TL;DR: Key takeaways
- Direct referral traffic from ChatGPT, Gemini, and Google AI Overviews is still small – but a growing share of buyers mention AI Search Engines as their primary research tool.
- To get cited inside AI-generated answers, B2B SaaS brands need high-authority mentions on LinkedIn and in traditional PR, not just classic backlinks.
- Schema markup and technical SEO are the price of entry: if LLMs can’t parse your pages cleanly, you won’t be cited.
- Attribution is shifting from referrer headers to self-reported sources (“How did you hear about us?”). Track it deliberately.
- The biggest risk in 2026 isn’t AI itself – it’s “lazy AI content.” Brands that keep a human touch will outperform brands that don’t, on both Google and LLMs.
The problem: your buyers research in ChatGPT, but your analytics can’t see it
A buyer types “best influencer marketing platform for European brands” into ChatGPT. The model returns a confident, paragraph-length answer with three named vendors and a few cited domains. The buyer clicks one of them.
In your analytics, that click shows up as direct traffic — or it doesn’t show up at all.
This is the gap every B2B SaaS marketing team is now trying to close. Eugen Knippel, VP Global Voice at Kolsquare, put it plainly in our interview: referral traffic from AI engines remains a tiny slice of acquisition, but the share of buyers naming AI as their starting point keeps climbing. The pipeline is real. The measurement is broken.
So the question isn’t whether to optimize for AI search. It’s how – when the channel is mostly invisible and the playbook is still being written.
What is Generative Engine Optimization (GEO)?
Generative Engine Optimization (GEO) is the practice of making a brand’s content discoverable, parseable, and citable by AI engines like ChatGPT, Perplexity, Google AI Overviews, Google AI Mode, Gemini, and Microsoft Copilot.
GEO overlaps with classic SEO — both reward authoritative, well-structured content — but the signals diverge in three important ways:
- The “ranking” surface is a generated answer, not a list of blue links. You either get named or you don’t.
- Citations come from a wider pool of sources than Google’s index, including LinkedIn posts, press coverage, niche forums, and structured data.
- Click-through is decoupled from visibility. Your brand can be the answer without anyone visiting your site.
Kolsquare’s approach to GEO is a useful blueprint because it treats AI search as a citation problem, not a traffic problem.
The three layers of Kolsquare’s authority playbook
1. Earn citations where LLMs actually look — LinkedIn, Influencers and PR
Classic link-building optimizes for Google’s PageRank. GEO optimizes for the corpora that train and ground LLMs. That corpus skews heavily toward two surfaces most B2B teams under-invest in:
- LinkedIn. Posts from credible operators get pulled into LLM training data and retrieved by web-grounded models. A consistent personal brand from the company’s executives compounds over time.
- Traditional PR. Tier-one outlets, trade press, and industry publications continue to feed the citation pool that AI engines weight as trustworthy.
The implication for B2B SaaS: a single quote in a top trade publication can outperform fifty mediocre backlinks for AI citation purposes. So can a thoughtful LinkedIn post from your VP that gets quoted in a roundup six months later.
“We invest in places where the citations carry weight. LinkedIn presence and earned press are doing work for us that classic SEO links no longer do alone.”
— Eugen Knippel, VP Global Voice, Kolsquare
2. Make your site machine-readable: schema markup + technical SEO
If a model can’t parse your page, it can’t quote your page. That makes technical SEO the price of entry for GEO:
- Schema markup (Article, Organization, Product, FAQPage, HowTo, VideoObject) tells LLMs what each chunk of content is.
- Clean HTML and semantic headings make extraction reliable.
- Fast, crawlable pages ensure AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, OAI-SearchBot) actually reach the content.
- Up-to-date sitemaps surface fresh content quickly — important because LLMs increasingly retrieve in real time.
This is unglamorous work. It’s also the reason some brands get cited and others don’t, even when their content is comparable.
3. Replace referrer tracking with self-reported attribution
The hardest shift is measurement. Referrer headers from AI engines are unreliable, often blank, or grouped into “direct.” Kolsquare’s response is the response most mature B2B teams are landing on:
- Add a “How did you hear about us?” field to demo and signup forms, with explicit ChatGPT / Perplexity / Gemini / Copilot / AI Overviews options.
- Cross-reference self-reported source with first-touch URL and branded search lift in Google Search Console.
- Use a dedicated tool to monitor whether and how your brand appears in AI answers across engines.
