The CEO of Dreamdata says stop trying to game AI search

Nick Turner runs Dreamdata now. Three months into the CEO role after proving himself for a year as the CRO. He has 20 years in marketing technology, much of it in SEO and content marketing.

He's been watching the AI search chaos unfold with an odd sense of déjà vu. It looks like 2000 all over again—multiple search engines, different algorithms, everyone scrambling to figure out what works. Except this time the algorithms change every few hours and nobody can track much of anything.

More and more companies are starting to wonder, how much pipeline is ChatGPT actually driving?

You probably don't know. Nobody really does.

Less than 0.5% of Dreamdata's pipeline comes from ChatGPT, or does it?

Nick pulled Dreamdata's numbers. ChatGPT has verifiably influenced less than half a percent of their pipeline. That's roughly $70,000 out of significantly more.

But he knows that number is wrong. Customer calls mention ChatGPT. Sales hears it in discovery. The influence is there but they just can't track it.

Throughout B2B industries, there is a similar story. The problem isn't that AI search isn't happening. The problem is marketers are flying blind.

Tracking AI search is harder than early SEO. Google gives you the same results for the same query, day after day. Consistent. Measurable. Optimizable.

Type a search into ChatGPT at 10 AM. Type it again at 2 PM. You'll get completely different answers. It's generated fresh every time, predicting the next word based on whatever content it finds relevant in that moment.

Now multiply that across seven different LLMs. Each with their own algorithm. Each crawling sites differently. Each producing different results for identical queries.

You want to track that? You'll need API credits for every test query. Across every model. Multiple times per day because the results keep changing.

What worked in 2000—tracking rankings across AltaVista, Yahoo, Excite, Ask Jeeves, and Google—was plenty complicated. This is exponentially worse.

What twenty years of SEO taught Nick

Nick's approach contradicts what most "AEO/GEO experts" are selling: Stop trying to game the system.

"You're immediately talking about how do I game the technology that's used to index and surface this content to our users," he says. "You should be having the same content for any channel for the most part if you're targeting your user well."

What has consistently worked across 20 years? Long-form content written by credible authors.

Not because length matters. Because depth signals specificity.

When you write 3,000 words about a technical problem, you're naturally targeting a narrow ICP. You can't fake that level of detail without actual expertise. The content self-selects for the right audience whether you're optimizing for Google or ChatGPT or whatever comes next.

Nick points our that Google factors in the author as part of the algorithm. Author credibility matters. The person writing needs to be believable. Their expertise needs to be demonstrable.

Generic AI-generated content fails here. It regurgitates from other websites. No original insight. No credibility. Just recycled information that won't stand out in any algorithm.

In 2000, you could keyword stuff. You could game multiple search engines because the algorithms were simple. Today's LLMs are more sophisticated, but they're also more fragmented. You could probably game ChatGPT right now if you wrote specifically for it.

But why would you? The channel is still tiny.

When Nick filters Dreamdata's content analytics to show blog posts that reached SQL stage, he sees 30% of deals influenced by specific pieces of content. That's tied to actual revenue numbers. Not traffic, not rankings, but pipeline.

Their platform integrates CRM data with Google Search Console, referrer URLs from ChatGPT, and activity tracking across 12+ systems most B2B companies use. Email marketing platform. Paid search. Content management. Chat bots. CRM.

Each system only sees part of the customer journey. Dreamdata stitches it together to show which content actually influences deals.

You can map individual prospects to specific search queries. Filter by closed-won deals in the last 60 days. See that 24 deals worth half a million came from organic search as the first touch channel.

It's not a complete solution to the AI tracking problem. But it's better than guessing.

Nick also predicts one LLM will eventually specialize in search. They'll make it their core competency and build the tooling SEOs need. It would be something like Google Search Console for their platform. We’re seeing these platforms popping up now but they’re still young.

Until then, focus on what you can control. Write genuinely useful content for your actual ICP. Establish author credibility. Connect whatever tracking exists to revenue.

The question isn't "How do I write for ChatGPT versus Google?" The question is "How do I prove my content drives pipeline?"

The parallel to early search engines isn't just history, it's a playbook

Google won partly because they made it easy for content creators to measure success. Any LLM that wants to dominate search will need to solve the same problem.

Early adopters who figure out attribution now gain a competitive advantage. When better tools arrive, they’ll already have proven strategies while competitors are experimenting. The channel will mature. One platform will emerge as the leader. LLM platforms will build measurement tools because they'll need content creators to feed their algorithm.

But here's what most people miss. Paid search doesn't work unless organic works. People need to trust the platform's results. They need to believe the content is credible before they'll tolerate ads.

Nick thinks ads are coming to LLMs. The opportunity is too big to resist. ChatGPT has massive daily active users. How do you monetize free users if not through ads?

When that happens, attribution becomes critical. You'll need to connect AI-influenced touches to actual pipeline. Not just traffic. Not just clicks. Revenue.

Connect content to CRM data or keep guessing what works

Connect your content performance to CRM data. Most content strategies die in the gap between "this got traffic" and "this influenced revenue."

Track what you can track. Referrer URLs from ChatGPT. Activity from LLM bots crawling your site (Botify has good research on this. The bandwidth these bots consume has skyrocketed). Google Search Console queries that led to conversions.

Then map those touches to deals in your CRM.

Start there, even if AI search remains hard to measure. Because when the tooling catches up, you'll already know what works.