Everyone loves a warm introduction and a good referral. Let’s look at the current mess state of AI recommendations.

Is AI recommending you?

94% of Americans ask AI for recommendations on which software, clothing, or ice cream to buy. They no longer ask Google first. My own mother asks ‘the Chat’ for shopping recommendations! But how do you get ChatGPT or Claude to recommend you?

Gone are the days when a company could obsess over optimizing their content for Google. There is no longer one platform making most product recommendations. Each AI platform uses a different algorithm to decide what to recommend and those algorithms are rumored to change several times an hour.

It’s too early to despair. Start by identifying what questions you want your company to be recommended for as a solution. Forget search terms…think in terms of questions. For example, ‘How do I replace my faucet?’ might be a good search question for a plumbing service or a faucet manufacturer. Creating those fancy schema for your webpages are likely a waste of time with newer AI models. Make ‘How do I replace my faucet?’ a page on your website. You get bonus points if the page is directly accessible from the main navigation.

Webpage content seems to be a very small factor (~4%) in the algorithms used by the different AI platforms. Today (August 12, 2026) the choice of recommended solution is far more heavily influenced by what people are saying about your company and product on Reddit, LinkedIn, and the various review sites such as IndieHacker to CNBC. Start by searching those sites for people complaining about you. Reddit is notorious as the place where people go to complain so start there.

Make it someone’s job to search Reddit and LinkedIn every day for posts about your brand, company or product and respond to every single one of them. Started a Reddit forum for your brand, product, or company if you have not yet already done so. This builds a crowd of people recommending you in other forums….which leads to AI recommending you.

I think of AI platforms as middle school children whispering ‘I heard that she finishes her homework in 10 minutes…how is she doing that?’ Shaping the gossip will pay back many times as people recommend you to other people…and AI recommends you to more people. It also sends a clear message to your current customers that you value them and care about their success.

Takeaway: Brand perception is the #1 factor driving AI - and human - recommendations when someone is searching for the solution that you offer. Build a trusted brand that is based on respect, honesty, and communication.

CorZen agents can build a personalized LinkedIn strategy, identify the top performing topics that your audience is searching for, and even formalize your manner of speaking so that AI-generated content sounds like you. Each agent has a different specialty but they all share one thing in common: you can download their results as a markdown (md) file, a PDF, or both.

Both formats are available because they are used in different scenarios. AI engines such as Claude, Perplexity, Kimi or ChatGPT prefer md files. Each AI platform can read an md file but won’t automatically read all of the content within a PDF file. (Most) humans prefer to read PDF files. They contain images, are formatted, and are easy to scan.

Both files contain the same information. Matching the file type to the application ensures that the content within the file is used rather than ignored.

Takeaway: For best results, match the file type to the audience who will ‘read’ it.

Same file. Each AI platform used different amounts of content. It’s best not to assume that the ‘PDF successfully imported’ message means that your content was imported for use in your next prompt.