The 3 Rs of AI Search: #3 Brand Recommendation

To be honest, as late as the fall of 2024, I was pretty doom-and-gloomy about what AI meant for the open web.

I had long guided my clients in writing generous, helpful, expertly written content that answered the burning questions their ideal clients typed into the Google search box.

It worked. It ranked, impressed readers, and reliably brought them clients.

Then AI companies gobbled up all the text, trained their models on it, and started publishing patently derivative content. On demand. Mostly without attribution. (At best with a few measly footnote links that it was obvious that no one was going to bother to click on.)

I’m an academic. It smelled like plagiarism. I wasn’t happy about it.

But above all, I worried that going forward, SEO just wasn’t going to be a smart investment for the small business and small websites that I like working on.

---

But then, in October 2024, two things happened.

1. ChatGPT was given access to live search results. Soon after, all the other LLMs got access to search indexes as well.
2. My clients—especially the ones had regularly enjoyed leads from Google—started enthusiastically telling me they were getting leads from ChatGPT.

In case it’s not clear, the first thing caused the second thing. After October 2024, good SEO became good GEO.

The first two Rs (Brand Recognition and Brand Reputation) are mostly about damage control, ensuring that brand-aware clients aren’t turned off and turned away by what chatbots say about your brand.

But you know what’s even better?

Brand Recommendations: Getting the bots to recommend your company to people who don’t know about your company yet.

How brand recommendations happen

AI Mode’s UX prompts people to “ask anything” … so they do.

… what should I eat for dinner?

… how can I tell my in-laws to stop using screens around my toddler without it being awkward?

…. what should I buy and who should I buy it from?

Chatbot recommendations are surprisingly valuable—they work like referrals. Why? Because the instinct to anthropomorphize AI is strong. As synthetic text has become more human-like, people trust it more (1).

As people develop pseudo-relationships with their chatbot assistants, they’re more likely to follow through on a purchase decision when a chatbot recommends it.

Just like a real referral.

Are Chatbots Recommending You?

Essentially, there are two ways to find out if chatbots are recommending you.

1. If you’re on a call with a prospective client and you ask “how did you hear about us?” and they say “ChatGPT!” … well, then you’re getting recommended by the bots.
   
   And if you ever hear that … please dig in. Don’t just respond with “Cool!” and move on with the conversation. Say “Cool! In what context? What kind of conversation were you having when it recommended me?”
   
   One of the biggest problems we have in AI search is a complete lack of visibility into what’s happening in other people’s chats. Anytime you find out that a lead came to you via chatbot, you have a unique chance to get real insights into how people in your industry are using ChatGPT.
2. Now if you haven’t got your first ChatGPT lead yet, there’s still a way to check how you’re doing on the brand recommendation front.
   
   As with all the other three Rs, simply start making your own data.

Turn on a VPN, enable incognito mode in your browser, and start chatting with bots using prompts inspired by this list:

• Who is the best [industry category]?
• Who is the best [industry category] for [specific persona]?
• I’m a [persona] and I’m trying to [problem]. Which [specific category] would be the best fit for for my [problem]?
• Who are the top [industry category] for [persona] with [problem]?

Take notes. Which brands are getting recommended more frequently and why? What key industry-specific sources seem to be influencing the results? What types of information are available online about site and brands are getting recommended vs. other site and brands (including your own) that aren’t?

Encouraging more and better chatbot recommendations

AI chatbots only have two knowledge sources:

1. Their training data: Large language models learn by repetition. If your brand is mentioned consistently across the training data, the model will learn to associate your brand the words and phrases that commonly co-occur with your brand name.
2. Live search results: Right now, getting mentioned industry-specific “best XYZ” lists online is the surest path to regular recommendations. Domain authority matters far less in AI search than it does in traditional search.

What really matters now isn’t backlinks, it’s brand mentions. Brands that get mentioned most frequently (especially when they’re mentioned on lists that rank well in traditional search) are the ones that are winning the most reliably.

Now, I can hear your collective groan from all the way over here, in northern Canada. (From my igloo. As I hang out with polar bears.)

“Ugh. The ‘new SEO’ involves getting mentioned in “best of” lists online? Ugh!”

I hear you. I don’t love doing cold outreach either.

But don’t fret. There’s hope.

The silver lining? Personalization.

If you’re like me and you don’t want to do outreach, I do have one glimmer of hope to share.

If you’ve come to any of my talks lately, you know how deeply personalized AI search is.

If you’re a solopreneur career coach, you’re probably never going to get on the list for “best career coach” … but what if that’s not your goal?

What if you got super-specific by aiming not for “best career coach” recommendations but for targeted chatbot recommendations laser focused on your true ideal client?

• “best career coach for moms returning to work after a 5-year gap to raise kids”
• “best career coach for new grads navigating their options in an industry facing disruption by AI”
• “best career coach for female directors preparing to apply for c-suite positions in the next 2 years”

Now, only AI power users are putting really detailed prompts like these into the “say anything” box. But the way the machine works is that people are likely to get similar results, even if they don’t include a wordy self-introduction into the prompt.

The AI has access to your location, your search history and your chat history. Quietly, behind the scenes, it’s creating a “user embedding” that stores detailed information about each user. So even if someone doesn’t give a an explicit account of who they are and what they want in their prompt, the personalization step in how AI answers are formulated means they’re getting very personalized results … whether they realize it or not.

In traditional search, the opportunity for small brands in SEO was never the high-volume, high-competition topics (”best career coach”).

The opportunity was in getting specific about who you serve best (your ideal client) and how you solve their very specific problems. Finding and targeting the keywords that your ideal clients and only your ideal clients were putting into the Google search box.

As AI-organized results take over the Google SERPs, the results are more personalized and fragmented then ever.

Which means that Google/OpenAI can (and are) playing matchmaker with greater precision, connecting up specific people to niche providers who really are likely to be the best fit for them specifically.

It’s not winner takes all.

The small guys? They’ve got a fighting chance.

And that’s how my clients are getting leads from ChatGPT.