Microtraining SearchBrand #2

Participants:

Hussam Sufan (SearchBrand.ai)

Gonzalo Aburto (Metrix Agency)

Alejandro Momberg (Metrix Agency)

🤝 Collaboration SearchBrand.ai + Metrix Agency

Date: October 28, 2025

Duration: ~60 minutes

TL;DR — The Essentials

  • AI traffic (ChatGPT, Gemini, Perplexity) has conversion rates 20-40% higher than site average

  • Although the volume is marginal today (0.01-1%), it is growing exponentially and spikes during commercial events

  • You must measure new metrics: Share of Voice, LLM mentions, and contextual visibility

  • You can track this traffic in GA4 by creating a custom channel group with a regular expression

  • Traditional SEO and AI SEO are complementary, not contradictory

  • Most effective content: clear sections, structured FAQs, self-contained snippets, complete thematic relevance

  • 80% of AI traffic comes from ChatGPT, followed by Perplexity and Gemini

The evolution of traffic: from clicks to zero-click

We are experiencing a fundamental shift in how people consume information online. What used to be a comparison journey with multiple tabs open (search for a hotel → compare 5-7 options → check reviews → compare prices → decide) is now compressing dramatically.

The new behavior: The user asks a specific question to ChatGPT or Gemini ("beachfront hotel in Cancun with breakfast included and pet-friendly") and the model takes them directly to the ideal product page, skipping the entire exploratory funnel.

2005-2015: Classic SEO

Optimization based on keywords, backlinks, and meta tags. Google only showed text results.

2016-2019: Semantic Search

Google evolves to understand meanings, not just words. RankBrain and intention comprehension appear.

2020-2022: Conversational Era

Voice searches and more elaborate queries. Short keywords ("running shoes") become full questions ("what are the best running shoes for asphalt at night?").

2023: Generative AI Explodes

ChatGPT and other LLMs generate no-click answers. The era of "zero-click" traffic begins, where people resolve queries directly in the model without visiting websites.

2024: Search Generative Experience (SGE)

Google launches AI Overviews. SEO professionals notice drops in CTR: impressions keep rising but clicks fall. AI takes the top spot in informational searches.

2025: Consolidation

AI traffic stops being "experimental" and becomes constant. Companies begin to measure Share of Voice, LLM mentions, and create specific AEO (Answer Engine Optimization) strategies.

The problem we face today: Clients ask, "Why is my CTR falling if my impressions keep rising?" The answer: AI is taking the top spot in informational searches, and users resolve their queries without clicking.

The new metrics you should measure

Traditional metrics (sessions, users, conversions) are still important, but they no longer tell the whole story. With the emergence of LLMs, new critical metrics appear:

1. Share of Voice

Of the total answers that ChatGPT, Gemini, or Perplexity generate about your industry, what percentage mentions your brand?

Practical example: If in your industry there are 1,000 monthly queries related to your product category, and your brand appears in 150 of those answers, your Share of Voice is 15%.

Why it matters: This metric is like keyword search volume in traditional SEO, but applied to AI mentions. It doesn't generate direct traffic, but it builds brand presence and "top of mind" in the models.

2. LLM Mentions

How many times do models cite you? In what context? What are they saying about you?

This goes beyond counting mentions: you need to understand if these mentions are:

  • Positive: "X is the best HR software for SMEs"

  • Neutral: "X is an available option in the market"

  • Negative: "X has had problems with..." (critical reputational issue)

Real reputational risk case: A toothpaste brand appeared on ChatGPT with false information ("X causes cancer"). The company had all the certifications and safety tests, but could not control what the model said. This is a reputational risk that did not exist before.

3. Search volume on AI platforms

Of the total searches about your industry, how many are done on ChatGPT vs Google vs Perplexity?

Similar to traditional keyword research, but applied to each platform. This helps you prioritize where to focus optimization efforts.

4. Contextual visibility

This is the most difficult metric to quantify but extremely valuable: how is AI positioning your brand in the broader context of your industry?

It does not generate direct traffic, but it does build authority. It's like being top of mind for consumers before they reach your site.

Change of mindset: We no longer just measure what happens on our website (lower funnel, conversion). Now we must also measure what happens OUTSIDE the site, on platforms where brand perception is built (upper/middle funnel).

