Now, examine your actual content inventory and how it might be perceived or used by generative models:
- Identify Key Content Pieces: Make a list of the content that aligns with common user questions or interests in your niche. This could be blog posts, knowledge base articles, product pages with detailed info, whitepapers, etc. These are the pieces likely to be pulled into AI answers when those topics come up. For each, ask: does this content directly answer questions? If you have content that’s very narrative or meandering, consider that an AI might skip it in favor of a competitor’s succinct Q&A style content. Flag content that could be reformatted or summarized.
- Search Your Content in AI Engines: Perform some tests: take a few sentences from your top pages and ask ChatGPT or Bing Chat about that topic (not the exact text, but the topic). See if the AI’s answer includes information that clearly came from your content. If it does and cites someone else, that means your competitors’ similar content might be outranking you in the AI’s mind. If it doesn’t include info from your page at all, then either the AI didn’t see your content or didn’t find it valuable. As an example, if you wrote “The Complete Guide to Widgets,” try asking the AI “What are widgets used for?” or something covered in your guide. This exercise can highlight whether your current content is being recognized.
- Brand Presence in AI Outputs: Ask generative AI directly about your brand, products, or key figures in your company. For instance, “What is [Your Company]?” or “[Your Product] vs [Competitor]”. Also try more indirect queries: “What are the top companies in [your industry]?” and see if you are mentioned. Document these responses. Are they accurate? Up-to-date? If the AI has outdated or incorrect info (e.g., an old CEO name or a completely wrong description of your services), that’s a red flag. It means the underlying data (like Wikipedia or news articles) might be outdated or insufficient, which you should address in your Entity strategy (Step 3).
- Existing Citations and Links: Use tools like Ahrefs or Majestic to see where your site is already referenced or linked on the web. Why is this relevant to GEO? Because if authoritative third-party sources mention you or cite your research, AI might pick up on that. For example, if Wikipedia cites your blog for a fact, that’s very valuable – many AI models give weight to Wikipedia data, and via it, your content could influence answers. Document any such high-value citations. Also search the web (Google it, or use Bing) for your brand name + keywords like “report” or “study” to find if any of your content has become a reference point. If none of your content is being cited elsewhere, an AI might have less “reason” to present your site as a source.
- Content Gaps: Identify topics in your domain that users care about where you have little or no content. Use keyword research tools and also think of natural language questions. For example, if you’re in finance and people often ask AI “How can I save more for retirement?” do you have a post that directly addresses this question? In the GEO age, covering user questions comprehensively is key. If gaps exist, note them – this informs your content creation plan (Step 4). Also check if there are questions where only competitors’ content gets cited by AI. If, say, Perplexity or SGE frequently shows competitor X’s blog for certain queries, that’s a gap to close.
- Content Depth and Quality: Evaluate whether your content is truly helpful and authoritative. AI systems use signals of quality much like Google’s algorithms do (they even use Google’s indices as a base). Thin or fluffy content will be ignored. Look at engagement metrics: high bounce rates or short time-on-page for a piece may indicate it’s not satisfying users, so likely an AI won’t find much value either. Also, is your content up-to-date? If you have great content from 2018, consider updating it with 2024 data and republishing – fresher information is more likely to be included by AI looking for current info.
- Use of Sources and Tone: Analyze if your content currently includes outbound citations, statistics, or expert quotes. Interestingly, adding these elements can both improve user trust and appeal to AI. A research study showed that including credible citations, quotes, and statistics in your content can significantly boost its likelihood of being included in generative answers (improving visibility by over 40%). Does your content do this? If not, mark it for improvement. Also check the tone: overly promotional content (“Our product is the best!”) is less likely to be used verbatim by an AI compared to objective, informative content. You might need to balance content geared for conversion with content geared for information/education (which is what AI will favor).
Performing this content audit from an AI perspective will highlight which pages to focus on and how to enhance them. You might discover, for example, that you have a fantastic guide buried in your site that hasn’t gained SEO traction but could be perfect for AI – you’d want to surface and optimize it. Or you might realize many of your pages answer the same question (duplicative content), which could confuse AI retrieval; you’d then consolidate those.