AI to develop content ideas for a wide variety of marketing channels — be it social media, newsletters, blogs, or videos. For the examples shared throughout this article, I’ve used ChatGPT.

A simple AI prompt you can use to find content ideas is:

Write a list of 10 [content type] ideas aimed at [audience] who are [problem they are trying to solve/goal they want to achieve]

In this prompt, the content type should relate to the channel you are creating content for, such as a blog. The audience will align with your target audience. The problem or goal should relate to what your brand does.

For a dog training school, the prompt might look like this:

Write a list of 10 blog post ideas aimed at new dog owners who are trying to train their puppy.

Screenshot of a ChatGPT conversation. The inputted prompt reads 'Write a list of 10 blog post ideas aimed at new dog owners who are trying to train their puppy'. ChatGPT's response shares a numerical list of 10 blog post ideas including '10 Essential Commands Every Puppy Should Know' and 'House Training 101: Tips for Potty Training Your Puppy’

If you put this prompt into Chat-GPT, you will be presented with a list of potential blog post ideas such as “puppy separation anxiety: tips and tricks” or “10 essential commands every puppy should know.”

Work through this list to decide if any of those ideas are viable blog posts. Some might be perfect as they are, while others might need fine-tuning to make sure they align with your content marketing strategy.

Create a hub-and-spoke strategy from those content ideas

Drill down further into each idea to create the bones of a hub-and-spoke strategy. For each idea presented, note what the topic is.

For the blog post idea "10 essential commands every puppy should know,” you can confidently assume the focal topic is “puppy training commands.”

There is likely an array of blog posts you can write about puppy training commands, making this a potentially powerful hub topic for your hub-and-spoke strategy. Rather than doing lots of manual research to find suitable spoke blog post ideas, you can use AI to speed up the process.

Use the following AI generative prompt to ask Chat-GPT for spoke content ideas based on the hub topic:

I am writing a series of [content type] for a [business type] aimed at [target audience].

For the hub topic “[hub topic],” create a table of spoke content ideas.

Applying this to the puppy training commands keyword, the prompt would look like this:

I am writing a series of blog posts for a puppy training school aimed at new dog owners.

For the hub topic "puppy training commands," create a table of spoke content ideas.

Screenshot of a ChatGPT conversation. The input reads 'I am writing a series of blog posts for a puppy training school aimed at new dog owners. For the hub topic 'puppy training commands', create a table of spoke content ideas.ChatGPT's response features a two-column table with the headers 'Blog Post Title' and 'Content Description'. Under these headings, you can see rows of blog post ideas with accompanying descriptions such as 'The Basics: Essential Puppy Commands' with the description 'Introduce fundamental commands like sit, stay, and come. Explain their importance and how to teach them effectively. Include troubleshooting tips.’


In response to this prompt, Chat-GPT generates a table of more specific blog post ideas related to the hub topic “puppy training commands.” This will help you spin out a series of valuable content that strengthens your content visibility for this topic.

Use AI for keyword research

Love it or hate it — keyword research is essential to developing on-site content strategies. It tells you what topics people care about, how popular those terms are, and how difficult it might be to rank in SERPs.

Generic AI tools can’t do in-depth keyword research, so you should still use a keyword research tool like Moz’s Keyword Explorer. However AI can help with the early stages of keyword research.

If you’re starting keyword research from scratch, use AI to generate a list of keywords. Take this a step further by asking the AI tool to group those keywords by search intent. This prompt will look something like this:

Create a list of keywords for [business type/topic]. Present the keywords in a table grouped by search intent.

Sticking with our dog training school example, the prompt would be:

Create a list of keywords for a dog training school. Present the keywords in a table grouped by search intent.

Screenshot of ChatGPT conversation. The input reads 'Create a list of keywords for a dog training school. Present the keywords in a table grouped by search intent.' The ChatGPT response features a three column table split by 'Search Intent: Informational', 'Search Intent: Navigational', and 'Search Intent: Transactional'. Under each row, there are numerous keyword ideas.

You’ll then be presented with a neat table of keywords you might be able to use. The keywords likely won’t be perfect, but it’s a great starting point.

Go through the list and research each keyword in Moz’s keyword research tool to get more data, including metrics such as Difficulty and Monthly Volume. You can then use your keyword research to see which keywords you do (or don’t) want to use and find other viable keywords.