Available tools in the market

Metrix Agency explored multiple options before deciding on its tools stack to measure AI:

Tool What it offers Advantages Limitations SEMrush Traditional SEO + basic AI module Mature tool, integration with other metrics Slow update for AI metrics, does not cover all needs SERanking Traditional SEO with attempts to incorporate AI Competitive pricing Very limited AI functionality HubSpot AI module within its marketing suite Integration with CRM and automation Requires full HubSpot ecosystem SearchBrand.ai Specialized in measurement and optimization for LLMs Identifies content gaps, competitive analysis, automatic content generation Focused only on AI (does not replace traditional SEO tools) Google Analytics 4 Traffic and conversion measurement Free, first-party data Only measures traffic that reaches the site, not contextual visibility

Why Metrix chose SearchBrand.ai

Gonzalo Aburto (G2 Digital and Media at Metrix) explains three key reasons:

1. Identification of content opportunities:

"The tool identifies questions that the audience asks about our products/services where we are not yet positioned. This gives us a clear roadmap of what content to create."

2. Competitive analysis in AI:

"We can see where we appear vs. the competition in AI answers. If my competitor is on ChatGPT for a certain search and I am not, I know exactly where I have a gap."

3. Automated generation of optimized content:

"Before, we used SEO Writer and other external tools. Now we can generate content directly on the platform, optimized for AEO. It's faster, more efficient, and reduces operational costs for the team."

How to structure content to appear in AI

Based on over 100 active SEO service clients (30 specifically analyzing AI), Metrix identifies these patterns of effective content:

1. Clear and short sections (one topic per block)

LLMs do not take the complete content of your site. They extract specific paragraphs and mix them with paragraphs from other sites. This is why every section must be self-contained.

Bad example:

Our services include consulting, implementation, and support. We do all this with cutting-edge technology and an expert team. Contact us for more information.

Good example:

## Generative AI Consulting We help companies identify generative AI use cases applicable to their business. Includes: process analysis, opportunity identification, and implementation roadmap. ## Implementation of AI Solutions We develop custom solutions with generative AI as a core component. Technologies: OpenAI GPT-4, Claude, Gemini. Average time: 6-12 weeks.

2. Complete thematic relevance (being a reference in a TOPIC, not just a keyword)

In traditional SEO, you could rank for "human resources software" with a single well-optimized page. In AEO, you need to demonstrate complete thematic authority.

Example: If you sell life insurance, you need content about:

  • What is life insurance?

  • How to choose the best life insurance?

  • How to read a life insurance policy?

  • How to quote life insurance?

  • Differences between types of life insurance

  • When to get life insurance?

  • Comparison: life insurance vs health insurance

Why it works: LLMs evaluate your "domain authority" for a topic based on comprehensive coverage. If you only have one product page, you compete with sites that have 20-30 articles on the subject.

3. Structured FAQs and lists

LLMs generate answers in bullet point and list formats. If your content is already structured this way, it is easier to extract and cite.

Proven fact: Alejandro Momberg (SEO Manager at Metrix) reports that "a large part of our citations in AI Overviews come from structured frequently asked questions with FAQPage schema markup."

4. Self-contained snippets

Don't make the reader (or the LLM) need to read 4-5 paragraphs to understand an idea.

Each paragraph must work independently.

Bad example:

As we mentioned earlier, this relates to what we said in the previous section about implementation. All of this connects with our unique approach...

Good example:

The implementation of generative AI in companies takes between 6-12 weeks. This timeline includes: requirements analysis (2 weeks), development (4-8 weeks), and testing (2 weeks).

Real cases: Data from the trenches

Metrix Agency shares real data from two clients to illustrate the current behavior of AI traffic:

Case 1: Home appliance e-commerce (Cyber Monday 2025)

Context: Tableware and kitchen appliance client during Cyber (October 27-29, 2025).

10x peak

Increase vs average

20

Purchases in 6 months

3.1%

AI conversion rate

2.2%

Average site conversion

Key insights:

  • The volume of users from AI is marginal on normal days, but spikes 10x during commercial events

  • Of 40,000 transactions in 12 months, only 20 came from AI (0.05% of the total)

  • BUT: The AI conversion rate is 3.1% vs 2.2% average = almost 1 percentage point improvement

  • This means that although the volume is small, the quality of traffic is superior

Conclusion: Users arriving from AI are closer to the purchase decision. The model takes them directly to the product page they need, shortening the funnel.

Case 2: Health retail in Chile

Context: Health service client with online scheduling (last 12 months).

833

Appointments from AI in 12 months

~800K

Total conversions/month

26%

AI conversion rate

21%

Average site conversion

Key insights:

  • 833 appointments in 12 months = ~70 per month from AI channels

  • The AI conversion rate is 5 percentage points higher than the average (26% vs 21%)

  • Sustained month-over-month growth, with peaks on specific dates

  • The volume is still small, but the trend is exponential

Conclusion: We can scale conversion volumes with less traffic volume. AI traffic converts better because users have already been "pre-qualified" by the model.

