Most content audits start with a spreadsheet. Mine start with a question: “Where does this site show up when someone isn’t looking for it by name?“

That question shapes everything:

• Which pages I look at first
• What I flag as high priority
• What I end up recommending

It’s also why I scope content audits as discoverability audits rather than pure SEO audits. The two overlap, but they’re not the same thing anymore. A site can have technically sound SEO and still be nearly invisible in the places buyers are actually making decisions, places like AI-generated summaries, answer engines, and LLM-cited sources.

I’ll walk you through an example using a client site audit I recently completed. The site I audited belongs to a B2B SaaS company in the AI productivity space. They make a meeting notetaker that automatically captures, transcribes, and summarizes calls, and they’ve built real brand recognition in a crowded category.

But their organic presence doesn’t match their product reputation. They have a handful of competitor comparison pages, a resource hub with a couple of PDFs, and no blog. The whole site is running on branded search and whatever traffic their comparison pages can capture. That’s what I’m working with, and it’s a pretty common starting point for SaaS companies that grew fast on word-of-mouth and product-led growth before anyone thought seriously about content.

Here’s how I approach a discoverability audit for this client.

Before I look at a single page, I spend time with the site as a buyer would. I read the homepage, the pricing page, the solutions pages. I’m trying to understand: what is this product, who is it for, and what do they want someone to do when they arrive?

This sounds obvious, but it matters for the audit because discoverability problems are often positioning problems in disguise. A solutions page that’s not ranking isn’t always a keyword problem—sometimes it’s because the page is written for someone who already understands the product, not someone who’s searching for a solution to a problem.

For B2B SaaS, I’m specifically looking at:

• How the product is framed on the homepage (brand-first vs. problem-first)
• Whether solution pages map to buyer pain points or to internal team names
• Whether there’s a content program at all (and if not, what’s filling that gap)

That last one tends to be the biggest finding. A surprising number of B2B SaaS companies have no blog, no educational content, and no indexable articles. Their site is entirely dependent on branded search and whatever competitor comparison pages they’ve built. That’s a fragile content strategy, and it’s one of the first things I flag.

Once I understand the intent, I map what actually exists. Essentially, I build a mental model of the site’s content architecture:

• Owned pages: Homepage, product/features, solutions, pricing, integrations, comparison/competitor pages
• Resource content: Blog, case studies, guides, glossary, educational hub — or the absence of these
• Trust content: About, security/compliance, customer logos, testimonials
• Utility content: Help center, terms, privacy, status pages

The distinction between these categories matters because they have different jobs and different optimization approaches. A glossary page and a competitor comparison page are both “content,” but one is building topical authority over time and the other is targeting high-intent transactional queries. Treating them the same way is a mistake.

I’m also noting what’s missing from this map. A B2B SaaS site with a full competitor comparison section but no educational content has made an implicit bet that buyers are already category-aware when they arrive. Sometimes that’s right. Often it’s leaving a lot of top-of-funnel traffic on the table.

This is the step most content audit guides cover well, so I won’t spend too much time here. I’m looking at:

Are they present, specific, and intent-matched?

Auto-generated meta descriptions are a common miss. Google writes one if you don’t, and it can pull the wrong thing. I prioritize this fix on high-traffic and high-intent pages, like the homepage, pricing, comparison pages, and solutions pages.

Footer pages like Terms of Service don’t need meta descriptions — they’re not pages you’re trying to rank, and the effort is wasted.

Is the primary keyword in the H1, or is the H1 a brand tagline that sounds good but doesn’t match what anyone would search for? Both can coexist, but the title tag needs to carry the keyword weight if the H1 won’t.

Thin comparison pages and solutions pages are a common SEO gap I find on B2B SaaS sites. A page that’s mostly logos, ratings, and testimonials isn’t giving Google (or an LLM) enough signal to rank it confidently. The comparison pages that win typically include:

• A feature comparison table
• A clear “who should use this” framing
• An FAQ section
• Enough prose to establish context

Even if it’s not visible on the live side, finding “lorem ipsum” placeholder text is embarrassing. Missing images, unfilled template sections, and broken links or pop-ups are worse. These are often a site build (or rebuild) artifact.

They’re also an easy fix, and a fix that matters more than people realize, because it affects both crawler trust signals and the brand perception of anyone who lands on the page.

This is the layer most audits miss entirely, and increasingly, it’s where I spend the most time.

To make this concrete: the B2B SaaS site I referenced earlier—the AI productivity tool with the strong competitor comparison section and no blog—did a reasonable job on traditional SEO fundamentals. The comparison pages had decent structure, the title tags were functional, the site wasn’t technically broken. By standard audit criteria, they had a passing grade.