Using AI for keyword research is great for getting the cogs turning, taking you from a blank page to a list of keywords. Grouping by search intent lets you take that rough keyword research one step further, making sure you’re covering various user journey stages.

Use AI to write blog post outlines

Okay, by this point, you should have a bunch of target keywords and blog post ideas. You’re almost ready to start putting pen to paper. But before you write your blog post, you need to plan a first draft.

You can use AI to help you write the first draft of blog post outlines. AI-generated first drafts won’t be groundbreaking. You will need to cast your eyes over it and optimize it with your amazing brand, customer, and market insights and knowledge.

While they aren’t perfect, they give you a good blog post template you can then tweak and improve.

Input this AI prompt to generate blog post outlines:

Write a blog post outline for a blog post titled "[blog post title].”

The main keyword for this blog post is "[main keyword]"

You guessed it, we’re sticking with the puppy training example! Based on the puppy training research we’ve done so far, the AI prompt might look like this:

Write a blog post outline for a blog post titled "10 essential commands every puppy should know.”

The main keyword for this blog post is "puppy commands"

Screenshot of ChatGPT conversation. The input reads: ‘Write a blog post outline for a blog post titled ‘10 essential commands every puppy should know.’ The main keyword for this blog post is ‘puppy commands’ ChatGPT's response provides a blog post outline split into sections


Chat-GPT will then give you a blog post outline split into sections, including the introduction, main content sections, and a conclusion.

As this post focuses on the “10 essential commands,” Chat-GPT has given a listicle-style blog post outline. It has also shared a list of additional resources at the end you might want to reference in the post.

You can then review the outline that Chat-GPT has provided and apply your topic knowledge, adapt based on your experiences, and carry out additional research to improve the outline before you start writing the article.

How to use AI in the writing phase of content marketing

The sidekick powers of AI don't stop when you start writing. When you’re writing content, the process will be predominantly human-led. Your tone of voice matters. As do your personal experiences, opinions, and insights. Those are the things that make your content unique and that AI can’t do very well. So don’t replace your uniqueness with generic AI content.

You can, however, use AI as a sub-writer to assist with the writing phase of content marketing projects.

Use AI to adapt your tone of voice

If you’re writing content and want to make sure it has a particular voice, you can use AI to tweak the tone of your writing. This can help make sure your content aligns with your brand voice and the page intent.

If you’re writing content for a sales landing page, you will want that content to sound more persuasive than an educational blog post. AI can help you adapt your content to make it better suit its intended purpose.

Use AI to adapt your content’s tone of voice by using generative AI prompts such as:

  • Make the following text sound more [adjective]

  • Improve the following text by using a [tone of voice type] tone of voice

For a sales landing page, these prompts might look like this:

  • Make the following text sound more persuasive and encouraging

  • Improve the following text by using a persuasive and confident tone of voice

Screenshot of a ChatGPT conversation showing an exchange between the user and ChatGPT to improve the tone of voice of the inputted text

Remember, the output from AI depends on the input. You can improve these prompts by providing more details. You could, for example, ask AI to use metaphors or similes in its version or to use simple vocabulary and short, snappy sentences.

You could even tell it your brand values to make sure those are taken into consideration for tone of voice content.

AI doesn’t always get the tone of voice right. Often, it will exaggerate the voice. Whatever AI provides, you will need to edit manually. But it can be great for finding inspiration and unraveling the thread for words you could use or changes you could make.

I never copy and paste the output given by AI. Instead, I’ll look for words, sentences, or content styles that stand out to me. I’ll then refine my content manually based on any learnings gleaned from the AI version.

Starting the paragraph with the word “Enroll,” for example, encourages action and lets the reader know the intention of the content. This is far more persuasive than the softer, descriptive opening line “In these classes…” used in the original content.

This feature is far from perfect, but it can be useful if you’re experiencing a classic case of writer’s block.

Use AI to generate headlines

The headline is one of the most crucial elements of a blog post. A compelling, well-written headline gets clicks. It drives traffic to your blog post and gets more eyes on your content.

Yet, headlines can be the hardest thing to write. Do a quick Google search, and you will find hundreds of formulas for writing compelling headlines or tools to improve your existing headlines.

When writing headlines, you should account for your target keywords and the intent of the content. If you’re writing an educational blog post that teaches readers how to do something, you will likely use phrases like “how to.”

The only issue is that having an endless list of “how to” articles on your blog starts to look boring and repetitive. You can mix this up by asking AI to generate potential headlines for your blog articles.

