SEO / GEO Gap: How to Convince Your Clients Without Pushing Them Away

A client calls me on a Tuesday morning. « Stéphane, I want page one of Google. »

Status quo bias: why your clients are locked in a comfortable loop

A human prefers what they master. Especially when their revenue depends on it. When a founder sees his site in position 3 for « Italian leather sofa, » he sleeps better. That SERP line is tangible. It cost him sleepless nights, campaigns, meetings. So when you talk to him about GEO, he hears the noise of a trendy buzzword. And he tunes out.

The DOSE framework, as Guillaume Attias teaches it at BMO Academy, captures this well. D for Trigger. O for Organization of the system. S for Stepping out of the action tunnel. E for Evolution of beliefs. Here, status quo bias clings to the first stage. The trigger hasn’t been powerful enough to break the routine.

Yet 47% of the clients I audit each month have never heard of AI visibility. Not because they’re closed-minded. Because they manage inventory, a team, cash flow. They don’t have time to chase the latest Search Engine Journal studies.

So how do you trigger that shift without scaring them? By speaking the language of their P&L. No jargon. No « sales funnel. » Just numbers they understand in three seconds.

Take another client. An online apparel store, 1,200 models, founder running it solo. One day I place two charts on his desk. His Google traffic curve: flat, +4% per year. The AI citation curve in his sector: +820% in 14 months. He looks at me. Says nothing. Just asks: « How long to catch this train? » That day, status quo bias exploded in silence.

You don’t need to convince a client that AI exists. He sees it in the news. He uses it personally. Show him how that reality intersects his business. With data from his own site, not promises.

Show him a conversational query that led to a transaction. Show him new competitors on his keywords. Show him especially what he loses if he doesn’t move now. Not with alarm. With positive projection. Imagine he captures 10% of future AI traffic instead of leaving it to competitors. What’s that worth?

GEO isn’t a replacement for SEO. It’s an additional channel. And any channel you neglect is a lever you hand to your competitors without a fight.

The three steps I use to shift a founder from SEO to GEO

I never pitch a new strategy as a rupture. I add a layer. I build another floor on solid foundations. Here are the three phases I’ve applied across 15 e-commerce mandates.

Phase 1: Observe without judging. For the first two weeks, I scrape logs, Search Console, external citations. I hunt for weak signals. A citation on Bing Chat, a sudden spike from Copilot, a mention in a generated summary. For a client specializing in tools, I uncovered 218 sessions from Gemini in one month, with a 2.7% conversion rate. The founder had no idea. That was the moment.

I don’t tell him: « You need to invest in GEO. » I show him: « Here, 218 visits, 5 orders, €290 average order. Zero acquisition cost. Your Google Ads campaign costs you €8.40 per click. » The proof is in his own reporting.

Phase 2: Show the silent evolution. I build a timeline. January: 5 AI citations. March: 14. June: 47. September: 98. This pace follows an unwritten rule: volume doubles every six months on transactional verticals. I also show how Google click-through rate erodes on informational queries. 37% of searches with no clicks, per a SparkToro study. That’s invisible value transfer.

The client then sees two paths: double down on pure SEO and accept slow decline, or expand to generative agents and capture new audience entirely.

Phase 3: Build a prototype for €2,400. No big rollout. Just focus on 20 key product entities. I create semantic cocoons designed for conversational questions, enriched with structured data tuned for language models. Within 90 days, a bio cosmetics client generated €7,230 in revenue solely via Bing Chat and Copilot. Cost? €2,400 in content and architecture. Return? Immediate. No ads, no links.

At this point, the client stops asking why GEO. He asks how to amplify the flow.

These three phases don’t cost a fortune. They don’t replace anything. They add a capture layer on an ecosystem Google itself is pushing. Because yes, the legacy search engine increasingly serves conversational answers. Adapting to it keeps you in the game.

The time frame: make them feel what they risk losing (without scaring them)

Nobody likes hearing they’ll lose traffic. Counterproductive. But everyone understands potential gain. Time framing lets you anticipate the future instead of suffering it.

Take a concrete analogy. In 2016, Google introduced featured snippets. The first sites to structure their data claimed position zero. Result? Average click gains of +32%, per Search Engine Land analysis. Those who waited two years watched competitors own that real estate. History rhymes with AI.

Today, a conversational agent doesn’t work like a snippet. It synthesizes multiple sources. It rarely cites more than three or four domains per answer. Sites that optimize topical authority first become preferred sources. Others vanish from responses.

So I ask my clients this: In 18 months, if 30% of informational queries run through a conversational agent, what share of that traffic do you want?

They always answer: « As much as possible. » Then I ask: « And today, how many resources are you dedicating to it? » The silence that follows says more than ten pages of argument.

Time framing shouldn’t be a sword of Damocles. It should be positive growth projection. Every month without GEO work is a month your competitors embed themselves in language models. Every piece you write thinking only of Google is an asset invisible to 40% of future searchers.

