TikTok Was Supposed to Kill Google. ChatGPT Is Actually Doing It.

For the last two years, you could not attend a marketing event without someone declaring that TikTok was going to kill Google. Gen Z doesn't search with Google anymore. Visual search is the future. If your business isn't on TikTok, you're building on a disappearing foundation.

Good story. Mostly wrong.

New data from Adobe Express, surveying 807 consumers in January and February 2026, shows that the actual threat to Google isn't TikTok at all. It's ChatGPT. And the gap is widening fast.

What the Numbers Actually Show

Adobe Express asked consumers which platforms they preferred over Google for search. The results:

• ChatGPT: 14% of consumers prefer it over Google
• TikTok: 7% of consumers prefer it over Google
• ChatGPT is now twice as likely as TikTok to replace Google

That's not a rounding error. ChatGPT's preference rate doubled TikTok's in the same survey period.

The platform helpfulness rankings tell a similar story. When consumers were asked which platforms were most helpful for finding information:

• Google: 85%
• Reddit: 29%
• ChatGPT: 26%
• YouTube: 24%
• TikTok: 16%

ChatGPT, despite being a fraction of Google's age, is already more helpful to consumers than YouTube. TikTok sits at the bottom of that list.

And the Gen Z angle that powered the TikTok-will-kill-Google narrative? Gen Z's preference for TikTok over Google collapsed from 8% to 4% between the two survey periods. The cohort most likely to have made TikTok search a permanent shift is already moving on.

Why TikTok's Search Moment Never Quite Arrived

To be fair to the TikTok thesis, it wasn't entirely fiction. Younger users did (and still do) use TikTok to discover restaurants, find tutorials, and research lifestyle purchases. But the key word is "discover." TikTok is good at surfacing things you didn't know you wanted. It's far weaker at answering questions you actually have.

When you need to know how to structure an invoice, evaluate a payroll software, or figure out whether the Productivity Solutions Grant covers a specific tool, you don't open TikTok. You search. And increasingly, that search is going to ChatGPT rather than Google.

The distinction matters for businesses. Discovery-mode traffic is real, but it's harder to convert and harder to attribute. Question-and-answer traffic, the kind ChatGPT is capturing, is high-intent by definition. Someone asking ChatGPT "which digital marketing agency in Singapore is good for SMEs" is much further down the decision journey than someone scrolling past a TikTok video.

The Scale Problem Nobody Talks About

It's worth putting ChatGPT's growth in context. According to Backlinko, ChatGPT reached 900 million weekly active users by December 2025, up from 300 million in December 2024. That's a 200% increase in one year.

No platform in the history of the internet has grown its active user base that quickly at that scale. And unlike most social platforms, ChatGPT usage skews toward people making decisions: researching purchases, evaluating vendors, learning new skills, doing work.

For businesses, this means the audience using ChatGPT isn't just bigger than it was six months ago. It's better. These are not passive scrollers. They're active searchers with specific intent.

Meanwhile, investment in TikTok among SMEs is cooling. Adobe Express found that 38% of SMEs plan to increase TikTok investment this year, down from 53% in 2024. The FOMO cycle that drove TikTok adoption may be reaching its natural ceiling.

What This Means for Your Content

This data changes how you should think about your content investments. Not dramatically, and not all at once. But the direction is clear.

Depth beats discoverability. TikTok-optimised content tends to be short, visually driven, and designed to stop someone mid-scroll. ChatGPT-optimised content is the opposite: substantive, specific, and structured to answer real questions. If someone asks ChatGPT about your industry, will any content you've published be worth citing? If the honest answer is no, that's the gap to close.

Authority signals matter more, not less. Google's E-E-A-T framework (Experience, Expertise, Authoritativeness, Trust) was designed for traditional search, but it turns out to be exactly what AI models look for when deciding whether to surface your content. Articles written by identifiable people with demonstrable expertise, linked to from credible sources, and updated regularly are far more likely to be retrieved by ChatGPT than thin, keyword-stuffed pages. This isn't new advice. It's just now doubly important.

