You ask ChatGPT, “what’s the best tool for AI content optimization?” And your competitor’s name pops up. Your brand? Nowhere in sight. That stings. It’s a quiet, invisible loss happening thousands of times a day. Your team is crushing it with SEO, but in the new world of AI-driven answers, you're becoming invisible.
This isn't a future problem; it’s happening right now. And if you’re not actively checking where you stand, you’re falling behind. The good news is, you can get a clear picture of the damage in a single afternoon, with tools you already have.
To run a keyword-level AI recommendation audit, first list 10-20 high-intent keywords your customers would use. Then, systematically query each keyword across AI models like ChatGPT, Gemini, and Perplexity in a clean browser session. Document every result—whether your brand, a competitor, or nothing is mentioned—in a spreadsheet to map your visibility gaps.
What You’ll Need (No Budget Required)
Before we dive in, let’s get your toolkit ready. The beauty of this process is its simplicity. You don’t need a new subscription or a credit card. You just need a few hours and the focus to see it through.
Here’s your checklist:
A Spreadsheet: Google Sheets or Microsoft Excel. This is where you’ll build your audit log. Simple is good.
Your Top Keywords: Not your entire SEO list. We'll get to that. For now, think about the top 10-20 phrases a customer who is ready to buy might type into an AI.
Access to AI Models: You'll need accounts for the big players. Make sure you can access ChatGPT, Google’s Gemini, Perplexity, and Claude. Free tiers are perfectly fine for this.
A Clean Browser: Use Incognito or Private mode for every single query. These models personalize results, and you want the cleanest, most objective view possible.
The 4 Steps to Your Afternoon Audit
Alright, let’s get to it. Block off your calendar for a few hours. Turn off Slack notifications. This is deep work that will pay off immediately. I’ve seen teams make this way too complicated, so we’re going to keep it dead simple and actionable.
Step 1: Build Your High-Intent Keyword List
This is the single most common mistake I see. Teams grab their top 50 traffic-driving SEO keywords and call it a day. The problem? Most of those are top-of-funnel, informational queries like “what is generative AI.” It's a completely hidden traffic problem because you're looking at the wrong map. You need to audit the keywords that signal purchase intent.
Think like your customer. What would they ask an AI if they were looking for a solution like yours?
“best [your category] tools for small businesses”
“[your competitor] vs [your brand]”
“how to solve [pain point] with software”
“[your category] pricing”
Start with 15 keywords. You can always add more later, but starting small keeps you from getting overwhelmed.
Step 2: Set Up Your Audit Spreadsheet
Open up that blank spreadsheet. Don't overthink it. Create these columns:
Keyword: The query you're testing.
AI Model: ChatGPT, Gemini, Perplexity, etc.
My Brand Mentioned?: A simple Yes/No.
Competitors Mentioned?: List any competitors that show up by name.
Screenshot/Link: Grab a screenshot of the response for reference.
Notes: Was the recommendation positive? Was your brand listed first? Last? Was the context wrong?
This simple structure is all you need to start seeing powerful patterns.
Step 3: Run the Queries (The Right Way)
Open a new incognito window. Log into your first AI model, say, ChatGPT. Now, one by one, enter your keywords from the spreadsheet. For each one:
Type the query exactly as it is on your list.
Read the entire response.
Fill out the corresponding row in your spreadsheet. Add 'Yes' or 'No', list competitors, and add any quick notes.
Take a screenshot and drop it into a folder.
CLOSE THE WINDOW.
That last step is critical. Start a fresh, clean session for every single AI model to prevent one query from influencing the next. Yes, it's tedious. But it’s the only way to get reliable data. Repeat this process for all your keywords across all the AI models you want to track.
Pro-Tip: Don't just copy-paste keywords. Try phrasing them as natural questions. Instead of just “[your category] tools,” try “What are some good tools for [your category]?” This often yields more conversational, recommendation-style answers.
Step 4: Analyze the Carnage (and the Opportunity)
After a couple of hours, you'll have a spreadsheet filled with data. Now for the fun part. Start looking for patterns. Ask yourself:
Which AI model is your best friend? Which one is your enemy? You might find you're a star on Gemini but a ghost on Perplexity.
Which competitor owns the conversation? If one competitor's name appears over and over, they have successfully trained the AI models. That's your target.
Are there specific keyword themes where you’re invisible? Maybe you show up for “pricing” keywords but never for “best alternative to” keywords.
This spreadsheet is now your roadmap. Every “No” in the “My Brand Mentioned?” column is a content gap. It's a specific, targeted topic you need to create content about to start showing up. This is a far more powerful strategy than just guessing what to write next.
What to Try First
You’ve done the work, and your spreadsheet is probably a sea of red. Don't panic. This is actually a good thing. You've moved from blissful ignorance to having a concrete, actionable plan. In the zero-click era, this audit is your new baseline for survival.
Here’s what you do next:
Prioritize the Gaps: You can't fix everything at once. Focus on the high-intent keywords where your top competitor is mentioned but you are not. These are the most painful losses and the biggest opportunities.
Don't Just Write, Answer: For each gap, the goal isn't just to write a blog post with that keyword. It's to create a definitive piece of content that directly answers the user's query so well that an AI has no choice but to reference it.
Think Continuously: This manual audit is a powerful snapshot. But the AI landscape changes daily. What works today might not work next month. This is where automation becomes not a luxury, but a necessity.
Doing this audit manually is a fantastic start and will put you ahead of 90% of your competitors. But once you see the gaps, you'll realize the real work is creating and publishing the content to fill them, and then tracking it all over again. A team I know did this manually for one quarter and it nearly broke them. It highlighted the scale of their AI invisibility problem, but the manual process of fixing it wasn't sustainable.
Now that you have a map of your AI visibility gaps, the next question is how to close them efficiently. This is exactly what we built blogawesome to solve. Instead of manually tracking and writing, you can see how our platform automates the entire process, from finding the gap to shipping the content that fixes it.
Frequently Asked Questions
How often should I do this audit?
If you're doing it manually, aim for once a quarter. AI models update constantly, so a yearly check-in is not enough to stay competitive. A quarterly audit gives you a good cadence to spot major shifts in visibility without it becoming your full-time job.
What if my brand is never mentioned at all?
Don't despair! This is actually a clean slate and a huge opportunity. It means no AI has a negative perception of you; you're just invisible. Start by targeting the lowest-difficulty, highest-intent keywords from your audit and create stellar content that directly answers them.
What's the difference between this and a normal SEO audit?
A traditional SEO audit focuses on how search engines like Google rank your web pages. An AI recommendation audit focuses on whether language models like ChatGPT cite your brand as a solution in their generated answers. The two are related, but not the same. You can have great SEO and terrible AI visibility.