Scaling content with AI sounds like the perfect solution until you realize one-click content won’t cut it if you want to rank. 

Lean teams trying to produce more with fewer resources quickly discover that AI drafts lack depth, miss audience needs, and fall short of search quality standards.

I believe there’s a better way to use AI to streamline the entire content process, from ideation to publishing, without sacrificing quality or control.

In this post, I’ll show you how I use AI to scale content marketing tasks like ideation, briefing, and optimization. But first, take some time to read my previous article, as it lays the foundation for this piece.

LLMs can turn audience data into actionable content ideas. Here’s how I approach content ideation with AI.

Feed user personas into ChatGPT

In the previous article, I mentioned leveraging the UserPersona Generator for audience insights. Start with detailed user personas based on your site’s Google Analytics (GA) data. Then, use the User Persona Generator to create personas that reflect audience behavior, interests, and needs. 

For Pooch & Mutt, this revealed segments like health-conscious dog owners and those interested in sustainable pet products​.

Use a prompt like:

Based on this user persona, suggest 10 blog topics that address their specific needs and interests    

Identify trending topics

Use Reddit, Google Discover, and competitor content to surface popular themes. The easiest way to scrape competitor content and keywords is to use the Domain Overview feature in Moz Pro to identify themes for the website. For example, these are popular themes for Tails.com, a competitor of our example site, Pooch & Mutt.

You can take it a step further by using the Ranking Keywords By Site Feature in Moz Keyword Explorer to scrape competitor pages and keywords they rank for.

However, traditional tools are not the only way to get competitor data. You can do some research on Reddit with Google Gemini using the prompt below:

"What are the top discussion topics in dog nutrition on Reddit this month?"   

It flagged discussions around hypoallergenic diets and homemade dog treats. I advise you to cross-reference results, as Gemini can hallucinate data​.

Run content gap analysis

With personas and trending topics in hand, identify gaps in your content. Use your content audit sheet to find underrepresented themes. For Pooch & Mutt, I found a lack of content around seasonal dog care despite being a popular topic in competitor blogs.

You can also use the Keyword Gap feature in Moz Pro to run a content gap analysis. First, navigate to Competitive Research in Moz Pro and select Keyword Gap. Next, enter your website URL alongside competitors and analyze it based on your chosen country.

You’ll find content gaps with opportunities to improve or create new content.

Export and add to your list of content ideas.

Pro tip: 

Avoid using ChatGPT for keyword research since the data will not be accurate. Instead, use Search Console and third-party keyword research tools like Moz Keyword Explorer to validate keyword opportunities​.

Once you have content ideas, create content briefs to help writers stay on track and deliver high-quality content. LLMs can automate this process while allowing customization based on your brand’s needs. 

I created this workflow using Monzo (an online UK bank) as an example.

Create a personalized content brief GPT

Start by setting up a GPT specifically for content brief creation. Define the structure you want the brief to follow, including:

• Primary keyword/topic 
• Secondary keywords (optional) 
• Content type
• Target country
• Target audience
• Content purpose/context
• Products/features /services discussing (optional)

In the above example, I chose "How to Save for a Holiday" as Monzo’s blog topic because it is relevant to Monzo's audience. 

For the content brief, the GPT output included conversion-focused CTAs, such as signing up for the Monzo newsletter and using the Monzo savings pot​.

Add contextual information

The brief improves when you feed the GPT additional context. For Monzo, this included:

• Existing content and audience insight
• Information about the Monzo savings pot
• Blog posts that matched the desired tone and format
• Content exclusions, such as holidays for people already traveling or those in debt

Adding contextual layer helped the GPT generate more relevant and actionable recommendations​. Here’s a sample instruction to help you create a content brief GPT.

Export the brief

Once generated, the brief included everything the writer needed to produce quality content, including:

• Title tag optimized for SEO
• H1 and URL aligned with the topic
• Meta description designed to drive clicks
• Relevant terms the writer should incorporate
• Subheadings and structure  with clear instructions
• Recommended media such as screenshots, infographics, and user testimonials

In the Monzo example, the brief even recommended adding screenshots of the savings pot and user testimonials from customers who had successfully saved for holidays​.

Pro tip:

Always amend the content brief as needed. AI can automate the process, but human oversight ensures the brief maintains brand standards and aligns with your goals.

Once your content draft is ready, use ChatGPT to edit it and ensure it aligns with brand guidelines.

Here’s how I approached AI editing with Monzo as an example.

Create a content editing GPT

Set up a dedicated GPT for editing tasks, separate from your briefing GPT, to avoid cross-task confusion.

Here are the two GPTs I made for the following tasks:

• Monzo Content Briefer
• Monzo Editing GPT

 I created three tasks within the editing GPT:

• Align content with brand tone of voice
• Compare content against the brief to flag any missing elements
• Sub-editing for clarity, flow, and accuracy​

Run brand tone of voice checks

To maintain Monzo’s brand voice, I uploaded their tone of voice guide into the editing GPT’s knowledge base. Next, I pasted my draft into the editing GPT and ran the “Content Review According to Tone of Voice” task.

The GPT flagged overly complex phrases and suggested simpler alternatives. It also referenced Monzo’s guidelines, such as using “helpful, plain English” and avoiding jargon​.

Compare content against the brief

I uploaded the draft and the original content brief, clearly labeling each in my prompt. The GPT compared them section by section, highlighting where the draft met or missed the brief.

For example, the brief for Monzo’s How to Save for a Holiday post included Monzo’s savings pot as a solution. However, my original draft mentioned savings tips but omitted Monzo’s product. The GPT caught this and recommended the product, ensuring alignment with the content goals​.

