Measure what your content says about your business once it’s indexed… and whether AI cites it. The likes and impressions from the first 48 hours in the feed stopped being the number that matters.
That’s the short version of a conversation I had with John Mulligan of ActionCoach on his podcast. He asked the question every business owner asks eventually: “What should I actually be looking at?” The answer has changed more in the last year than in the ten before it.
The metric moved out of the feed
AI and search are compressing into one thing. What you write on your website, what you post on social, what shows up in your newsletter… all of it gets indexed, and AI pulls from all of it when someone asks a question your business could answer.
So the question is no longer “how did this post do?” The question is “when this post gets indexed, what does it say about my company?” Does it match the website? Does it sound like the same business? Does it answer anything a real client would ask?
A post that gets 40 likes and says nothing specific is doing less for you than a post that gets four likes and clearly explains who you help and how. The second one gets cited. The first one evaporates.
AI is talking about you behind your back
Search used to be a list of links. Now it’s a conversation, and the conversation has context.
Think about how much your own AI tool knows about you. Mine has my health notes, my recipes, my client list, my calendar. When I ask it a question, it answers with all of that in mind. Your prospective clients are doing the same thing. When they ask “who should I hire for this?”, the AI already knows their industry, their budget, their frustrations… and it answers accordingly.
The part that surprises people most: the AI may recommend you, describe you, and compare you to someone else without that person ever clicking over to your website or your social media. There are conversations happening about your business that you will never see.
That’s why consistency matters more than volume now. If your website says one thing, your LinkedIn says another, and your Instagram captions sound like a different company altogether, the AI has nothing solid to repeat. It will go find someone it can describe clearly.
The last-five-posts test
Pull up your last five posts and read them in a row. Do they sound like anyone else in your industry? Could a competitor have posted them word for word?
This is the first thing I look at with a new client, and it comes up constantly. The posts are fine. They’re grammatically clean, on topic, posted on schedule. And they’re interchangeable with everything else in the feed, because the same AI tools wrote them from the same generic prompts.
The feed algorithms are AI too. They can tell when content is generic, and they treat it accordingly. So when a business owner tells me reach is sliding, calls are down, and even impressions are softer than they were, the first question is how the content is being written… and whether it still sounds like the business it came from.
Keep using AI. Give it something specific to work with: who you help, what you’ve seen, what you actually think. That’s the difference between content that gets indexed as “a marketing company” and content that gets indexed as your company.
How to tell if AI is sending people your way
You can see it in your website analytics. When someone clicks over from ChatGPT, Perplexity, Gemini or Claude, it shows up as a referral source. If those visits are appearing, two things are true: the AI recommended you, and the person liked what it said enough to go look for themselves.
That second part is the one I care about. The way clients actually arrive has shifted. Someone hears about you (from a referral, from a podcast, from an AI answer), then they go research you. They read what you’ve been saying online. They decide whether you sound like you know what you’re talking about. Then they call. Social media’s job in that sequence is credibility, not discovery.
Social platforms don’t yet show you AI-driven clicks the way websites do, so the website is where to watch. If the number is zero, that’s useful information too. It usually means the content isn’t giving the AI anything clear enough to recommend.
For clients who want the full picture, I run an AI audit: what’s indexed, what it’s saying, and where the website and the social accounts disagree with each other. Most businesses have never looked.
Direct it. Don’t let it direct you.
Ask any AI tool “what’s the best social media strategy for my business?” and it will answer with complete confidence. The problem is where that answer comes from: everything published online up to a year or so ago, averaged together, with no idea what your business actually needs.
It will tell you with great certainty that this is the way. It is usually describing what worked for someone else, last year.
AI is good when you direct it. You bring the client knowledge, the industry patterns, the opinion. It brings the speed. Flip that around and you get content that sounds like everyone, measured by numbers that no longer mean anything.
So the metric that matters now is a simple one to say and a harder one to earn: when AI describes your business to someone who’s never heard of you, is it right… and is it worth repeating?
____
Frequently Asked Questions
Do likes and impressions matter at all anymore
They tell you how a post performed in the feed for a day or two. They don’t tell you what that post says about your company once it’s indexed, which is what AI uses. Treat them as a signal, not the scorecard.
How do I know if AI is recommending my business?
Check your website analytics for referral traffic from AI tools (ChatGPT, Perplexity, Gemini, Claude). Visits from those sources mean you were recommended and the person wanted a closer look.
Should I stop using AI to write my content?
No. Give it more to work with. Your client patterns, your opinions, and your actual voice are what keep AI-written content from sounding like everyone else’s.
What if my website and my social media say different things?
That’s the first thing to fix. AI pulls from all of it, and inconsistency gives it nothing solid to repeat. An AI audit will show you where the gaps are
