Zbyněk Fridrich is a freelance SEO specialist and last year’s winner of the Best SEO Project award in the Czech Republic.
With 17 years in the field, he has one hard rule for every new client: “If you want to work with me, you need Semrush.”
To position his clients for success, he analyzes how AI systems recommend brands—and where his clients are missing from those recommendations.
In this guide, I’ll share his exact workflow and what it produced for WorkLounge, a co-working brand that doubled its organic visibility and AI traffic in five months.
The Two-Phase Approach to Search Visibility
Zbyněk’s workflow covers two main steps that leverage AI sentiment and AI prompt insights from Semrush’s AI Visibility Toolkit:
- Phase 1: Get control of your sentiment—understand what AI is saying about your brand right now, and fix it.
- Phase 2: Find new opportunities—explore prompts where you should be appearing but aren’t and close the gap.
“The goal is to have maximum control over what AI says about us today. Only in the second step do we look for new content opportunities.”
Skipping phase 1 is the mistake most brands make. If AI has the wrong picture of you, mindlessly creating more content doesn’t fix the problem.
Step 1: Analyze Sentiment in Brand Performance
Zbyněk starts by opening the Brand Performance tool, adding the client’s key rivals and target location, and checking how engines like ChatGPT, Google AI, and Gemini are describing the brand.
This is critical because visibility in AI search isn’t driven by “rankings” alone. It’s determined by whether AI models consider your brand authoritative and relevant enough to recommend.
“It’s not important to be in every answer. What matters most is sentiment and overall visibility.”
First, Zbyněk examines the brand’s overall sentiment in AI answers and checks how it compares to other players in the niche.
Then, he scrolls down to the Business Drivers report and explores the specific attributes, associations, and topics that keep coming up in AI outputs.
Next, he opens the Perception report to analyze specific perceptions—positive and negative—that influence AI positioning.
Finally, Zbyněk reviews the AI-generated strategic recommendations in the toolkit—actionable suggestions based on his client’s AI visibility data.
“For WorkLounge, the initial picture was clear—and damaging. AI consistently described the space as loud. Phone booths weren’t mentioned anywhere. Member access was described as 9–5, even though members can come and go 24/7. None of this was exactly accurate, but it’s what AI had to work with based on the existing website content.”
Step 2: Rewrite Existing Website Content
From here, Zbyněk identifies the key areas of focus—perception change points—and updates the client’s website to correct whatever AI is getting wrong.
If negative or inaccurate reviews or third-party sources are shaping AI’s perception of your brand, your website content needs to directly address and counter those narratives.
“I never work on new content if I haven’t fixed the content already on the website.”
For WorkLounge, that meant going through 90 pages of product and service content and rewriting them to give AI accurate information. For example, Zbyněk covered:
- Opening hours: Structured content now clearly distinguishes 24/7 member access from 9-5 public hours
- Quiet zones: Phone booths and quiet areas—already physically there—finally documented on the site
- Membership benefits: Product pages rewritten so the member vs. non-member experience is unmistakable
Step 3: Fix Technical Issues and Implement Structured Data
Next, Zbyněk uses the Site Audit tool to identify and fix issues that prevent search engines and LLMs from correctly reading the client’s site.
These fixes include:
- Adding structured data
- Improving page structure
- Enhancing content formatting
- Resolving issues with internal links
He also experiments with LLM.txt—a file that gives AI crawlers explicit instructions about how to interpret your site’s content.
It’s still an emerging practice with no definitive proof of impact, but Zbyněk has seen positive results with it for his clients.
Step 3: Use AI Prompt Data to Plan Content
Once sentiment is in good shape, Zbyněk pulls the prompts from the Narrative Drivers tool and uses them to plan new content.
These are the actual questions people ask AI tools about topics in the client’s space.
He then downloads the data to build a comprehensive audit and reporting file for his clients.
Finally, Zbyněk picks the 20-30 most relevant prompts per project. Each prompt becomes an FAQ block on the relevant product or service page or a new content piece.
For WorkLounge, this meant building FAQ sections around questions like what the access policy is for members, what quiet work options are available, and how the space compares to traditional offices.
To sum it all up, prompts also tell him where the gaps are. If a prompt isn’t surfacing his client in a strong position, that’s either a piece of content that needs to exist or one that needs to be fixed.
Step 4: Distribute Across Channels at the Right Time
From here, Zbyněk works on pushing this new content across all channels—blog, social, newsletter, link building—timed to when search demand for that topic peaks.
“AI tools read your entire digital footprint, not just your website. The more consistently and accurately a brand appears across trusted sources, the stronger the signal.”
For WorkLounge, content about quiet zones and phone booths went out everywhere at once, including an updated Google My Business profile. Social posts reinforced the same messaging. Newsletter content tied it together.
The timing matters too.
Zbyněk aligns content pushes with seasonal demand peaks for each topic—so the content lands when people are actively searching for it, not whenever the editorial calendar happens to free up.
Step 5: Track Performance in Parallel
To track the effectiveness of his efforts and iterate, Zbyněk monitors target prompts and keywords in the Position Tracking tool.
This allows him to see the organic impact across key search surfaces—from Google AI to ChatGPT.
For WorkLounge, AI Overview visibility went from 17% to 35% over five months—directly tied to the content changes made in phases 1 and 2.
And guess what? Organic traffic and rankings followed, too.
Traffic attributed to ChatGPT has also grown nearly 20x versus the prior period and continues to trend upward.
“This whole cocktail is aimed at improving visibility in LLM systems, increasing organic traffic, getting more mentions in AI overviews, and boosting the overall business performance of the website.”
Step 6. Generate Client- and Exec-Ready Reports
Finally, Zbyněk uses Semrush’s My Reports feature to export the AI Visibility and SEO data into clean, shareable PDF reports.
These go straight into client presentations—showing sentiment scores, prompt visibility, strategic tips, and progress over time without requiring clients to log into the tool themselves.
It’s also how he gets buy-in. When clients see AI describing their brand negatively in a polished report, the case for fixing it makes itself.
The Results: What Five Months of AI Visibility Work Look Like
WorkLounge started this process in September. By January, every metric Zbyněk tracks had moved:
- Sentiment score: 67 → 82. The most fundamental shift. AI tools went from describing WorkLounge as a loud, 9-to-5 office space to accurately representing what the brand actually offers—quiet zones, 24/7 member access, and a range of workspace options.
- AI Overview visibility: 17% → 34%. The share of relevant keywords where WorkLounge appears in Google AI Overviews doubled. This is a direct consequence of the content changes—AI now has accurate, well-structured information to pull from.
- Organic traffic: up. Better AI sentiment leads to more accurate AI recommendations, which leads to more branded searches, which leads to more organic traffic. The chain is consistent across all of Zbyněk’s clients running this workflow.
“I never do something only because of AI. I treat AI visibility as part of SEO.”
The New Organic Visibility Playbook
Zbyněk’s approach isn’t complicated. He uses AI sentiment data to find what’s broken, fixes the content, then builds on top of a clean foundation. The results follow.
What makes it work is treating AI visibility and SEO as one strategy—not two separate workstreams.
Fix the narrative AI tells about your brand, and everything else improves with it: organic rankings, AI Overview appearances, and referral traffic.
That’s exactly what Semrush One is built for. It combines traditional SEO tools with AI visibility data in one connected workflow—so you can track prompt visibility and keyword rankings, run site audits for both search engines and AI crawlers, and report on all of it without switching tools.
If you want to run the same workflow Zbyněk uses, that’s where to start.
