Large Language Models (LLMs) (like ChatGPT and Claude) and Answer Engines (like Google AI Overviews or Perplexity) are becoming key entry points for buyers who are researching solutions. But not all queries are created equal. The more niche or specialized the topic, the more likely vendor content will be featured and referenced in the response. For companies looking to influence how they show up in these systems, this is an important opportunity. This applies for B2B and B2C brands, but the prevalence of non-vendor content is likely to be higher in B2C searches, other than in very niche topics.

Why Niche Topics Drive Vendor Coverage

In broader or more established categories, LLMs tend to surface more third-party sources like analyst sites, established media outlets, and review platforms. Vendor content typically makes up around 25–30% of references.

But shift to narrower, verticalized, or newer topics, and the ratio flips. In these areas, vendor content can account for 60% or more of what shows up in model responses. That’s because third-party coverage simply isn’t as deep or abundant in new or highly specific categories.

Real-World Examples

Here’s what this looks like in practice. I recently ran these three searches in ChatGPT. You can see how the percentage of vendor references (utilization of vendor information in the response and a link to vendor sites) shows up in each response:

  • CRM (general): “I’m in charge of selecting a new platform for our company. Which CRM software should I include in my short list?”
    → Vendor references: 25% | Non-vendor: 75%
  • CRM for Healthcare: “I’m in charge of selecting a new platform for our company, a mid-sized healthcare company. Which CRM software should I include in my short list?”
    → Vendor references: 40% | Non-vendor: 60%
  • CRM for a Local Business: “I’m in charge of selecting a new platform for our company, a small local flower business. Which CRM software should I include in my short list?”
    → Vendor references: 83% | Non-vendor: 17%

Boosting references in LLMs example

The same pattern shows up in searches for customer data platforms (CDPs). A general query will surface more independent sources, while enterprise- or industry-specific questions skew heavily toward vendor content.

Other Niche Searches 

SUVs, Cloud GIS solutions, Synthetic Data, Skateboard Shoes

This pattern holds across different industries and tools:

  • Reviews & Rankings
    For broad queries like “Top-ranked SUVs” (non-vendor 100%) or “Best email marketing platforms” (67% non-vendor), LLMs skewed heavily towards third-party sites (e.g., analyst reports, review aggregators).
  • Cloud GIS
    When I looked into cloud GIS providers, ~78% of references came directly from vendor sites (Esri, CARTO, Google Earth Engine, GIS Cloud, Felt, SuperMap). Independent sources were minimal (Wikipedia, Penn State). This again shows that in specialized tech categories where the coverage may be thin, vendors dominate the knowledge base.
  • Synthetic Data Providers
    In searches for synthetic data companies, around 70% of citations were from vendors like Gretel.ai, Tonic.ai, Synthesis AI, and MOSTLY AI, while the rest came from neutral sites (Wikipedia, analyst blogs).
  • Skateboard Shoes for Girls
    In searches for a very specific niche within a broader category, vendor content dominates (~73%) with the non-vendor being primarily retail sites. Likely no editorial content exists for this very specific topic. Of course, it may have zero or low search volume as well. So, it’s important to find and create content for niche topics that people are researching with some degree of volume (e.g. by using SEO keyword research tools).

These examples reinforce the point: niche, emerging, or technical categories leave a vacuum that vendors themselves fill with content, as long as they know to create it and they’ve optimized it to be discovered and easily digested by the LLMs or Answer Engines.

ChatGPT vs. Perplexity: Different Balance of Sources

One interesting observation is how ChatGPT and Perplexity differ in their reference balance.

  • ChatGPT often blends vendor content with analyst sources, but in niche categories, vendor content still takes the majority share (sometimes 60–70%).
  • Perplexity surfaces more citations overall, including forum posts, blogs, and news outlets. My impression is that it leans towards non-vendor content where possible. For example, you can see the following:
    • For a query about Enterprise CDPs, it was 100% dominated by non-vendor references.
    • For a query about synthetic data sources for AI training, vendor sites only represented 10% of the references. 
    • Cloud GIS solutions was a mix of 36% vendor and 64% non-vendor references.
    • For a very niche topic, like skateboard shoes for girls, it as 73% vendor sites (brands) and 27% non-brand references.
    • Best family SUVs was 100% non-brand references.

The takeaway is that no matter which system a buyer is using, niche queries or categories will tilt toward vendor-owned content.

“Publishing content for areas not widely covered but very important commercially to a vendor, can be a huge win in terms of visibility and qualified traffic from LLMs/AEs” -Tom Treanor

Companies that publish useful, detailed, and optimized content on those topics are more likely to be included.

Custom GPT banner

 

Implications for Your Content Strategy

If you’re in a new, specialized, or smaller category, you have a unique chance to punch above your weight in how LLMs present answers. But for any company – whether you’re in a specialized category or not, here are ways to maximize your visibility and references in LLMs or Answer Engines:

  • Create content around niche questions
    Think beyond broad solution overviews. Address specific use cases, sub-niches, key features, technical capabilities or build detailed pages on specific questions that your buyers are likely to ask. This is where traditional keyword research tools can help you find topic clusters with meaningful volumes of searches that will likely translate to LLM inquiries, as well.
  • Double down on vertical content
    If you serve specific verticals like (e.g. healthcare, retail, or financial services), make sure you publish detailed content for those audiences including how your solutions help them solve their specific problems, blog content and vertical guides, industry case studies and more. Vertical content is more likely to be referenced when LLMs try to answer specialized questions.
  • Optimize your site for LLMs (and search engines)
    Address specific questions. Write shorter, clearer sentences. Include FAQ on pages or in metadata for each page. Cross-link to related resources. Focus on clarity with page title, headlines and meta description. Make your site is compatible with how LLMs digest it, but don’t throw out traditional SEO.
  • Build credibility with third-party sites (including topical or vertical sites)
    While vendor content dominates in niche areas, getting cited on industry publications and analyst blogs remains important. These sites not only influence SEO but also help you appear in more balanced LLM answers. Do outreach, provide samples or demos to site owners/media, encourage reviews, invest in PR and do bylines/guest posts.
  • Build a Company FAQ page:
    Build a company FAQ page that LLMs can reference that details multiple aspects of your company including basic “about” information, key products, verticals served, etc. H/T to Rosemary Brisco at ToTheWeb for this one!

Don’t Neglect Your Company Content Strategy and Own Critical Niches

LLM optimization depends on a robust and data-driven content strategy and strong execution on the content production and optimization aspects. In addition, getting third party references is an important part of the strategy and this comes via creating great, unique and useful content that other sites want to reference, doing outreach to other sites, soliciting reviews, and robust PR strategies.

And every company wants to be found for the broad, high-volume questions—but the real opportunity is often in the niches. Specialized, vertical, or company-size–specific topics can tip the scale in your favor and give you an outsized presence in LLMs and answer engines. For your outreach or PR, getting featured on vertical sites or on topics that address specific niche audiences, capabilities or features can also help.

With a robust content strategy that goes both broad and deep into meaningful areas, you’ll not only strengthen your position in AI-driven responses but also meet your buyers where they are—searching for solutions tailored to their unique context.