5 Ways to Expand Beyond Pages Into Formats AI Can Learn From

Traditional SEO often meant focusing on web pages and blogs. AI, however, learns from and surfaces multiple content formats – text, videos, audio transcripts, code, and user-generated content across the web. 

To capture AI referral traffic, you need to spread your expertise across formats so that whatever medium an AI model or assistant draws from, your brand is present. 

1. Go Multimedia with Video & Audio (and Use Transcripts) 

AI models don’t just train on written blog posts – they’re always ingesting knowledge from videos, podcasts, and beyond (via transcripts and text associated with those media). So, extend your content strategy to video and audio formats. 

Create instructive videos (webinars, demos, interviews) and upload them to YouTube – not only is YouTube the world’s second-largest search engine, but AI systems frequently crawl and even cite YouTube content in answers. We’ve observed that “videos” were among the top 5 content types cited by Google’s AI overviews, which means a well-optimized video can put you in an AI summary where a plain web page might not appear. 

Likewise, consider launching a podcast or converting some blog posts into audio form. Microsoft’s Copilot and other assistants might fetch information from popular podcast transcripts or audio if that content is surfaced on the web. 

The Crucial Part of Posting Videos and Podcasts: The Technicals

Always provide transcripts or textual summaries for your videos and podcasts. AI can’t (yet) “watch” or “listen” to raw media in real time, but it can read transcripts. When we host webinars or videos, we transcribe them and even repurpose the transcript as a blog article. This creates a text asset that AI can index. 

For podcasts, write detailed show notes or full transcripts on your site – not only do you cater to hearing-impaired humans and SEO, but you also hand AI models text they can learn from. 

In one instance, our team noticed that a direct quote from a podcast (present in its posted transcript) was used verbatim by an AI answer tool – a clear sign that transcripts extend your reach into AI responses.  

Lastly, multimedia content often earns higher user engagement (time on page, shares), which can indirectly signal quality to AI (and definitely to Google). And there’s a bonus: being on platforms like YouTube or Spotify also increases brand visibility. 

Users may ask AI, “Show me videos about [topic],” and if you have a YouTube presence, the AI might recommend your video. Overall, a multi-format approach ensures whether someone prefers to read, watch, or listen, your content is accessible – and every format can reinforce the others. 

2. Participate in Q&A Platforms and Forums 

Some of the content that trains AI models and powers their answers comes not from polished articles, but from Q&A discussions and forum posts. By being active on platforms like Stack Exchange, Quora, Reddit, or relevant industry forums, you can put your expertise (and subtle brand mentions) directly into the places AI is learning from. 

For example, if you provide a high-quality answer on Reddit to a question about your industry, that content might later inform ChatGPT’s knowledge or appear via Bing’s integration of Reddit data. In fact, both OpenAI and Google struck deals with Reddit to access its data for training, underscoring how influential forum content can be. We recommend identifying the top 2-3 Q&A communities in your niche and making it a habit to answer questions there under your real name or brand account.  

Reddit Answers matter – are you paying attention to their input?

When you do, be genuinely helpful and avoid overt self-promotion (most communities frown on spam). However, it’s acceptable to mention your brand or link to your content if it’s truly relevant and permitted by the forum rules. The goal is to seed the public web with authoritative answers that include your key insights and even your brand name in context. 

Over time, these mentions can accumulate. LLMs weight mentions of brands that frequently appear near certain topics in their training data. Our own strategy involved engaging in conversations on Reddit and niche forums, dropping knowledge and occasionally citing our success, which helped amplify our credibility and put Xponent21 into the broader AI SEO discussion online. 

Beyond training data, these platforms can also drive direct referrals. For instance, users of Perplexity AI often see references from Quora or StackExchange in the answer citations. By answering there, you increase the chance of getting cited in an AI-generated answer with a backlink to your post. It’s a win/win: you help real people and simultaneously plant seeds for AI visibility. Just remember to keep your tone conversational and your explanations clear – if your answer looks like something an AI would consider a concise solution that’s informative rather than a salesy call-to-action, you’ve done it right. 

3. Encourage Authentic User Reviews & Testimonials 

User-generated content (UGC) – especially reviews and testimonials – carries a lot of weight in AI recommendations. When an AI is asked for the “best project management software” or “top restaurants in Chicago,” it doesn’t just rely on branded content, it considers aggregated user opinions and reviews. 

Make sure your business has a healthy presence on major review platforms relevant to you (Google Reviews, Yelp, G2, Capterra, TripAdvisor, Amazon, etc.). Encourage satisfied customers to leave detailed reviews that mention specifics of your product or service. LLMs training on public data may pick up on recurring positive sentiments or descriptions from these reviews, influencing how the AI perceives your brand’s reputation.  

We’ve seen that LLMs often pull from trusted, authentic UGC that fits a prompt perfectly. For example, if multiple reviews mention your brand’s key strength (“the customer service was so responsive at X company”), an AI might incorporate that point when asked about you or compare you to competitors. Moreover, some AI search tools explicitly surface reviews – e.g., a ChatGPT plugin might fetch top Amazon reviews for a product query. 

