AI Mode & LLM Visibility for B2B SaaS Tips
The one with Antonis Dimitriou
What we talked about
- Communicating AI search changes to clients
- LLM-agnostic optimization
- Pre-training data vs RAG/web search visibility
- Product-feature matching in LLM recommendations
- Reddit, YouTube and off-site channels
- Listicle manipulation / black-hat GEO
- Quantity vs quality in LLM recommendations
- LLM spam, non-determinism and hallucination
- Structuring a $5K/month B2B SaaS SEO budget
- AI Mode vs AI Overviews
- Declining featured snippets
- DOJ docs: AI Overviews use fast search
The questions we dug into
Q1How do you communicate AI search changes to clients so they don't panic?
Show real statistics and studies on how search usage is shifting (e.g., ChatGPT vs traditional search), then clarify what will and won't move the needle.
Asked by Sara · answered by Antonis Dimitriou
Q2If a client wants to appear in ChatGPT, is it worth focusing on?
Stay LLM-agnostic: don't chase specific platforms. Follow a set of best practices and mentions/citations follow if executed well.
Asked by Sara · answered by Antonis Dimitriou
Q3Which LLM visibility practices differ from SEO?
About 90% is good SEO. Two components: create touch points (site, about page, social, off-site content) so your brand is in pre-training data, and rank for the synthetic fan-out queries used in RAG/web search. The other ~10% is matching product attributes the LLM associates with the query (e.g., white-labeling for agencies).
Asked by Nik · answered by Antonis Dimitriou
Q4Are off-site channels like Reddit, reviews and socials new, or were they always SEO?
Antonis: theoretically not new — search engines used co-occurrences — but now it's tangible because you can see ChatGPT and Perplexity querying Reddit. Sara: before LLMs, SEOs rarely included Reddit, YouTube or socials in their offering; it's still under-implemented (she bulk-optimizes client YouTube channels with a script).
Asked by Nik · answered by Antonis Dimitriou & Sara
Q5Do LLMs value quantity of mentions over quality, and can they spot quality?
Antonis: more touch points help disambiguation and OpenAI may use link-graph trust signals; generic queries are easy to game, but specific queries need specific, quality content. Sara: LLMs aren't ready for the coming spam; they're non-deterministic and always hallucinate, which makes Google look more reliable. Nik: his agency showed up in ChatGPT quickly after publishing a few listicles (own site, Medium, LinkedIn).
Asked by Nik · answered by Antonis, Sara & Nik
Q6How would you structure a $5K/month budget for a B2B SaaS client?
Cover basics (technical setup, E-E-A-T pages like About Us, foundational links), start with bottom-of-funnel pages then build topic clusters up the funnel to boost money pages, repurpose content to Reddit, personal social accounts, Medium and video, and add backlinks.
Asked by Nik · answered by Antonis Dimitriou
Q7What are your thoughts on Google's AI Mode?
Both find it random and hard to distinguish from AI Overviews. Antonis uses ChatGPT for deep research and blue links for verification. Nik thinks AI Overviews better meet user needs and hopes AI Mode stays a separate option.
Asked by Sara · answered by Antonis Dimitriou & Nik
Q8Will AI Mode use the same lower-quality 'fast search' system as AI Overviews?
Antonis: a faster system was likely adopted because SGE was slow, but Google will improve it. Nik: the point of both is the fastest possible results; expect ongoing updates like the core SERP.
Asked by Sara · answered by Antonis Dimitriou & Nik
Lines that stuck with us
- Antonis: 'SEO has to be LLM-agnostic.'
- Antonis: ~90% of LLM visibility is good SEO; ~10% is AI-search specifics like product-trait matching.
- Nik published a few listicles (own site, Medium, LinkedIn) and saw his agency recommended in ChatGPT.
- Sara: in the DOJ documents Google said most quality signals come from the page, not users.
- Study cited by Sara: only ~2% of users clicked into AI Mode unless prompted.
- Sara's newsletter: AI Overviews use 'fast search', returning lower-quality results by design.
- Nik: focusing only on LLMs is like ignoring 'the desert' for 'one piece of sand'.
Note: This page was AI-generated from the podcast transcript and manually reviewed for accuracy and context. While we’ve made every effort to ensure it reflects the original conversation, mistakes or omissions are still possible. If you spot something that doesn’t look right, please reach out and let us know.