Sub-edit for clarity and accuracy

Finally, I ran a full sub-edit within the editing GPT. The GPT flagged factual errors, awkward phrasing, and grammar issues. 

I tested it with exaggerated claims like “Monzo charges 30,000% interest,” and it immediately flagged them as incorrect. It also caught stylistic issues, like excessive exclamation marks and inconsistent formatting.

Before you hit publish, ensure your author bios reflect their expertise. Use LLMs to generate polished author profiles across platforms. 

Here’s how I do it with Copilot:

Generate a page summary

Start by opening Copilot in Microsoft Edge. If you’re working with a LinkedIn page, personal website, or any other author page, enter the CoPilot interface at the top right of the Edge browser and click Generate a Page Summary. 

This step is essential; without it, Copilot won’t scan the page properly or provide an accurate summary​.

Review skills, recommendations, and certifications

Copilot will highlight key details, including:

• Professional skills
• Endorsements and recommendations
• Certifications and memberships
• Interests and affiliations

It generates a snapshot of your professional background, making it easier to build a concise and authoritative author bio​.

Create the author bio with a prompt

Next, ask Copilot to generate an author bio using the following prompt:

"Use this page’s information and write a formal author bio. This should include the exact name, company credentials (make sure to include exact current job title for current role), education, experience, and if present, any awards and affiliations. Only if you see them on the page, you can also link to other social media profiles, portfolio, or other relevant websites that showcase their work and reputation. Do not mention their LinkedIn profile or social media in your output. Do not make up information not present on the page."

I used this approach to create my author bios, requiring minimal adjustments​.

Tailor for specific roles and platforms

Customize the bio further by specifying the role or focus area you want to highlight.

I emphasize technical expertise if it’s a bio for an SEO website. For a brand blog, I tweak it to highlight my content strategy and audience engagement skills.

Pro tip: 

Keep bios concise. A word count between 40 to 70 words works best for blog posts. Create separate bios tailored to each channel if you publish across multiple platforms.

Once your content is live, use LLMs to repurpose it to extend reach and value across platforms. 

Here’s one way I approach content repurposing using GPTs and VEED.

Create a repurposing GPT

I built an Ultimate Content Repurposer GPT to simplify repurposing across channels. 

Add your content URL or paste the draft into the GPT alongside your target audience details. The GPT analyzes the content and extracts three key themes relevant to that audience. It then outputs content ideas tailored for different platforms.

For example, I repurposed an article about integrating UX into marketing. The GPT generated LinkedIn posts, Twitter threads, and email newsletter content, each adapted for platform-specific engagement​.

Turn written content into video

I used VEED's Video GPT to repurpose blog posts into short-form videos. I pasted the article line, and the GPT extracted key takeaways. 

It then created a brief for a TikTok video titled "Bite-Sized UX Marketing Tips," pulling quotes from expert opinions in the original article. 

After confirming the brief, VEED generated the video project.

I recommend editing the output before publishing, as automated video generation can include minor errors. Also, if you're using a free plan​, be aware of watermarks.

When your content is live, use LLMs to analyze performance and find areas for improvement. Integrate tools like Microsoft Clarity and custom GPTs to track performance and ensure your content meets E-E-A-T standards. 

Here’s how I approach it:

Review user engagement with Microsoft Clarity heatmaps

I love using Microsoft Clarity for real-time heatmaps that show how users interact with key pages. It helps me understand how users engage with CTAs, site navigation, and content depth.

My goal is to increase conversion rates, so I prioritize money pages, product pages, and high-traffic blog posts. For example, a Clarity report revealed users were dropping off before reaching a key CTA on a product page. I used this insight to move the CTA closer to the top, which improved conversions​.

Evaluate content quality with the Search Quality Rater Evaluator GPT

I used the Search Quality Rater Evaluator GPT to ensure content meets Google’s E-E-A-T (Experience, Expertise, Authoritativeness, and Trustworthiness) standards.

Paste your content URL or full text into the GPT. It evaluates the page using Google’s rater guidelines and provides feedback across six categories:

• Expertise: Check if the content is written by knowledgeable authors with subject-specific experience.
• Authoritativeness: Assess whether the site or author is recognized as a trusted source in the industry.
• Trustworthiness: Verify if the content is transparent, factually accurate, and free from misleading claims.
• Content quality: Ensure the content is detailed, up-to-date, and provides value to readers.
• User experience: Evaluate site navigation, layout, and readability for a smooth user journey.

Optimize content for ‘Experience’

If you want to review a page for content experience, use my EXP Boost GPT to get feedback on the following:

• Text analysis: Check if the content demonstrates first-hand expertise.
• Image analysis: Evaluate whether visuals support the content and reflect real-world usage.
• Authorship analysis: Assess whether the author’s credentials are clear and authoritative.

Analyze review-style content

I created the High-Quality Review Analyzer GPT to ensure your content meets Google’s review guidelines. Paste your content URL or text into the tool, and it will analyze the post, highlighting areas for improvement.

I used it to compare two review articles. The GPT identified which one provided more trustworthy information based on criteria like first-hand experience, expert quotes, and transparency. It also suggested improvements, like clarifying product specifications and adding comparison tables​.

Pro tip: Always prioritize optimization efforts on your most important pages, such as product pages, feature pages, and high-conversion landing pages.

You now have enough LLM workflows and GPTs to build a robust content engine from research to execution. However, automation doesn’t equate success.

You need a validation system and quality control process to ensure content adheres to brand guidelines while providing immense value to your audience.

I encourage you to read Aleyda Solis’s article, in which she shares some validation workflows to ensure quality when using LLMs for SEO and content tasks.