Consider ways to supply data to AI systems beyond just standard web content:
- Participate in Q&A Platforms: Sites like Stack Exchange, Quora, Reddit (in informative subreddits) – these have historically been mined for training data. If you, as an expert, answer questions there (and mention your brand where appropriate, though carefully to avoid seeming spammy), those answers might become part of the AI’s knowledge. For example, a high-quality Quora answer you provide might later be paraphrased by ChatGPT when someone asks a similar question. This is indirect GEO – you’re optimizing content on third-party platforms to influence AI outputs. It also has the side benefit of direct visibility on those platforms.
- Provide Structured Knowledge to Models: This is more future-looking, but initiatives like OpenAI’s plugins or Google’s Dataset Search are ways to feed data. If you have a lot of data or a knowledge base, creating an API or dataset that can be ingested by AI assistants (with your branding on it) is a route. E.g., releasing a well-structured dataset or CSV of common industry metrics with your site credited. It could be that an AI finds that data and uses it. Or, if AI assistants allow custom integrations (like voice assistants did with “skills”), you might eventually create a plugin that ensures your content is used for relevant queries (e.g., a travel company might have a plugin so that “What’s the best time to visit Paris?” uses their data).
- Optimize for Specific AI Platforms: If you know a lot of your audience uses, say, Bing Chat, test your queries there often. Bing allows some level of choice in answers (balanced, creative, precise). See if one mode picks your content more. If yes, perhaps it likes more detailed (creative) or more factual (precise) writing. Tailor accordingly. Similarly, if you have a product and you know ChatGPT plugins like Instacart or Kayak could mention it, see how to get listed or optimized for those ecosystems.