Traffic distribution by source

Based on analysis of multiple clients, Metrix identifies this typical distribution:

Source % of AI traffic Note ChatGPT / OpenAI ~80% Absolute market dominance Perplexity ~10-15% Second place, growing rapidly Gemini ~5-8% Slow but steady growth Claude / Copilot / Others ~2-5% Marginal, but worth tracking

Practical implication: If you have limited resources, prioritize optimization for ChatGPT. It represents 80% of the traffic. But don't completely ignore Perplexity and Gemini.

Tutorial: How to track AI traffic in Google Analytics 4

Currently, when you check your acquisition reports in GA4, AI traffic appears scattered across different channels: sometimes in "Referral", sometimes in "Unassigned", sometimes mixed with others.

The solution: Create a custom channel grouping that unifies all LLM traffic.

Step 1: Access channel group settings

In your Google Analytics 4 property:

  1. Click on the Admin icon (gear) in the bottom left

  2. In the "Data display" column, select Channel Groups

  3. Locate your default channel group

Step 2: Create a new group based on the default one

Instead of creating a group from scratch, we will copy the existing one:

  1. Click on the three dots next to the default group

  2. Select "Create new channel group"

  3. This will copy the entire existing structure (organic search, paid, direct, etc.)

Why copy instead of create new? This way you keep all the traditional channels and only add the AI channel. It allows you to compare AI vs Organic vs Paid in the same report.

Step 3: Add the AI channel with a regular expression

Now you will add a new channel that captures all LLM traffic:

  1. Name the group: "AI Channel Group" (or whatever name you prefer)

  2. Click on "Add new channel"

  3. Channel name: "AI" or "LLMs"

  4. Configure the condition:

    • Dimension: "Source"

    • Operator: "matches regex"

    • Value: (copy this exactly)

chatgpt|openai|perplexity|copilot|gemini|claude

What does this regular expression do?

  • The | symbol means "or" (logical OR)

  • It will capture any session where the source includes any of these words

  • It is case-insensitive

  • You can add more sources by separating them with |

Step 4: Save and apply the new group

  1. Click on "Save"

  2. The new group will be available in your reports

  3. Go to Reports > Acquisition > Traffic acquisition

  4. At the top, change the dropdown from "Default channel group" to your "AI Channel Group"

Step 5: Analyze performance

Now you will see the "AI" channel alongside:

  • Organic Search

  • Paid Search

  • Direct

  • Referral

  • Social

  • Etc.

Metrics you can analyze:

  • Sessions and users

  • Conversion rate

  • Completed events

  • Revenue (if you have e-commerce configured)

  • Percentage of AI channel participation vs other channels

  • Average session duration

  • Pages per session

Step 6 (Optional): Advanced segmentation

If you want to see the breakdown by each individual LLM:

  1. Go to "Explore" in the GA4 sidebar menu

  2. Create a new free-form exploration analysis

  3. Add dimension: "Source"

  4. Filter by your regular expression or manually select: chatgpt, perplexity, gemini, etc.

  5. Compare metrics between each source

Typical insight: You will find that ChatGPT represents ~80% of AI traffic, followed by Perplexity (~15%) and Gemini (~5%). This helps you prioritize optimizations.

Want to automate all this?

SearchBrand.ai is launching direct integration with GA4. Instead of manually configuring channel groups, the platform will automatically connect and show you AI traffic along with contextual visibility metrics.

Join the waitlist →

Frequently asked questions

How do I track ChatGPT, Gemini, and Perplexity traffic in Google Analytics 4?

You need to create a custom channel group in GA4 using a regular expression that identifies AI sources. The process is: 1) Go to Admin > Data display > Channel Groups, 2) Create a new group based on the default, 3) Add a channel named "AI" or "LLMs", 4) Configure the condition "source matches regex" and include: chatgpt|openai|perplexity|copilot|gemini|claude. This will group all traffic from these sources into a single measurable channel.

What is Share of Voice in the context of AI and LLMs?

Share of Voice in AI measures the estimated percentage of times your brand appears in the answers of ChatGPT, Gemini, Perplexity, and other LLMs compared to the total searches related to your industry. It is similar to traditional keyword research, but applied to mentions in AI-generated responses. This metric does not directly generate traffic, but it indicates how present your brand is in the "top of mind" that these models build.

What is the typical conversion rate of traffic coming from AI compared to other channels?

Actual data from Metrix Agency shows that AI traffic has conversion rates 20-40% higher than other channels. In an e-commerce for home appliances, the AI conversion was 3.1% vs 2.2% site average (almost 1 percentage point improvement). In health retail, it was 26% vs 21% average (5 percentage points higher). This happens because LLMs lead users directly to product or conversion pages, shortening the exploratory funnel. Users arriving from AI are closer to making a purchase decision.

