8 SEO tasks you shouldn’t do manually

Take a look at your SEO to-do list. Most of it is probably the same rote, repetitive tasks. Turning everyday work into faster, easier outputs is easier than ever with AI.

Besides obvious tasks like note-taking and team reminders, you can automate tasks such as content audits, page outlines, and keyword research.

Start with simple strategies to save time on the repetitive work you do every day, then expand into using AI tools for automation. Always do a final check yourself, rather than trusting 100% of your work to LLMs, which rarely get things exactly right.

Identify automation opportunities

One simple way to decide what you can automate is to ask yourself: Would I assign this to an intern?

The types of tasks you’d give a new employee are ideal for automation. Let the intern (or AI) do 70% of the work, like research and a rough draft. Then complete the final 30% by giving feedback (or improving your prompts), and finalizing and publishing the draft.

Based on a typical SEO intern job description, these are some examples of tasks that could be automated:

• Analyze data and identify trends around traffic, engagement, and rank/visibility.
• Ensure SEO best practices are used when updating content.
• Create detailed reports for stakeholders about SEO performance, trends, and recommendations.
• Identify content gaps and duplicate content.
• Scale SEO-optimized templates across series or topics.
• Build an editorial calendar and share the strategy playbook.
• Document prompts, templates, and QA standards.

Other ideas to find automation opportunities:

• Audit your existing workflows and tasks.
• Review your onboarding and documentation.
• Ask your team which tasks they hate most, and why (sometimes the issue is different, but still fixable).
• Ask AI what it can handle.

Automation won’t fix these core challenges:

• Broken systems: You might miss some issues if you don’t know what to look for or how to fix them.
• Incomplete assets: You’re only as good as your data. If you don’t have all the tracking or performance numbers you need, it’s unlikely you’ll get complete results.
• Lack of resources: It’s great to run an audit, but do you need a ticket to make changes? How long will approval take?

8 opportunities for SEO automation

1. Content calendar

One of the first things I automated was the team content calendar, or at least the first draft.

I used UNIQUE, MAXIFS, IFERROR, and VLOOKUP formulas to paste different reports into a master sheet and generate a list of which pages were due for an update based on our update schedule. Some SEO experts suggest most content should be refreshed at least every 1-2 years, especially since LLMs favor freshness signals.

Add a performance audit with some VLOOKUPs, and put it all in a custom GPT to complete your draft. Then tweak your draft based on your goals and strategy.

Prompt example 

Based on the sitemap, performance report, and last quarter’s content plan, give me a table of the pages that are due for an update. Include columns for the URL, title, writer assigned, sessions, bounce, conversion rate, and notes.

Now add to the table the list of pages from the performance report that have dropped more than 30% in sessions or conversion rate. For any pages that are also in the previous update list, add the performance data to the notes column, but don’t include it as a separate row. Format it like this: Sessions -XX% L90D, conversion rate -XX% L90D.

Time saved: 8 hours per quarter.

2. Keyword and prompt research

While Ahrefs and Semrush content gap analysis reports are helpful, they can also include a ton of data to wade through. There’s often irrelevant info, such as branded keywords you don’t want to target.

AI can help generate keywords and prompts for you to track or target. To get LLM prompt ideas, start with some of your longest-tail keywords, which you can find in Google Search Console by exporting all keywords to a page and sorting by length. Then ask AI tools to refine your list and generate similar ones.

One watchout: AI tools often struggle to understand user intent, short- vs. long-tail queries, and when to target specific keywords tied to intent.

For example, it may suggest targeting the word “cats” on a local vet website homepage. While that word may appear often across pages, it’s not necessarily one you want to target. You’re unlikely to rank for the “cats” SERP or AIO for a small local site, but you could rank for “cat vet” or “cat care.”

Prompt example 

You’re an SEO analyst working on an update for this page. Using this Ahrefs report with competitor page analysis and keywords, identify the 20 most important keywords our page should target, ranked by MSV and relevance, and include those numbers with each keyword in a table. 

Don’t suggest targeting any branded keywords. Then give me a list of 10 ways you’d improve this page to better target the keyword, and why, with exact quotes from the copy you’d change.

