Why Observability Alone Fails

If you’re not getting cited by ChatGPT, your “AI SEO” strategy isn’t working, no matter what your dashboards say. Most of it is observability theater: dashboards, charts, synthetic prompts — and zero actual placement.

In this episode, we chat with Shawn Schneider, founder of Eldil AI, about what actually determines whether your company shows up in ChatGPT answers. The short answer: LLMs don’t reward more content, clever prompts, or prettier dashboards. They reward a small set of trusted third-party sources — and most brands aren’t mentioned in any of them.

Shawn breaks down why observability alone creates a false sense of progress, how to identify the specific citations that dominate your category, and how to turn that insight into real placements through outreach and negotiation. We also unpack why Google Search Console is still the best signal we have for AI-driven queries, how to prioritize the one citation that actually matters, and what the first 30–90 days can look like when you do this correctly.

Guest

Shawn Schneider — founder of Eldil AI, a GEO / AI SEO platform focused on identifying and securing the citations LLMs rely on most; helps brands and agencies win visibility in ChatGPT by targeting the power-law sources that shape AI answers.

Guest Links
LinkedIn: https://www.linkedin.com/in/shawn-schneider-61b2b5207/
Company Website: https://www.eldil.ai/

What You’ll Learn

• Why most GEO / AI SEO observability tools are meaningless without actual placements 
• The only thing that reliably improves AI search visibility: citation placements
• How to use Google Search Console to surface AI fan-out queries
• Why synthetic prompt data is still unreliable (and what to trust instead)
• The power law of citations: why only 1–3 sources actually matter
• How Eldil turns citation discovery into outreach and negotiated placements
• What 30–90 days can look like when you secure the right citation
• Which industries should invest heavily — and which should ignore this for now
• Why ChatGPT dominates referral traffic compared to other LLMs
• What happens when ads arrive inside AI search results

Timestamps

• 00:00 — GEO, AI SEO, AEO: noise vs. reality
• 00:21 — Why observability tools don’t move the needle
• 03:55 — Where GEO tools get their data (and why it’s messy)
• 07:16 — Using Google Search Console as a prompt proxy
• 09:40 — The three pillars: technical, content, authority
• 12:07 — Citations as the dominant ranking lever
• 13:07 — The power law: thousands of citations, one winner
• 19:07 — How fast results actually show up
• 20:39 — When building your own citation content makes sense
• 30:41 — Which business models win with GEO
• 37:11 — ChatGPT ads and the future of AI search
• 41:32 — Where to find Shawn and closing thoughts 

Key Topics & Ideas

1. Why dashboards feel good but don’t create outcomes.

• Most tools are essentially “Google Analytics for LLMs”
• ChatGPT referrals rise naturally as usage increases
• Charts go up even if you do nothing
• Without placements, observability is just vanity

2. The three common approaches in the market today:

• Guessing prompts with LLMs
• Clickstream data sourced from Chrome extensions and brokers
• Synthetic prompts without transparency

Eldil uses Google Search Console + Analytics as the best available proxy for real intent.

3. How to spot AI-generated fan-out queries:

• 50+ character queries
• High impressions
• Low or zero clicks

These often represent LLMs expanding short prompts into long-form searches.

4. The three pillars: Technical, Content, Authority

• Technical — can an LLM crawl and understand your site?
• Content — does useful information exist?
• Authority — does anyone credible back it up?

Authority is the multiplier most teams ignore.

5. What actually shapes AI answers:

• Citations are not backlinks, they are semantic explanations
• LLMs repeatedly return to the same trusted sources
• Third-party listicles and niche blogs dominate citation share

6. The Power Law of Citations

• 10k–15k citations may exist
• 200–300 matter
• 1–3 actually move the needle

If you’re not in those, content volume won’t save you.

7. The real workflow:

• Identify high-value customer questions
• Extract dominant citations
• Rank them by weight
• Contact site owners
• Negotiate placement
• Monitor AI visibility and referral traffic

This is where most tools stop — and where Eldil focuses.

8. How many placements do you need?

Surprisingly few.

• You don’t need 100 placements
• You need the right one
• Then expand into adjacent verticals

This is concentrated betting, not spray-and-pray SEO.

9. Why GEO feels different from traditional SEO:

• You are inserting into sources that already rank
• Changes can show up in weeks, not years
• Meaningful referral growth often appears within ~60–90 days

10. Who Should (and Shouldn’t) Do This

Best fit:

• High-ACV B2B SaaS
• Long buying cycles
• High-LTV e-commerce (supplements, skincare)
• ICPs that already live in ChatGPT

If your customers do not use LLMs yet, start elsewhere.

11. Why ChatGPT is the main event

Based on Eldil’s data:

• ChatGPT referrals dwarf Perplexity and others
• For most companies, this is where focus belongs
• Smaller channels still matter for high-ticket sales

12. What’s coming next

• Paid placements inside LLMs
• Organic plus paid becoming a one-two punch
• Citation inventory getting expensive fast

The window for cheap dominance will not last.

Sponsor
Today’s episode is brought to you by Graphed – an AI data analyst & BI platform.

With Graphed you can:
Connect data like GA4, Facebook Ads, HubSpot, Google Ads, Search Console, Amplitude
Build interactive dashboards just by chatting (no Looker Studio/Tableau learning curve)
Use it as your ETL + data warehouse + BI layer in one place

Ask:
“Build me a stacked bar chart of new users vs. all users over time from GA4”
…and Graphed just builds it for you.

👉 Get a free trial worth $500 plus a personal discovery call to set up your first dashboard: https://graphed.com/