What new metrics should I measure with the advent of LLMs?

In addition to traditional metrics (sessions, users, conversions), you should measure: 1) Share of Voice: percentage of your brand's presence in AI responses, 2) LLM Mentions: how many times models cite you and in what context (positive/neutral/negative), 3) AI search volume: of the total queries in your industry, how many are made in these channels, 4) Contextual visibility: off-site positioning that doesn't generate direct traffic but builds brand authority. These metrics complement, not replace, traditional ones.

Is AI traffic already significant or is it still marginal?

In absolute volume, it is still marginal (typically 0.01-1% of total traffic), but it shows three critical trends: 1) Sustained and exponential growth month over month, 2) Dramatic peaks in commercial events (Cyber, Black Friday) where it can multiply 5-10x, 3) Superior conversion rate (20-40% better than site average). Although the volume is small today, the trend indicates that it will be a larger channel in 6-12 months. Don't wait until it's 10% of your traffic to start measuring it.

Should I do traditional SEO or AI SEO? Are they contradictory?

They are complementary, not contradictory. Traditional SEO is still the base because LLMs currently extract information from Google and Bing search results (via SERP API). If you are not indexed in traditional search engines, it is very difficult to appear in AI. Best practices for technical SEO (structure, indexing, crawling, semantic HTML, structured data, FAQs) benefit both traditional and AI positioning. The difference lies in also optimizing for extraction and citation, not just for clicks. Alejandro Momberg of Metrix summarizes it: "They are the same technical bases, but with 20% differences in execution."

What type of content works best to appear in AI responses?

The most effective content has: 1) Clear and short sections with one topic per block, 2) Complete thematic relevance (being a reference in an entire topic, not just a keyword), 3) Structured FAQs and lists (LLMs generate answers in bullet point format), 4) Self-contained snippets (that don't require reading 4-5 paragraphs to understand the idea), 5) Structured data (FAQ, product, reviews schema markup), 6) Tables and comparisons, 7) Content focused on a specific niche to increase thematic authority. Metrix reports that a large part of their citations in AI Overviews come from well-structured FAQs.

Should I be concerned about what is said about my brand on social media and forums for AI purposes?

Yes, absolutely. LLMs cite sources beyond your website. Reddit, for example, represents a significant portion of citations. This has two implications: 1) Opportunity: Authentically participating in relevant forums can increase your visibility in AI, 2) Reputational risk: Negative or false information in these spaces can be cited by models. Brands need to monitor not only what they say on their site, but what is said about them across the web. This includes Instagram, Reddit, TikTok, specialized forums, and media outlets.

What length should my content be to optimize for AI?

The recommended minimum is 1,000 words, but more important than length is semantic structure. Use H2s for main sections, short paragraphs (3-4 lines maximum), and ensure each block is self-contained. LLMs go through a process called "chunking" where they divide content into chunks. If your content is already well-structured, you facilitate this process. Think about writing so that a machine can extract, reuse, and display your content as an independent answer.

Can I use SearchBrand.ai with my traditional SEO tools (SEMrush, Ahrefs, etc.)?

Yes, and it's the recommended setup. SearchBrand.ai does not replace traditional SEO tools; it complements them. Metrix Agency uses SEMrush for traditional SEO + SearchBrand.ai for AI metrics. The advantage of SearchBrand is that it is 100% specialized in AI: it identifies content gaps for LLMs, analyzes competitors in AI responses, generates optimized content for AEO, and will soon integrate with GA4 to automatically display AI traffic.

Final recommendations from Metrix Agency

Based on 14 years of experience in digital marketing and managing over 100 SEO clients, Gonzalo Aburto and Alejandro Momberg share these final recommendations:

1. Measure now, even if the volume is small

Gonzalo Aburto: "The volume is still marginal today, but it's growing exponentially. If you wait until it's 10% of your traffic to start measuring, you'll be late. Implement the channel group in GA4 TODAY and start building your baseline."

2. Let data guide your strategy

Alejandro Momberg: "We are in an exploratory phase. What works for one brand may not work for another. Be meticulous with measurement: create reports, identify insights, let the data tell you what's working and what isn't."

3. Don't abandon traditional SEO

Traditional SEO remains critical because:

  • LLMs extract information from Google/Bing results

  • If you are not traditionally indexed, it is very difficult to appear in AI

  • The technical bases are the same (80% overlap)

  • Optimizing for AI also improves your traditional SEO

4. Focus on niches and thematic authority

Gonzalo Aburto: "Don't try to talk about too many things at once. Define your niche, build comprehensive content on that topic. Thematic authority is your competitive advantage in AI."

