Welcome to another edition of Context Window from Novela AI Marketing Simulations!
An interesting thing happened in Dublin last night. Hundreds of people gathered in O'Connell Street for a Halloween parade that didn't exist. They'd seen it mentioned online and, well, that was enough.
You'll be waiting a while, lads
Speaking of trust in online information, OpenAI just made its biggest move yet: ChatGPT Search launched last night for Plus subscribers. It's a significant moment - not just because it puts ChatGPT in direct competition with Google, but because it fundamentally changes how people will find and verify information online.
Sure, ChatGPT Search isn't a "Google Killer" right out of the box. Google still had 90% of global search traffic in September 2024 and it is making its own moves to upgrade the search experience. However, when we add in new tools like Perplexity AI and Meta's moves into search, the direction of travel is obvious. Search is getting chatty.
ChatGPT Search: Room for Improvement
The timing is telling. Just as we're grappling with how easily misinformation spreads (top o' the mornin' to ye, Dublin), we're racing to put AI between people and their information sources. In my marketing classes, I'm already seeing this shift - students have made ChatGPT their default starting point for everything from basic facts to academic research.
It doesn't matter that academic databases exist specifically for research. They've found their comfort zone, and they're staying there. Two years ago, they were skipping academic databases and using Google instead.
🔍 The Clear View: The Power of Defaults
Traditional search engine marketing (SEM/PPC) offered brilliant transparency:
- Labelled advertisements (they are increasingly hard to spot, admittedly)
- Organic results you can evaluate - Trackable traffic patterns
- Somewhat visible authority signals
- Measurable user behaviour
But something fascinating is happening with conversational search.
Take a recent story about Londoners attempting to game AI systems by writing fake restaurant reviews, trying to keep tourists away from their favourite spots.
Users tend to trust these outputs, precisely because they feel more personal, more authoritative. We see it at Novela, too. We have an AI co-pilot in our simulations and we quickly learned that it needed to be accurate all the time, because students would follow whatever it told them to do.
This mirrors what behavioural economists like Daniel Kahneman call the "law of least effort" - we tend to gravitate toward the cognitively easier option. Just as Google became the default choice for an entire generation, we're watching ChatGPT become the default for the next. If its search works well, it's easier to continue the conversation there than to hop over to Google.
🔮 Through The Glass: The Implications
This default bias creates fascinating dynamics:
For Marketing Practitioners:
1. The Eternal Truth
- Users have always chosen convenience
- ChatGPT is winning the same way - it's just a conversation, and everyone knows how to have one of those. What's more, it's a one-sided conversation where the robot listens and does whatever we ask.
2. What Has Changed
- We're not just outsourcing search, we're outsourcing synthesis
- The 'path of least resistance' now leads to AI doing our thinking. ChatGPT's new models will take us much further in this direction
- The stakes are higher because the interface is more persuasive
3. Trust Signals Are Evolving
- Traditional: Backlinks, domain authority, click-through rates
- Emerging: Citation frequency, synthesis-friendly formatting, "quotability"
4. Commercial Reality
- Google adding ads to AI Overviews isn't just about competition
- It's about shaping behaviour before new defaults solidify
- The battle isn't just about technology, it's about habit formation
For Marketing Educators:
This is where my classroom experience becomes particularly relevant. I'm watching brilliant students:
- Skip academic databases for ChatGPT summaries
- Trust AI synthesis over primary sources
- Choose convenience over comprehensiveness
And trust me, I am seeing a lot of very similar-looking assignments as I work through my grading this week.
The challenge is about developing what behavioural scientists call "metacognition" - thinking about how we think and make choices.
👀 Lens Shift
ChatGPT Search isn't just about building a better search engine. It's about changing what people expect from search itself.
Google dominates traditional search because it perfected a delicate balance: giving users what they want while building a massive advertising business. Their results had to be good enough to keep users coming back, but not so complete that users wouldn't need to click.
ChatGPT Search and its rivals disrupt this balance. By providing synthesised answers, it's teaching users to expect complete solutions, not just links. It's moving the battlefield from "finding information" to "understanding information" - and that's territory where Google's ad-first model becomes more difficult to defend.
We saw this tension when Google launched AI Overviews earlier this year. They had to be careful not to make their answers too complete, or they'd cannibalise their own ad business. Now they're trying to thread the needle by adding ads to AI Overviews, but that only highlights the contradiction: How do you monetise a conversation without breaking it?
💡 Back to Basics: The Query is Still King
Let's get atomic for a moment. Behind all this change, some fundamentals remain stubbornly constant: People still have needs. They still search to fulfill those needs. They still buy things and go places.
The query - that moment of expressed intent - is still marketing's golden opportunity. Whether someone types "best restaurants in London" into Google or has a chat with GPT about dinner plans, they're still signaling the same basic intent.
But here's where it gets interesting:
- Traditional search: Query → Results → Evaluation → Action
- AI search: Conversation → Synthesis → Trust → Action
The steps to purchase haven't changed, but the cognitive load has shifted. AI isn't removing the need for marketing; it's changing where and how we apply it.
🎯 What This Means for Marketing Skills
This is why at Novela, we're rethinking how we prepare the next generation of marketers.
But equally, we can't get distracted by the shiny new tools. The core marketing skills - understanding human behavior, crafting compelling messages, building brand trust - are more vital than ever.
The key is developing marketers who can:
1. Think in systems, not just steps
2. Understand patterns, not just platforms
3. Master principles, not just processes
4. Shape conversations, not just content
⚡ The Opportunity
The marketers who thrive won't be those who master ChatGPT search (though that helps). They'll be those who understand:
- How intent manifests in conversation
- How trust builds through synthesis
- How to be citation-worthy, not just clickable
- How to shape thinking, not just search results
This is why our simulations matter more than ever. We're not just teaching tools or techniques
- we're developing the cognitive frameworks needed to navigate this new landscape. Because while AI might handle the queries, humans still need to understand the questions behind them.
Those hundreds of people in Dublin last night defaulted to trust, because that's what humans do. As we build these new AI search systems, that's something we'd do well to remember.
Comments