SEO Competitor Analysis in the Age of AI Overviews: What Should You Measure?

If you have been doing competitor analysis the same way since 2022checking organic keywords, comparing backlinks, taking SERP screenshots, and making reportsyou may be missing how search has changed.
AI Overviews, ChatGPT Search, Perplexity, and Gemini are now influencing many searches. They do not work exactly like traditional Google search. A competitor may not rank first in Google but can still get the click, appear in an AI answer, or get mentioned as a source.
This means that a #1 Google ranking does not always bring the same amount of traffic it used to.
So, what should you measure now?
This post explains the important things to track and gives you a simple workflow for doing some of it automatically. Checking ChatGPT manually every Monday may work for a few keywords, but it becomes difficult when you have hundreds of keywords to monitor.

Why your old dashboard is lying to you
Three important SEO metrics have changed:
Impressions - When an AI Overview answers a search directly, many people may not scroll down to the normal search results. Your impressions may stay the same or even increase, but that does not mean more people are interested in clicking your website. Comparing this month's impressions with last year's numbers may not give you a fair picture because search results work differently now.
*CTR *- In the past, ranking #1 usually meant getting a good number of clicks. Now, an AI answer can appear above the normal search results and give users the information they need without visiting a website. Research from Ahrefs has found that organic CTR for some position-one informational searches can drop by up to 58% when an AI Overview appears. So, if your CTR reports still use old benchmarks from 2023, they may not show the real situation.

Ranking #1 means winning. - This is no longer always true. Ranking high on Google and being mentioned in an AI answer are now two different things. You can rank above a competitor in the normal search results while that competitor gets mentioned, linked, or used as a source in the AI answer.
The opposite can also happen - A smaller website that does not rank on page one can still get mentioned by AI tools if its content is clear, useful, and easy for AI systems to understand.

This does not mean that traditional SEO metrics are useless. They are still important. But you need to look at them differently and add something new to your reports: AI visibility.

The metric stack that actually reflects reality in 2026
1. AI Citation Share
Instead of only asking, "Do we rank?", also ask, "How often does AI mention us?"
For each important search query, check whether an AI Overview appears. If it does, record which websites are mentioned or cited in the answer.
Do this for your 10–20 most important topics, not only for keywords that directly make money. Many AI Overviews appear for informational searches, which are often used by people at the beginning of the buying journey.
Track each competitor as a percentage.
For example:
Competitor A: cited in 8 out of 10 AI answers = 80% citation share
A competitor that appears in 8 out of 10 AI answers has a level of authority that a normal rank-tracking tool may not show.

2. Authority Outside Your Website
AI systems often look at information from different sources before deciding what to trust.
A brand that is mentioned on Reddit, Quora, review websites, and industry forums can build authority outside its own website. Traditional SEO tools such as Ahrefs and Semrush may not show all of these signals clearly.
A simple way to check this is to search for your competitor's brand name along with relevant industry terms on Reddit and Quora.
Look for real discussions and genuine mentions, rather than posts created by the company itself.
If your competitor is regularly being discussed by other people while your brand has almost no third-party mentions, that is an important gap. On-page SEO alone may not solve it.

3. Look at Content Structure, Not Just Schema
There has been a lot of discussion about whether structured data, such as JSON-LD, helps websites get cited by AI.
However, research from Ahrefs found that adding schema did not create a measurable increase in AI citations. Websites using schema often had better content, more backlinks, and stronger technical SEO, which may have been the real reason they performed well.
When checking a competitor's pages that are being cited by AI, focus on the content itself.
Look for:
Specific facts, numbers, and data that can easily be quoted
Clear definitions that answer a question on their own
Original research or clearly explained methods
Content that answers one question directly instead of hiding the answer inside a very long article
The main question should be: "What makes this page easy for AI to understand and use?"
Don't focus only on the competitor's schema or technical markup.