Time saved: 15 minutes per page.

3. Internal linking

Internal links to your top pages are important for site crawling. The simplest way to automate internal linking is to export an Ahrefs backlink report that shows pages with few internal links, then input that with your sitemap into a GPT. Don’t select “Group Similar Links” so all URLs appear in the same report.

Ahrefs also has a feature for internal link ideas. It’s not great at suggesting pages with slightly different intent, but it can be a useful starting point for simple sites. Encourage your editorial team to use this workflow when working on drafts.

Prompt example 

Using this internal link report, sitemap link, and linking best practices documentation, suggest 5-10 pages I could link to. Only include pages with fewer than 10 internal links. Include only highly relevant pages. Don’t include any pages with “author” or “about” in the URL.

Time saved: 10 minutes per page.

4. Outlines and briefs

If your team works from outlines, briefs, or tickets, start by documenting a basic template for different tasks. Then set up a GPT to create stronger outlines and even add alerting.

For example, input a template or content brief example into an AI tool, along with an example page, then add specifics based on the draft your writer or freelancer is working on. You can save additional time by linking your keyword research or internal linking GPT to fill in intent gaps.

A good rule of thumb: Each custom GPT should have one major focus, then link to supporting GPTs. This avoids creating bloated tools that aren’t effective across multiple requests.

Automation example:

• Create a Jira survey that automatically generates a ticket.
• Use a custom ticket GPT to refine it before backlog grooming.
• Set up a Slack channel with the Jira app to send notifications when tickets open or close.

Time saved: 20 minutes per ticket.

5. Brand standards and compliance

If you deal with strict compliance or frequent legal reviews, work with those teams to create a custom business GPT. Let writers run high-risk drafts through it to catch simple issues before final review.

Prompt example 

Does this page use the correct terminology for the client? Highlight passages that go against editorial, compliance, and brand standards. Include a one-sentence explanation, a citation, and page number from the standards, and three better alternatives.

Time saved: 10 minutes per page.

6. Data validation and reports

Manual data validation is time-consuming. There are simpler ways to analyze and flag anomalies automatically.

For example, use conditional formatting in Sheets to highlight rows with more than a 10% deviation, or apply a color scale.

You can also use a GPT to diagnose why data doesn’t match across reports or align with your test hypothesis.

Prompt example 

Review this page performance report and data troubleshooting guidelines. Identify rows that skew the average the most. Export a table with those rows, and include a column explaining why each may cause issues and how to investigate further.

Time saved: 1 hour per 100 rows.

7. Metadata and schema

When creating or optimizing content, write or edit title tags and meta descriptions from scratch. Even if Google rewrites metadata in the SERP, it still helps signal what the page is about.

To optimize metadata, review top-performing combinations on similar pages. CTR is useful, though higher-ranking pages often get higher CTR regardless.

For schema, such as review or FAQ, manual creation increases the risk of errors. Even if schema isn’t automated, you can use formulas to generate it, then validate after publishing.

Prompt example

You’re a website writer updating a page. Turn these FAQs into a valid FAQ schema using the sample provided. Remove unnecessary formatting, keep links, and convert bullets into paragraphs.

Time saved: 10 minutes.

8. Formatting and shortcoding

If you use HTML, inline CSS, or shortcodes, automate as much as possible to reduce errors and inconsistency.

The simplest approach is using Excel or Sheets functions to concatenate code and content or format tables for your CMS.

For more advanced needs, create a formatting GPT that applies best practices and suggests improvements.

One watchout: If your AI tool can’t read live pages, you’ll need to paste the code first.

Prompt example

Using the formatting reference and brand guidelines, format this page using best practices. Include links, tables, CTAs, and engagement elements. Then check for issues like incorrect wording, missing closing tags, or stray brackets.

Time saved: 15 minutes per page

Make automation work for you, not the other way around

Automation doesn’t need to be complex or take time away from other work. The time savings compound with repeated use, freeing your team to focus on strategic work that requires human judgment.

Encourage your team to identify ways to reduce manual effort and experiment with custom GPTs to streamline repetitive tasks.