Structure is a competitive advantage. ChatGPT and other large language models are better at retrieving and citing content that is logically structured: clear headings, concise answers, factual specificity. FAQ sections, how-to guides, numbered frameworks, and comparison tables all perform well in AI retrieval contexts. The irony is that this is also what makes content useful for human readers. Good content has always worked. AI search just makes the penalty for bad content more immediate.

Singapore Businesses Have a Structural Advantage Here

There's a reason the English-first content angle matters so much. Research by Peec AI found that 43% of ChatGPT's "fan-out" queries (searches that draw from multiple sources) default to English-language content, regardless of the user's native language.

Singapore's business environment is English-first by default. Your website, your blog, your case studies: if they are written in clear, professional English, they are already optimised for the dominant retrieval language of AI search. Businesses in markets with fragmented language environments (mainland China, Japan, parts of Southeast Asia) face a significantly higher barrier.

That advantage only holds if the content is actually good. Publishing English-language marketing fluff does not help. Publishing English-language content that genuinely answers the questions your buyers have, with specificity and authority, puts you in a strong position as AI search continues to displace traditional search volume.

We covered this in detail in our earlier piece on why English-first content gives Singapore businesses an edge in AI search.

The Practical Shift: Three Things to Prioritise

If you're allocating marketing attention over the next six months, here's where it goes based on where search is actually heading:

1. Audit your existing content for AI-readiness. Go through your top 10 pages. Ask: if someone asked ChatGPT a question this page answers, would it get cited? Look for thin sections, unsubstantiated claims, and content that hedges everything. Make it specific. Make it citable.

2. Build your Google Business Profile aggressively. Local search intent is still largely captured by Google Maps and the Local Pack, and it is not going away. For businesses serving a specific geography, GBP reviews, posts, and correct categorization drive a disproportionate amount of high-intent traffic that AI search tools do not yet adequately replace.

3. Don't abandon TikTok entirely, but recalibrate. If TikTok is generating real leads or sales for your business right now, keep it going. But if you've been investing in it out of FOMO with unclear returns, the data now gives you permission to reallocate. The marginal dollar spent on a well-structured article that ChatGPT will cite is likely to compound more over the next two years than the marginal TikTok video.

One More Thing: Where Your Content Lives Matters

Depth and authority aren't enough on their own. AI models can only cite content they can access. That means:

Paywalled content doesn't get cited. If your best thinking is locked behind a lead form or login, ChatGPT can't surface it. Consider publishing at least a portion of your most substantive work publicly.

Technical accessibility matters. AI crawlers need to be able to reach your pages. If your robots.txt is set to block all AI crawlers as a blanket measure, you may be preventing legitimate AI search tools from indexing your content. There's a difference between blocking training crawlers (reasonable) and blocking retrieval crawlers (costs you AI search visibility). Our AI crawler guide covers exactly how to draw that line.

Schema markup helps. Structured data gives AI models clearer signals about what a page is about, who wrote it, and when it was published. FAQ schema, Article schema, and Organisation schema are worth implementing if they're not already in place.

None of this is complex. It's largely the same SEO hygiene that competent websites should already have. The difference is that AI search amplifies the gap between websites that have it and those that don't.

What to Watch

The Adobe Express data is a January/February 2026 snapshot. The trajectory it shows has been consistent for over a year, but the specific numbers will shift. ChatGPT is not the only AI search tool in the mix: Perplexity is growing, Google's own AI Mode is expanding, and Microsoft's Copilot integration into Windows and Edge is quietly building search volume.

The underlying point holds regardless of which AI tool eventually dominates: search is fragmenting, AI is capturing a growing share of it, and the content that performs best is authoritative, structured, and specific. That's a more durable bet than optimising for any single platform's algorithm.

If you want to understand how AI crawlers are currently reading your website (and what you can allow versus block), our guide on AI crawler settings for Singapore SMEs walks through the practical decisions. And if you're seeing traffic shifts related to Google's AI Overviews, we covered what the recent changes mean here.

The channel landscape is changing. The fundamentals of useful content are not.

---

Magnified helps Singapore SMEs build content that works for both human readers and AI search. See how we approach SEO and content strategy or get in touch to talk through your situation.