Lily Ray recently shared on her blog analyses showing that sites already regularly cited in AI agents started building brand and entity authority long before ChatGPT exploded. That’s the result of long-term presence strategy.

To convince a client, show the path. Not the cliffs. Show the gap between capturing 2% of AI citations and 12%. Price it with his own margins. Converted to net euros, that gap becomes a strategic budget advance. Suddenly, GEO isn’t consultant fashion. It’s rational growth leverage.

€7,230 in 90 days: when GEO outperforms SEA, with zero ad spend

The case I mentioned earlier deserves deeper look. A natural cosmetics site, 370 products, 5,200 organic monthly sessions in October 2024. €63 average order value. The founder bet everything on Instagram and Google Ads. That was his only strategy.

We built a first prototype on 19 key entities. Semantic cocoons designed for conversational answers, with enriched structured tags for agents: concise descriptions, Q&A, trusted sources linked to product sheets. No fluff content. Useful text, definitions, machine-readable comparisons.

In 90 days, here’s what we got:

• 1,247 sessions from AI agents (Bing Chat, Copilot, Gemini)
• 113 conversions tracked through the e-commerce funnel
• €7,230 in revenue directly attributed to those clicks
• Implementation cost: €2,400 (semantic research, structuring, deployment)
• Return on investment: +301% in three months

The client didn’t spend a cent on ads for these sales. He didn’t increase Google bids. He simply made his content readable by machines seeking the best answer to an intent.

What struck me was that these AI visitors aren’t « curious browsers. » They arrive with a specific ask. « What moisturizer for atopic skin? » « Organic solid shampoo comparison. » They buy. Often faster than standard organic visitors.

That same quarter, the client’s Ads campaigns showed a CPA of €14.70. Here, CPA is nearly zero. The only cost is semantic architecture investment. Once content is structured, it keeps generating visits month after month, with no daily upkeep.

This isn’t magic. I’m not claiming GEO will replace SEA. But for a brand building lasting presence, it’s a full channel. And this channel is nearly competition-free right now. It’s like SEO in 2004. When few were there, and every euro returned ten.

Now look at your clients. What do they pay per click on Google Ads? And what if a channel existed where cost per acquisition isn’t driven by auction, but by information quality? That’s the equation attracting decision-makers.

No magic content: what actually works for appearing in AI agents

Every month I see articles promising the GEO holy grail. « Secret ChatGPT prompts, » « revolutionary templates. » I’ll be straight: it doesn’t work. Clean architecture generates recurring citations. And I’ve been forging it since 2016 with semantic cocoons.

First building block: entities. Generative agents reason in entities, not keywords. A product, a brand, a characteristic, a problem. Your content must organize around these entities, with internal links clarifying each page’s meaning to machine learning. On the cosmetics site above, every product sheet links to an ingredient, skin type, application ritual. That semantic web made the site emerge as a trusted source for complex conversational queries.

Second building block: structured data for agents. FAQ markup, HowTo, QAPage, especially concise descriptions in blocks readable by LLM attention layers. The goal: deliver a ready-to-cite answer, no spam, source clearly identified.

Third building block: domain authority in a narrow theme. Language models privilege sources showing continuous expertise. A well-built cocoon creates dense interconnection signaling strong topical authority. You don’t need 10,000 pages. 47 tightly linked pages in one skill silo suffice to become the reference point for an AI.

I see with my clients that combining these three blocks produces a snowball effect. The more the agent cites you, the more visibility you gain, the more other agents cite you in turn. It sustains itself. Unstructured content stays invisible, regardless of writing quality.

Don’t fear simplicity. One opening paragraph answering « What is X? » packs more punch than 800 words of convoluted storytelling. Machines love clarity. So do humans.

Finally, measure. Use your analytics to trace sources like bing.com/chat, Copilot extensions, etc. Ask clients to install specific tags. Otherwise, this traffic remains a ghost. And you’ll never have the numbers to justify the next investment.

And your clients?

I’m not selling you the method. I’m showing you the pages.

In 18 months, when 30% of informational queries run through conversational agents, your clients will have three options:

• Already have structured presence and capture this audience before the rush.
• Watch competitors claim citations and spend ten times more trying to dislodge them.
• Tell themselves they should have listened to this signal when it was still weak.

The first option costs a few thousand euros and an architecture rethink. The second costs years of lag and vanished marketing budgets.

So when you pitch GEO to your next client, don’t talk AI. Talk revenue. Show them the €7,230 a comparable site generated in 90 days with zero ads. Show them the double-digit growth curve. Let the numbers speak.

The choice is theirs. But at least it’ll be informed.

And your clients—how long before AI starts eating their historic traffic without them seeing it coming?

Stéphane Jambu

SEO & AI Engineer

I build growth systems / AI / Neuroscience | 650+ clients · 80 LinkedIn testimonials · 30 years of expertise · 15 years of systems running without me.

Follow on LinkedIn