Use the following prompt to get AI to generate headlines for your blog posts:

Write a list of 10 potential headlines for a blog post about [blog post topic]. This blog post [brief description of blog post].

Using our puppy training school example, this prompt might look something like this:

Write a list of 10 potential headlines for a blog post about puppy commands. This blog post shares 10 essential commands any new puppy owner should know when training their puppy.

Screenshot of ChatGPT conversations. The user wrote the following input: Write a list of 10 potential headlines for a blog post about puppy commands. This blog post shares 10 essential commands any new puppy owner should know when training their puppy. ChatGPT responded with a numerical list of 10 headlines that could be used for this blog post.

As this example is a listicle, it makes sense that Chat-GPT includes the number in the headline. However, you can see a variety of headlines you could potentially use that sound more compelling than just “10 essential commands every puppy should know.”

The headlines also use power words like “must-know” and “ultimate” to entice readers to click through to the article. Meanwhile, words and phrases like “learn,” “teaching,” and “training 101” let the reader know the article has an educational intent.

These outputs offer a great first draft of some potential headlines. You can then fine-tune these headlines using well-known headline best practices.

This prompt doesn’t have to just be used for blog articles. You can adapt it for YouTube video headlines or email subject lines.

You can also use AI to generate headline formulas that you can use when writing headlines in the future. Do this by using a prompt such as:

Write 10 headline formulas that can be used to write blog post headlines. Display these as a table and make sure they cover a variety of content types such as [types of blog posts you typically share]

Screenshot of ChatGPT conversation. The user inputted Write 10 headline formulas that can be used to write blog post headlines. Display these as a table and make sure they cover a variety of content types such as [types of blog posts you typically share] ChatGPT's response features a three column table with headers for 'No.', 'Headline Formula', and 'Content Type'. In the table, it provides headline formulas. At the end of the table, ChatGPt says to 'Feel free to mix and match these formulas to create compelling and diverse headlines for your blog posts!'


As always, you can drill down further into this by asking AI to prepare several headlines for each article type. You can then tweak the formulas and keep these on hand next time you’re writing an article.

Use AI to summarize your writing

Every blog post should end with a conclusion that gives the TL;DR (too long; didn’t read) low down of the article.

Typically, your conclusion should summarize the article by sharing the key takeaways. It should succinctly wrap up the article, leaving a lasting impression on the reader.

You might also want to use the conclusion to encourage readers to take action. This might include prompting them to read related articles, encouraging them to get in touch, or directing them to additional resources.

If you’re using ChatGPT-4, you can input longer content. This makes it great for asking ChatGPT to summarize long-form blog articles. You can also use AI tools that have specifically designed conclusion generation features, such as WriteMe or LongShot.

Use the following AI prompt to get AI to summarize your content:

Summarize the following content and provide the key takeaways. This will be used to write the conclusion for the blog post about [brief description of the blog post]

For our puppy training school example, this prompt will look like this:

Summarize the following content and provide the key takeaways. This will be used to write the conclusion for the blog post about the 10 essential commands any new puppy owner should know when training their puppy.

Screenshot of a ChatGPT conversation. The user inputs 'Summarize the following content and provide the key takeaways. This will be used to write the conclusion for the blog post about the 10 essential commands any new puppy owner should know when training their puppy.' followed by a block of text about puppy commands. ChatGPT responds with a condensed version of the copy

In this example, I only shared a snippet of a blog post to give you an idea of how this might look.

You can even take this prompt a step further by asking Chat-GPT to provide the key takeaways as a bullet-pointed list.

Screenshot of ChatGPT conversation. The user writes ‘Thanks. Can you share those key takeaways as a bullet pointed list?’ChatGPT responds with a bullet pointed list, as requested

This quickly transforms a long-form article into a scannable list of key takeaways. You can then use this output to get a quick overview of the blog article so you can easily write your conclusion.

Use AI to write meta-descriptions

While they might only be short, writing high-quality meta descriptions requires skill. The right meta description can be the deciding factor in whether someone clicks through to your website or not.

Luckily, you can use AI to distill your on-page content into a couple of short sentences you can use as a meta description.

You can get ChatGPT to write meta descriptions by using the following AI prompt:

Summarize the following content into a 2 sentence meta-description

The optimal length for a meta description is between 50 to 160 characters. Run the AI-generated output through a word count checker to ensure it’s a suitable length. If not, you can either ask AI to generate a shorter version, or you can edit it manually.