By amplifying your positive UGC, you increase the chance an AI assistant will recommend you. The machine takes nods from humans, so the real human reviews are sure to influence how the machine recommends solutions.

One strategy to make the most of UGC is to integrate review acquisition into your workflow: after a successful engagement or sale, politely ask the client to share their experience on a particular platform. Even on your own site, showcasing testimonials (with permission) can help; ensure they’re marked up with review schema for extra machine readability. 

Beyond formal reviews, testimonials and case studies can also be fodder for AI. A well-crafted case study on your site, if picked up in search results, could be summarized by an AI to a prospective customer. 

We once noticed an AI answer citing a case study snippet that listed quantifiable results, treating it as a factual reference. The takeaway is to generate plenty of genuine, positive talk about your brand online. Not only will it influence human prospects, but it also serves as reliable data points for AI algorithms that highlight brands people trust. 

As a bonus, when users do click through from an AI referral, seeing strong reviews and testimonials will reinforce the AI’s recommendation, creating a trust loop that boosts conversions. 

4. Publish Original Research & Data Resources 

If you have access to unique data or insights in your industry, leverage that by publishing original research – think surveys, studies, or data analysis posts. AI models love reliable data points and often regurgitate statistics or findings they’ve seen across the web. By contributing novel data to the public domain, you increase the likelihood that your information gets cited in AI-generated content. 

For instance, if you run an HR software company and you publish an annual report on hiring trends (with shareable stats like “X% of companies plan to use AI in recruiting in 2026”), those stats might get quoted by blogs, news outlets, and eventually AI answers on related questions. 

Additionally, consider contributing to Wikipedia or Wikidata where appropriate. While you should not write an overtly promotional article, you can add well-cited information about your field or add your company to Wikidata (the knowledge base that feeds Google’s Knowledge Graph). AI systems ingest a lot of Wikipedia content, so being part of that ecosystem lends credibility and visibility. 

Similarly, releasing data sets or tools on open platforms (like a GitHub repository with a useful dataset, or an interactive calculator hosted on your site) can get your brand name and insights into places where AI might pick them up. 

For example, OpenAI’s models have been partly trained on content from sites like GitHub – if your organization open-sources code or datasets, an AI that was trained on that content might later mention your tool or findings when relevant. 

To maximize this, make your research easy to find and reference. 

• Use clear titles like “2025 State of [Industry] Report – [Company] Research”. 

• Include executive summaries (or TLDRs) that an AI could easily quote. 

• Use charts or infographics (with alt text descriptions) that bloggers might embed (spreading your data further). 

Being cited widely increases your authority signals. In the modern search era, the concept of backlinks is less important – what’s worth focusing on is the more abstract idea getting “mindshare” in the training data and real-time answer generation sources.

One authoritative study or insightful whitepaper can earn your brand dozens of mentions across the web, which in turn positions you as a go-to source for AI answers on that topic. (And of course, it earns genuine respect from human peers and prospects, strengthening your brand overall.) 

5. Build AI Integrations or Tools (to Embed Your Brand) 

This is a more advanced and unconventional strategy: create integrations with AI platforms so your content or service is directly available through them. For instance, develop a ChatGPT plugin that lets users query your knowledge base or use your product via ChatGPT. If you’re a travel company, a plugin for ChatGPT or Bard could allow those AIs to pull live data from your site (flights, hotels) and recommend them to users in conversation. Microsoft’s Copilot and other assistants are increasingly allowing third-party connectors; being an early mover here could put your brand at the forefront of AI referrals when users ask for something your tool can handle. Essentially, you’re making your own content/service a part of the AI’s function. 

Xponent21 created a custom GPT for MVQs – otherwise known as Most Valuable Questions. Releasing this free and publicly allowed us to share our expertise on FAQ strategy with any given brand because it is a blanketed approach that can add value with our digital agency’s specific skill set. 

Even simpler that a GPT, consider chatbots or AI assistants on your own platform that can be accessed externally. For example, if you create a public-facing chatbot (using your data) and list it in an AI directory or allow others to link to it, you might get referrals from curious users of AI aggregators. 

The goal is to embed your expertise directly into AI ecosystems. It’s not traditional referral traffic in the sense of a click from a website, but it earns you visibility and engagement that can lead to more formal referrals. 

For instance, an executive using Microsoft 365 Copilot might ask for a SWOT analysis template, and if your company has a popular published template or tool that Copilot finds, it could present it (with attribution). Or a developer might use an AI coding assistant that suggests a function from your open source code – introducing them to your brand. These are subtle, but in 2026 the lines between “traffic” and “presence” blur. We strongly believe in consistent brand presence across digital ecosystems – it’s a core tenet of our AI SEO Blueprint. 

The more touchpoints an AI has to encounter your brand – whether via plugin, integration, or content – the more you’ll be part of the answers it gives. Just ensure any integration truly adds value (spammy or gimmicky plugins will be ignored or removed, doing more harm than good). When done right, being woven into the fabric of AI interactions can yield high-quality leads and users who come through non-traditional channels but are highly primed by the AI’s implicit endorsement. 

This blog was originally published with Xponent21 – read more of my articles here.