Google vs ChatGPT: Two Ways to Ask the Same Question

We've all asked Google for advice at work. "How can AI help my company's communication or marketing?" feels like one of those questions we type into the search bar between meetings that leads us to spend half an hour looking at websites we have never seen before.

But ask ChatGPT that same question, and you'll get something quite different. Not just in format, but in how the information is selected, structured and delivered.

I tested it—and then analysed the results using Spark, a tool from Hotwire that gives visibility into how AI language models like ChatGPT "think".

What I found was a fascinating shift in how we all are and will seek and find information.

Google: the usual suspects (and a few paid guests)

The first page of Google's response to the question at the top of this article included 18 links. Five were sponsored, which you can't really miss, as they dominate the top of the page. The rest ranged from LinkedIn articles and Reddit threads to industry blogs and media.

Google's logic is well-known: it mixes keyword matching, content authority, and what's worked for other users. It's not bad at all. You get breadth, different formats, and links to click, read, and dive deeper into.

But you also get noise. Four of the links were forum discussions, often outdated. There's no cohesive narrative. It's like walking into a crowded room where everyone simultaneously shouts their version of the answer. You're in charge of filtering it all.

ChatGPT: a guided answer

When I posed the same question to ChatGPT, I received a clean, structured explanation covering eight ways AI can support marketing and communications—from content creation and campaign analysis to social media monitoring.

It was succinct but comprehensive. No ads, just the answer. It's synthesised, saves time, and feels more like a consultant than a shouting room.

But what's under the surface is even more interesting—and that's where Spark comes in.

A new lens on AI's black box

With Spark, I can see beyond what ChatGPT says and analyse how it does it. In my comparison with Google, it revealed that ChatGPT's process is more nuanced and multi-layered than most people assume.

When responding to the "How can AI help my company's communication or marketing?" question, it doesn't just run a keyword match or rank signals. It maps the question to broader themes, identifies subtopics that typically matter to professionals in marketing and comms roles, and then pulls insights from several categories of sources.

Here's how those sources break down looks like:

• ChatGPT selected high-authority outlets like Harvard Business Review and Adweek to shape the model's understanding.
• Analysis from Gartner and Pearson were heavily used as well. These sources had materials validating trends, providing forecasts, and offering comparative frameworks.
• Brands like Canva, HubSpot and Coursera were also featured; interestingly, the material selected was educational.
• Journals, textbooks, and Wikipedia were in the mix, too, and the data shows some were baked into the LLM's training.
• Spark also showed that ChatGPT responses considered potential personas. It assumes that a "CMO" might want strategic insights, a "Marketing Manager" might look for tools and use cases, and a "Brand Comms Lead" might be more focused on campaign ROI.

Only 8% overlap, really

The analysis revealed that only 8% of the results from ChatGPT overlapped with Google. That's a stunning variation for the exact same question.

This means that if you rely solely on SEO to be visible, you might not be showing up in AI-driven environments at all.

This matters because more people are turning to AI boots for search. Somewhere between 400 million and 800 million people are using AI tools monthly in some capacity (it's very hard to find a precise number). Gartner predicts that by 2026, traditional search engine volume will decrease by 25% as AI searches and agents increasingly replace them.

What Spark is quietly telling us

The data also shows that keywords matter, but are just the start. Appearing in AI answers is less about gaming an algorithm and more about building lasting authority: publishing valuable content in reputable places and ensuring it's meaningful, not just searchable.

That might mean rethinking some of your content strategies. It might mean investing more in analyst relations. It almost certainly means understanding the platforms and formats that LLMs favour.

Of course, Google and its billions of daily searches will not go anywhere soon. But the search giant is also integrating AI search into its own traditional tool, and the competition is, for the first time in 20 years, creating real alternatives by making AI tools part of the information ecosystem.

For brands, the lesson is that visibility is evolving. Understanding how AI sees your business is becoming as important as how Google ranks you.