Your existing content can be a goldmine if you know how to improve it. Finding the time is a challenge in itself. I’ve found that using Claude makes what can be an outsized task much more manageable.
If it feels daunting, you don’t have to build a massive content audit workflow from the start. With Claude, you can start by auditing a single article, iterate on it, and create skills you can reuse and refine over time. I’ve gone from one-off audits to a full library of skills that I refine every time I use them.
Claude can help you uncover topical gaps, flag outdated information, audit brand voice, and more. You just need to invest the time to iterate so the value of your Claude skills compounds.
Here are six types of content audits you can run with Claude. The first four work at the article level, so you can start with a single piece today.
Page-level audits
If you’re uncomfortable creating skills or workflows, page-level audits are a great starting point. These four audits work on a single article, with no content inventory, no data exports, and little setup required. At the end of each session, ask Claude to create a skill you can reuse for future page-level audits.
1. Brand voice consistency
Content libraries drift over time because of changes in brand voice, staffing, products, and services. A brand voice consistency audit can help you identify what needs to be updated so a piece better aligns with your brand guidelines.
Unless you have detailed brand guidelines with plenty of examples, let Claude extract your voice guide from high-quality content. This helps remove subjective language common in many brand guides, such as “conversational but authoritative” or “educational, not too formal.”
Pick three to five articles as your standard bearers. Download them as markdown files if you can, then send them to Claude with a prompt asking it to describe:
- How the articles typically open (for example, do they open with a direct claim, a counterintuitive statement, or a concrete scenario?)
- Sentence and paragraph construction: average length, range, when sentences get longer or shorter, how paragraphs tend to close, and so on.
- Personality dimensions: three to five “We say X, but not Y” pairings, each with a do and don’t example.
- Vocabulary: words and phrases to use, and words and phrases to avoid.
- What this brand never does: specific constructions, phrases, and conventions absent from every piece.
Instead of “conversational but authoritative,” you should get:
- Observations such as “Your articles open with a direct claim rather than a scene-setting paragraph, sentences average 15 to 20 words and rarely exceed 30, transitions are functional (i.e., ‘here’s why that matters’) rather than formulaic (‘furthermore’).”
- Example pairs such as “We’d say ‘the data shows three things,’ not ‘there are multiple factors to consider.’”
The goal isn’t to produce a voice guide your writers can use. It’s to create one an LLM can understand and apply.
Once you’re happy with the output, have Claude save it as a skill and evaluate an article with it. If it calls out things you don’t agree with, have Claude update the skill until you’re satisfied with the output.
You can use it to find brand inconsistencies in older content, check new content for alignment, and even start generating on-brand content, though you’ll still want to edit it.
Dig deeper: How to train Claude to sound like your brand
Be the brand AI recommends.
See where your brand appears in AI search, where competitors are winning, and what it takes to become the answer AI recommends.
See your AI visibility
2. Coverage comparison
If you need to improve content performance, a coverage comparison can help you identify topical gaps in your articles.
Use the Claude in Chrome extension to have Claude scrape content from the top three to five ranking pages for your target keyword. Ask Claude to analyze your content against the scraped content and highlight any gaps. You want to know:
- What your competitors are doing well.
- What you’re doing well.
- Where you can improve your piece.
If you prefer this information in a table, tell Claude to output it that way. Or, if you’d rather have a downloadable DOCX, ask for that.
If Claude recommends things you’ll never add to your content, make note of them when packaging the prompt into a skill.
3. Freshness audit
Old content adds up, and it’s hard to focus on refreshing it when you’re also creating new content. Creating a skill that identifies what needs to be updated so you don’t have to read every old article can save you a lot of time.
Feed Claude an older article and ask it to flag everything that’s time-sensitive: statistics with years attached, named tools or platforms, references to “current” or “recent” trends, and any claims that depend on a specific market or regulatory context that may have shifted. You’re not asking it to rewrite anything. You’re asking for a list of issues so you can update them.