Screenshot of a ChatGPT conversation. The user inputs ‘Summarize the following content into a 2 sentence meta description.’ followed by a block of text about puppy commands.ChatGPT responds with a condensed version of the copy that is two sentences long.The user then asks ‘Can you make it slightly shorter?’ChatGPT responds with a shorter variation of it's original response that is only 145 characters long.

In the above example, ChatGPT generated a meta description that was 195 characters long. As this is slightly longer than desired, I inputted the prompt, “can you make it slightly shorter?”

In response, ChatGPT cut the original output down to 145 characters while maintaining the key takeaways of the article.

How to use AI in the editing phase of content marketing

Before any content sees the light of day, it needs to go through a couple of rounds of revision. These revisions will either be conducted by the original writer or by editors, proofreaders, or content managers.

But reviewing your own content is hard. You’re too close to it. You know it inside out. I like to give myself one day between writing my first draft and proofreading it. This makes sure I’m looking at my content with fresh eyes.

But you don’t have to edit or proofread your content alone. Your AI content writing sidekick can come in useful during the editing phase of content marketing.

Use AI to clarify your writing

If you’re writing about complex topics, your content can easily become overwhelming for the reader. It might be laden with jargon or use niche analogies that only a small subset of people understand. The trouble with this is that complex content runs the risk of alienating your reader.

On the flip side, being too creative with your writing can also have a negative impact. Perhaps you’ve tried to write content that’s witty or overflowing with metaphors to create a vivid emotional reaction. Except it doesn’t quite land how you expected. The meaning and purpose of the content are lost in creativity.

Both scenarios are examples of how being “clever” can negatively affect the impact of your content. As writers, we live by the rule of being “clear over clever.” Yes, being clever is great. It turns heads, captures attention, and shows people you know what you’re talking about. But it should never come at the expense of your writing being clear and easy to understand.

If you’ve written something that you think might be clever rather than clear, you can use AI to clarify your writing.

AI prompts such as “Does this text make sense?” or “Make this text easier to understand” let you check and improve the clarity of your writing.

Screenshot of a ChatGPT conversation. The user writes ‘Does this text make sense?’ followed by a block of text about teaching your dog recall commands. ChatGPT responds to confirm the text makes sense and suggests it ‘could be slightly refined for clarity and flow’. ChatGPT then gives a revised version that could be used instead

You can use the AI output to see whether it has understood your content or if the original meaning has been lost. In its response, Chat-GPT will tell you what it interpreted the text as, and it might also provide suggestions for improving your copy. You can then use this information to rewrite your content so it is easier to understand.

Use AI to simplify your writing

I have the habit of being a bit of a waffler when I’m reading. My default writing style is to use all the words. So, when proofreading my content, I sweep my content for long sentences and large blocks of text. I will then highlight these and simplify them.

To speed up this process, you can use AI to simplify content so it’s shorter and more readable.

Some AI prompts you can use to simplify your content are:

  • Make the following text shorter and snappier

  • Explain [topic] to me like I’m 7 years old

  • Make the following text easier to read

Screenshot of ChatGPT conversation.The user inputs: ‘Make the following text shorter and snappier’ followed by a block of text about common dog responses to the doorbell ringing and the importance of teaching them the place command.ChatGPT responds with a shorter alternative that can be used instead

If we input a large block of text, you can see how the prompt “make the following text shorter and snappier” produces a simplified version of the original content. You can then comb through your content, making it short and more readable.

Final thoughts — AI can be a powerful content marketing sidekick

AI isn’t perfect, but it can be an invaluable asset for content marketers, acting as a powerful sidekick at every stage of the content creation process.

Guided by well-crafted generative AI prompts, you can optimize your content. From kickstarting content research with keyword research to creating hub-and-spoke content strategies, generating headlines, and clarifying your writing, AI can streamline your content marketing efforts across content research, writing, and editing.

Remember, you can’t rely on AI alone just yet. Your content still needs a human touch if you want it to outperform your competitors — especially when writing for YMYL topics and niches where E-E-A-T is crucial.

For the best results, pair AI’s capabilities with human creativity, experience, and expertise so you can stay ahead in the ever-changing world of content marketing.

Nenhum comentário:

Postar um comentário

Observação: somente um membro deste blog pode postar um comentário.