4. Branded and High-Intent Keywords
AI answers are taking away some of the attention from informational searches. Because of this, it is becoming more important to track searches that are closer to revenue.
Look at:
Changes in branded search volume for your business compared with competitors
Rankings and traffic for commercial and transactional keywords
Organic leads and conversions instead of only clicks and impressions
This helps remove some of the noise from searches that were unlikely to become customers anyway.
For example, 10,000 informational visits may look impressive, but 100 visits from people searching for a service and ready to buy can be much more valuable.

5. How Often Your Keywords Trigger AI Overviews
Look at the keywords that both you and your competitors are targeting.
Then check how many of those searches show an AI Overview.
This tells you how much the search landscape has changed for your target topics.
For example:
High AI Overview rate: AI citation visibility becomes more important.
Low AI Overview rate: Traditional Google rankings are still more important.
This helps you decide where to focus your competitor analysis instead of treating every keyword in the same way.

A lightweight way to track this without buying a platform
Tools like Semrush's AI visibility features, Rank Prompt, and SE Ranking can automate this process. But if you don't want to pay for another tool, you can also build a simple system yourself.
The basic process is simple:
Create a fixed list of prompts.
Run those prompts through each AI search platform on a regular schedule.
Record which websites are mentioned or cited.
Compare the results with your previous checks.
You can automate much of this with a simple script.
`# Pseudocode — replace this with the API client

you use for each AI platform.

import json
from datetime import date

PROMPTS = [
"best {category} tools for {use_case}",
"how to {task} without {common_pain_point}",
# Add your 15–30 main prompts here.
]

ENGINES = ["google_ai_overview", "chatgpt", "perplexity"]

def check_citations(prompt: str, engine: str) -> list[str]:
"""Return the domains cited for this prompt."""
# Connect to the relevant API or data source here.
...

def run_audit(prompts, engines, competitors: set[str]):
results = {"date": str(date.today()), "prompts": {}}

for prompt in prompts:
    results["prompts"][prompt] = {}

    for engine in engines:
        cited = check_citations(prompt, engine)

        results["prompts"][prompt][engine] = {
            "cited_domains": cited,
            "competitors_present": [
                c for c in competitors if c in cited
            ],
            "self_present": "yourdomain.com" in cited,
        }

return results

if name == "main":
competitors = {
"competitor-a.com",
"competitor-b.com"
}

data = run_audit(PROMPTS, ENGINES, competitors)

with open(
    f"ai_visibility_{date.today()}.json", "w"
) as f:
    json.dump(data, f, indent=2)`

**
A Few Things to Keep in Mind**
Start with a spreadsheet. For 15–20 prompts, a spreadsheet is enough. If you start tracking hundreds of prompts every week across several AI platforms, automation will save a lot of time.
Keep your old results. Don't only look at today's data. Save the results from every check so you can see which competitors are gaining or losing AI citations over time.
Follow platform rules. Not every AI search platform provides a public API. Before automating anything, check its terms and use approved APIs or third-party data providers when available. In some cases, manual checks may be the safer option.
The goal is not to build a complicated system. You simply want to track who gets mentioned, how often they get mentioned, and how that changes over time.

Old metric vs. new metric, side by side

Putting it together
A competitor analysis that's actually useful in 2026 has two tracks running in parallel:
Classic SEO tracks keyword gaps, backlinks, technical audit, content gaps. Still valid, just re-segmented so you're not drawing conclusions from AI-distorted impressions and CTR numbers.
AI visibility track citation share per prompt cluster, third-party entity mentions, and structural analysis of what's getting quoted and why.
The teams pulling ahead right now aren't the ones with the fanciest dashboard, they're the ones who noticed the second track exists at all. Most competitor audits still stop at classic keyword gaps, which means the AI visibility layer is wide open if you start measuring it before your competitors do.
If you're building this out, start small: pick 15 prompts that matter to your funnel, run them by hand across Google, ChatGPT, and Perplexity once, and just look at who shows up. You'll know within an afternoon whether this is a gap worth automating for your niche.