But when I ran the GEO evaluation, the picture changed. Querying several AI answer engines for category-level questions the product should have owned, like how bot-free meeting capture works, or what to look for in an AI notetaker for compliance-sensitive teams, the brand barely appeared.

A competitor with a weaker product but a robust blog and active community presence got cited repeatedly. My client’s content gaps weren’t just an SEO problem. They were a visibility problem in the places their buyers were increasingly going first.

That’s what GEO surfaces that traditional SEO doesn’t.

Generative Engine Optimization (GEO) is the practice of optimizing content to be cited, summarized, and represented accurately in AI-generated answers. It’s related to SEO but operates on different principles.

Google ranks pages. LLMs cite sources, synthesize answers, and build entity models. The content that gets cited tends to be the content that’s most clearly structured, most directly answers a specific question, and comes from sources the LLM has learned to treat as authoritative.

For my client’s B2B SaaS site, I evaluated GEO across three dimensions:

Are there pages on this site that directly answer the questions a buyer would ask an AI? Not “What is [product]“—that’s brand search. I mean questions like: “What’s the difference between a meeting notetaker and a conversation intelligence platform?” or “Do AI notetakers work without a bot in the meeting?”

If the answer to those questions isn’t on the site, the site is invisible in those answer moments — even if the product is the best answer to the question.

LLMs build models of what a brand is, what it does, who it’s for, and how it compares to alternatives. That model is built from owned content, third-party mentions, community discussions, and structured data. A site with no blog, no educational content, and no structured data is essentially asking an LLM to describe it based on whatever it scraped from G2 and Reddit.

I look for:

• Whether key entities (the product name, the category, key features) appear consistently and accurately across the site
• Whether there’s structured data (FAQ schema, product schema, organization schema) that helps LLMs disambiguate the brand
• Whether there’s any off-site entity reinforcement happening in the places LLMs treat as high-trust inputs, such as communities, review platforms, and third-party publications

This is the most tactical GEO layer, and it’s where understanding how AI systems actually process content pays off.

In RAG-based systems, which is the architecture behind tools like Perplexity, ChatGPT with browsing, and Google AI Overviews, content isn’t read the way a human reads it. It gets broken into small segments, typically a few hundred words each, and the model only receives the segments most relevant to the query.

If the answer to a buyer’s question is buried halfway through a 2,000-word article with no clear structure to signal where it lives, there’s a real chance it never gets retrieved at all. The model doesn’t read to the end. It retrieves the best chunk and stops.

That makes content architecture a retrieval problem, not just a readability problem. A page structured as one long narrative, even a well-written one, is harder for these systems to parse than a page with a clear question-and-answer format, a definition section up top, or a FAQ block that isolates discrete answers. The content doesn’t need to be shorter. It needs to be structured so that any individual section can stand alone as a complete answer to a specific question.

In this step, I flag pages that could be restructured with a FAQ section, a TL;DR summary, or a definition block, often without rewriting the full page. It’s one of the highest-leverage GEO improvements a site can make, and it’s almost always underutilized.

The output of all of this is a prioritized recommendation list, not a flat list of findings. The framework I use is a simple two-axis model: “How much impact will this have on discoverability?” and “How much effort does it require?“

High-impact, low-effort recommendations go first. Things like:

• Fixing placeholder text on live pages
• Adding FAQ schema to existing comparison pages
• Writing meta descriptions for the top 10 pages
• Tweaking a title tag

These are often one-hour fixes with real upside, and they demonstrate quick wins to my client, who is most likely already buried in work.

High-impact, higher-effort recommendations go in the middle:

• Building out the blog with foundational SEO and GEO content
• Deepening existing comparison pages
• Creating solution pages that map to buyer pain points rather than internal team names
• Converting PDF case studies to indexable web pages

Lower-impact items, and anything that requires significant technical lift without proportional content benefit, go at the bottom or get noted as longer-term considerations.

The goal is a document that anyone on the marketing team can pick up and know exactly where to start.

The sites that win organic discoverability over the next few years aren’t necessarily the ones with the most content. They’re the ones whose content is structured to be understood by crawlers, by LLMs, by the AI systems that are increasingly mediating the relationship between a buyer’s question and a brand’s answer.

Most B2B SaaS companies built their content strategy for a world where ranking on page one meant writing the most comprehensive blog post and building out a broad (but often not deep) scope of blog content. That world still exists, but it’s a smaller part of the picture than it used to be. Increasingly, the question is: “When someone asks an AI about your category, does your brand show up, and does it show up accurately?“

That’s what I audit for. And it’s a significantly different question than it used to be.