Contents
In the first article, we looked at where AI is useful in SEO and why the “one prompt, one finished page” approach does not work. In this article, we have collected ready-to-use prompts and AI tools that can help automate SEO work, from content creation to analytics and technical audits.
Here is that first article
Using AI in SEO: opportunities and risks
Scenario 1. Content and semantic automation
In this scenario, AI is used as part of the workflow. It does not replace the writer.
| Task | Where to do it | Example prompt |
|---|---|---|
| Topic research | Perplexity / AI search | Gather up-to-date information on [topic]. Give me 10–15 key points, each with a source link. Also list any disputed or debatable points separately. |
| Intent analysis | ChatGPT | I am optimizing a page for the query [query]. Describe the types of pages ranking in the top 10, the content format, and the search intents behind them. |
| Article structure | Claude | Create an article outline for [topic]. Format: H2/H3 headings with a short description of each section. |
| Writing sections | Claude | Context: I am writing an expert article for an IT blog. The audience is business leaders and marketers. Task: Write the section “Benefits of marketing automation.” Requirements: 800–1000 words, expert tone, include the keywords marketing automation, CRM system, lead magnet. Format: 3–4 subsections with H3 headings, each subsection 200–250 words. Examples: Articles by HubSpot and Salesforce on this topic. Constraints: Avoid filler and generic phrases. Focus on specific value. |
| SEO refinement | ChatGPT | Optimize the text: add related terms, improve the headings, and point out weak spots. |
| FAQ | ChatGPT | Generate 5 FAQs on [topic] for search snippets. |
Scenario 2. Analytics and reporting automation
| Task | Where to do it | Example prompt / formula |
|---|---|---|
| GSC export analysis | ChatGPT / Claude | Analyse this Google Search Console export. Find queries where impressions are growing but CTR is falling. |
| Finding reasons for a drop | ChatGPT | Why might the page [URL] have dropped? Give me 5 hypotheses. |
| Page segmentation | Google Sheets | =REGEXEXTRACT(A2,"https?://([^/]+)") |
| Title length check | Google Sheets | =IF(LEN(B2)>60,"Too long",IF(LEN(B2)<30,"Too short","OK")) |
| Business report | ChatGPT |
Create a monthly SEO report based on the following data:
Report structure: 1. Executive Summary 2. Traffic and ranking trends 3. Completed work 4. Issues identified 5. Plans for next month Style: professional, with specific numbers and clear conclusions. |
Scenario 3. Technical SEO automation
| Task | Where to do it | Example prompt |
|---|---|---|
| Technical audit | ChatGPT / Gemini |
Analyse the technical characteristics of the site [URL] and create a prioritised optimisation plan.
Check:
For each issue, specify:
|
| Schema.org | Gemini |
Create JSON-LD Schema.org markup for [page type: article / product / service / organisation] using the following data: [specific data].
Use current schema types and required properties. Make sure the markup is valid and follows Google recommendations. |
| Meta tags | ChatGPT |
Create a compelling meta description for the page [URL] with the primary keyword [keyword].
Requirements:
Or Create an SEO title tag for a page about [topic] with the keyword [keyword]. Criteria:
|
| Developer brief | ChatGPT | Write a development brief for fixing the issue [description]. |
Important: this is not a replacement for an SEO audit. It is a way to prepare briefs and standard checks faster.
Scenario 4. Competitor analysis and semantic research
| Task | Where to do it | Example prompt |
|---|---|---|
| Competitor analysis | ChatGPT |
Run a detailed SEO analysis of a competitor based on the following data:
Website URL: [URL] Top 10 organic keywords: [list] Main content categories: [list] Analyse:
For each point, suggest specific actions that could help outperform the competitor. |
| Gap analysis | ChatGPT | Which topics does the competitor not cover? |
| Clustering | ChatGPT |
I have a list of [N] keywords on the topic [topic]: [list]
Group them into topical clusters based on:
For each cluster, specify: Cluster name
|
| Cannibalisation check | ChatGPT | Find keyword cannibalisation in this list of URLs and keywords. |
Limitations and pitfalls
1. Hallucinations
AI can be confidently wrong. Facts, numbers, case studies, and research should always be checked.
2. Data validation
A model does not know which rows in a table are junk and which ones matter. Any analysis built on unclean input data will produce distorted output.
3. Service limits
API restrictions, query limits, and service instability still matter.
4. Over-reliance on tools
When the whole process turns into “the AI said it, so we did it,” several problems follow:
- analytical skills get weaker;
- there is no prioritisation;
- decisions are made without understanding why they matter for the business.