5. Experiment and share learnings

The AEO/AI SEO community is under construction. What works today can change in 3 months. The invitation is:

  • Experiment with different types of content

  • Rigorously measure the results

  • Share findings with the community

  • Stay updated with model changes

Immediate action checklist

  • Create an AI channel group in GA4 with the provided regular expression

  • Establish a baseline: how much AI traffic you receive today

  • Register on SearchBrand.ai to measure Share of Voice and mentions

  • Audit existing content: does it have clear sections, FAQs, self-contained snippets?

  • Identify your main niche and topics where you want authority

  • Create a content calendar focused on comprehensive thematic coverage

  • Implement FAQ schema markup on key pages

  • Monitor what is said about your brand on Reddit, forums, and social media

  • Set up monthly alerts to review AI traffic evolution

  • Compare conversion rate: AI vs other channels

The near future: What's coming

During the microtraining, participants discussed emerging trends worth keeping an eye on:

AI integration with commerce

Hussam Sufan mentions: "ChatGPT now integrates with Booking, Expedia, Spotify. I searched for hotels in Pichilemu and it showed me 5 out of 165 options. How does this impact the 160 hotels that DID NOT appear?"

Implication: LLMs are becoming commerce platforms. Being in the "top 5" answers will be critical.

AI voice assistants

Soon you'll say: "Alexa, I'm out of coffee" and the assistant will automatically buy you coffee. Which coffee will it choose? The one that is best positioned in its recommendation model.

Sentiment analysis in mentions

The natural evolution of tools will be to incorporate sentiment analysis: not just counting mentions, but classifying them as positive, neutral, or negative. This will be critical for reputation management.

Direct GA4 integration

SearchBrand.ai is developing direct integration with Google Analytics 4. Instead of manually configuring channel groups, the platform will automatically connect and show AI metrics along with contextual visibility.

Hussam Sufan: "We are in a privileged moment. We are pioneers in a wave that is forming. From here to 12 months, this will be a necessity, not an option. Let's act now while there is still a competitive advantage."

Conclusion: Let's build this together

This microtraining reflects the spirit of the emerging AEO community: we are all learning together.

There are no perfect recipes. There are best practices, experiments, real data, and a lot of uncertainty. But there is also absolute certainty: AI traffic is here to stay.

The numbers confirm it:

  • Exponential growth month over month

  • Conversion rates superior to traditional channels

  • Dramatic peaks in commercial events

  • Changing consumer behavior (60% already use AI for purchases)

The invitation from Metrix Agency and SearchBrand.ai is clear:

1. Start measuring TODAY. Implement the channel group in GA4, record your baseline, observe the evolution.

2. Experiment and document. Test different content types, measure results, identify what works for YOUR brand.

3. Share learnings. Join the community, participate in future webinars, contribute your findings.

4. Keep traditional SEO strong. It is the foundation upon which everything else is built.

The future is not AI or traditional SEO. It is AI and traditional SEO working together to build a holistic brand presence.

Join the SearchBrand community

Access future microtrainings, share best practices, receive updates on model changes, and connect with other professionals navigating this shift.

Join for free →

About the participants:

Hussam Sufan is the founder of SearchBrand.ai, a platform specialized in measurement and optimization for LLMs. With over 15 years of experience in growth marketing, he leads the construction of tools for the new era of AI positioning.

Gonzalo Aburto is G2 Digital and Media at Metrix Agency. With 14 years in digital marketing, he leads the agency's SEO strategy and its adaptation to the era of LLMs. He manages over 100 active SEO service clients.

Alejandro Momberg is SEO Manager at Metrix Agency, a specialist in technical SEO. He has dedicated the last few months to experimenting and implementing AEO strategies for dozens of clients, documenting results and building best practices.

About Metrix Agency:

Metrix is a digital agency with over 100 active clients in SEO services and digital marketing. They are at the forefront of adapting traditional SEO to the era of LLMs, experimenting, measuring, and sharing learnings with the community.

About SearchBrand.ai:

SearchBrand.ai is a platform specialized in measurement and optimization for AI answer engines (ChatGPT, Gemini, Perplexity, Claude). It allows measuring Share of Voice, identifying content gaps, analyzing competitors, and will soon integrate directly with GA4.

© 2025 SearchBrand.ai — Microtraining #2: Measuring AI traffic in GA4

Last updated: October 28, 2025. LLM behavior and AEO best practices evolve rapidly. Review this content every 60-90 days for updates.