If you have new products that may not have been mentioned in older articles, products you no longer carry, or similar changes, include that context in the input so Claude can flag content to remove or add.
Dig deeper: How to turn Claude Code into your SEO command center
4. AEO and AI retrievability
Answer engine optimization (AEO) focuses on making content visible in AI-generated responses. AI tools, including ChatGPT, Perplexity, and Google AI Overviews, tend to surface content that answers questions directly. If your article buries the answer under three paragraphs of preamble, or structures key information in a way that’s hard to extract, it’s less likely to appear in those responses.
Give Claude an article alongside the target query and ask it to evaluate:
- Whether the piece answers the question directly and early. LLMs scan from the top down and weigh the opening heavily.
- Whether key statements are specific enough to quote directly. An LLM is less likely to extract or cite vague language because it’s harder to confirm.
- Where an FAQ-style section would help.
- Whether the page has clear authority signals, such as outbound links to primary research, first-person experience, or specific examples.
Save this as a skill, and it becomes an extra editor to help improve AI visibility.
Get the newsletter search marketers rely on.
Library-level audits
These audits require access to your performance data or content inventory, either through a connector or a manual export.
5. Performance triage
When people think of a content audit, they tend to think of analyzing a content library to uncover performance issues. Performance triage is exactly that kind of audit.
Before you start, make sure you have access to your data via a connector such as BigQuery or the Semrush API. If not, export the data you normally use for large-scale content audits (traffic, clicks, engagement metrics, conversions, rankings, and so on) as input.
Ask Claude to prioritize: pages that have suffered meaningful performance drops in the past six to 12 months, pages with high impressions but consistently low click-through rates, and pages that have been live long enough to rank but never have.
Be clear about what constitutes a performance drop for your website, since traffic varies across sites and industries. Ask Claude to output a prioritized list of what’s worth investigating and why. From there, the page-level audits above provide the diagnostic framework.
If you’ve run this analysis before, give Claude the previous output so it better understands what you’re looking for.
Dig deeper: How to build a Claude Code-powered second brain for agency work
6. Topical gap analysis
Entities are a major part of AEO and semantic search. A topical gap analysis can help uncover whether you have enough content to build authority around the entities related to your brand.
The basic question is: What isn’t your content library covering that it should?
To get started, you’ll want a list of those entities as one of your inputs. For example, at my agency, we want to be known for SEO and AEO. If you have a clear list of services or products, you can use that instead of a list of entities.
Using Cowork or Code, have Claude analyze your sitemap and compare it to your target entities. Alternatively, if you have a Screaming Frog export with URLs, page titles, and meta descriptions, use that as an input for a more accurate analysis.
Ask Claude to identify topic clusters that are underrepresented or absent based on your target entities or products. If you want a prioritized list, you can use the Semrush MCP to have Claude check search volume for potential keywords.
Not every gap is worth filling. Filter the results to match your audience’s needs, and reiterate any changes you make to Claude to create a skill that matches your desired output. Your final list can then feed directly into your content creation workflows or be handed off to your content team.
If AI can’t find you, customers won’t either.
Track your visibility across AI search, uncover missed opportunities, and grow your presence where customers are asking questions.
See your AI visibility
Don’t try to audit everything at once
Content audits stall not because teams lack data, but because the scope feels too large to start. Pick one audit and one article. Run it, save the skill, and use it again on the next piece.
For me, iteration is part of the fun. I enjoy taking a skill and improving it, then chaining it with other skills to uncover even more content opportunities. I hope you’re able to build one of these this week.
Contributing authors are invited to create content for Search Engine Land and are chosen for their expertise and contribution to the search community. Our contributors work under the oversight of the editorial staff and contributions are checked for quality and relevance to our readers. Search Engine Land is owned by Semrush. Contributor was not asked to make any direct or indirect mentions of Semrush. The opinions they express are their own.