The point isn’t to perfectly attribute every dollar to a model. It’s to confirm that the channel exists, justify the investment, and track directional movement.
🔭 This is exactly what we built OtterlyAI for — tracking brand mentions and citations across ChatGPT, Perplexity, Gemini, Google AI Overviews, and Copilot, so you can see what AI answers say about you (and your competitors) without guessing. Run a free GEO audit →
Why “lazy AI content” is the biggest GEO risk in 2026
Eugen’s strongest opinion in the interview was about quality, not tactics.
The temptation, with Claude or ChatGPT or any model, is to use them as content treadmills: generate, lightly edit, publish, repeat. The output looks competent. It passes a casual read. It even ranks for a while.
But two forces punish lazy AI content:
- Google’s algorithm updates continue to demote generic, low-effort content — and detection is getting better.
- LLMs preferentially cite sources with original perspective, named experts, fresh data, and clear voice. Sludge gets paraphrased away.
Kolsquare’s discipline is to use AI as a thinking partner, not a ghostwriter:
- Iterate custom instructions constantly so the model reflects current industry knowledge, brand voice, and ICP nuance.
- Keep a human in the creative loop for every customer-facing piece — opinion, story, point of view.
- Use AI to accelerate research and drafting, not to replace the editorial judgment of the people closest to the customer.
The brands that win on both Google and LLMs in the next 18 months will be the ones that use AI to publish less, but better.
A 5-step starting GEO playbook for B2B SaaS marketers
If you want to apply Kolsquare’s approach to your own team this quarter, here is the order of operations we’d recommend:
- Audit your current AI visibility. Find out which prompts already mention your brand (and which mention competitors instead). OtterlyAI’s brand radar does this across the major engines.
- Fix the technical foundation. Confirm your key pages have Article, Organization, FAQPage, and Product schema. Make sure AI crawlers aren’t blocked in robots.txt unless you mean to block them.
- Build an earned-citation pipeline. Pitch your executives into trade press. Help them post consistently on LinkedIn with point-of-view content, not announcements.
- Add a “How did you hear about us?” field to every form with explicit AI engine options. Start measuring next week, not next quarter.
- Set a quality bar on AI-assisted content. Define what “human touch” means at your company — original data, named expert, clear opinion — and apply it before publishing.
Frequently asked questions
What is GEO (Generative Engine Optimization)?
GEO is the practice of optimizing a brand’s content and digital footprint to be discovered, parsed, and cited by generative AI engines like ChatGPT, Perplexity, Gemini, Google AI Overviews, and Copilot.
How is GEO different from SEO?
SEO optimizes for ranking in a list of links on a search engine results page. GEO optimizes for being named or cited inside an AI-generated answer. The two share many fundamentals — quality content, technical hygiene, authority — but GEO weights LinkedIn presence, PR, structured data, and original perspective more heavily.
Does AI search actually send referral traffic?
Today, direct referral traffic from AI engines is small for most B2B SaaS brands. But a growing share of buyers cite AI as their starting point, and that influence shows up in branded search lift and self-reported attribution — not in referrer headers.
How do I get my brand cited by ChatGPT?
Earn high-authority mentions where LLMs look (LinkedIn, trade press, structured data on your site), make your content machine-readable with schema markup, publish original perspective and data, and monitor your visibility with a GEO tool like OtterlyAI.
What schema markup matters most for AI search?
Article, Organization, FAQPage, Product, HowTo, and VideoObject schema cover most B2B SaaS use cases. The goal is to tell AI engines unambiguously what each piece of content is and who published it.
How do I measure GEO performance?
Combine three signals: (1) self-reported attribution from forms, (2) branded search lift in Google Search Console, (3) direct visibility tracking in tools like OtterlyAI that monitor whether AI engines mention your brand in answers to relevant prompts.
Track your AI search visibility with OtterlyAI
If this interview resonated, the natural next step is to find out where your brand actually stands in ChatGPT, Perplexity, Gemini, Google AI Overviews, and Copilot today.
OtterlyAI monitors brand mentions, citations, sentiment, and share of voice across the major AI engines — and gives you the GEO recommendations to act on what you find.
Run a free GEO audit →
See how OtterlyAI works →
Read more on GEO →
About the guest
Eugen Knippel is VP Global Voice at Kolsquare, the European influencer marketing platform. He leads Kolsquare’s global brand and content strategy, with a focus on how AI is reshaping discovery, attribution, and content quality in B2B SaaS.
