[aka, Why falling in love with the first person you meet might be a mistake]
All AI platforms are not created equal. In fact, they do not even rely on the same sources to answer your questions. Some draw primarily from live web search and citations, while others depend more heavily on training data, licensed content or private knowledge bases. So, if you ask the same question or give the same prompt to five different AI platforms, you are likely if not surely to get five different responses.
And that could be a problem if you are one of those people (Dylan Field calls them “brave”) who is content to prompt, read and run.
So let’s start with the obvious: the way people discover information online is changing quickly. While most still rely on Google (or Bing? Or TikTok? Or YouTube?) in search of information and/or entertainment, and some just randomly float through the internet universe pinging from one website to the next, a growing number of users are turning to AI (whether they realize it or not).
And because AI does not act like Google, simply serving up a bunch of ads (er, sponsored links) and endless pages of websites to vet, but instead provides actual answers synthesized from instantaneous scrapes of multiple sources, there is a tendency for or even a likelihood that the average user might accept its seemingly flawless answers (responses) as gospel.
And this, I would like to suggest, is a huge mistake for a host of reasons. As I already indicated, since different AI platforms – Claude, Gemini, Perplexity, ChatGPT, etc. – all use different sources, they are all likely to give different responses. Different platforms use different combinations of web crawls, proprietary data, licensed datasets and live retrieval, so the same question can produce different answers across tools. Even when two systems both use the web, they may differ in how they rank sources, determine freshness, and decide what to cite.
Does AI use the best sources?
Meh. AI systems tend to favor content that is easy to parse and useful to summarize. They prefer clear headings, direct answers, structured sections, FAQ blocks and comparison tables. Freshness also matters, especially for timely topics, because AI-driven search often prioritizes current information when users ask questions about evolving subjects. So, is it the best data or just the easiest and most convenient data?
The good news is if you keep prompting – especially if you ask discerning questions – you are more likely to eventually get a more meaningful response. But even after multiple prompts, there is no guarantee that the final response is the best response. It could be… or it might not be.
What’s the point?
Marketing agencies and organizations are quickly figuring out how to take full advantage of AI, hosting primary content on a public, fast, well-structured website, then reinforcing it with authority across trusted external properties. They know that if they want AI to associate their brand with a topic, they just publish the core explanation on their own site, support it with bylines or mentions on respected third-party sites, and make sure the content is accessible to crawlers.
And there are some pretty simple instructions to follow:
- Answer the exact questions visitors would ask.
- Use structured, scannable formatting.
- Build topical depth instead of isolated pages.
- Earn mentions and backlinks from trusted third parties.
- Keep content fresh and clearly attributed.
It’s all about becoming a credible, widely referenced source in the open web that AI systems can access.
The Implications?
Your first prompt should not be your final output. In all likelihood, thanks to the “artificial” intelligence built into current AI models and the craftiness of marketers who are figuring out how to bastardize AI responses for their own desires, you are not going to get a good answer on the first try… or maybe even on the tenth attempt.
So be careful. Be vigilant. Be skeptical. Be human. Use AI freely. Test different platforms regularly. Keep up with the constant changes in the space. And trust no one until you are certain that you can.
