📖 In This Issue
Featured Snippets: (News & Resources)
Cover Story: My 2027 AI Search Predictions
Operator of Interest: Samantha Torres
Learn This: Latent Space
📰 Featured Snippets (News & Resources)
Google is testing removing AI Mode buttons among other AI first pathways from their search experience. Personally, I think this is to appease investors in the short term that are nervous about Q3’s decline in AI Mode revenue. This likely doesn’t mark a permanent change though.
Google rolled out Gemini 4 Hydrogen. They claim it is “the biggest model upgrade to Search yet.“ But, quite honestly the areas they claim it improves already were really good. So I am not sure we will notice that big of a difference on a per query basis. But I am sure result quality will improve in the aggregate.
ChatGPT is experimenting with displaying what appears to be traditional search results along with generated text. My guess is that this will help with their new ad products better than conversational responses.
Jakub T. (jcubic) Jankiewicz argues that all AI generated code should be open source. He rightly points out that most (if not all) of the training data, that systems like Codex or Claude use, are scrapped directly from open source projects that require some form of attribution in their license. Also, IMO because AI has made software development hyper accessible and ubiquitous, software shouldn’t be a commercial commodity anymore to begin with.
My 2027 AI Search Predictions
It's that time of year again when everyone in the SEO industry dusts off their crystal balls and starts making predictions about what the next twelve months will bring.
And with AI continuing to disrupt how people search for information, there's no shortage of possibilities.
Over the past year, we've seen Google accelerate its AI search products, an explosion of new tools promising AI visibility, and a seemingly endless debate over whether SEO is being replaced by GEO, AEO, or whatever acronym someone invents next.
But while much of the industry is focused on what AI might make possible, I'm more interested in what the companies building these systems are incentivized to do, and what that means for the businesses relying on them.
With that in mind, here are my five predictions for AI Search in 2027.
1. Google Will No Longer Have a Default Search Experience
For more than two decades, Google Search has followed a relatively consistent model. You enter a query, Google processes it, and you receive a search engine results page. The composition of that page has changed dramatically over the years, but the underlying experience has remained recognizable.
I think 2027 will be the year that model begins to disappear.
Instead of treating traditional Search, AI Overviews, and AI Mode as distinct experiences, I predict Google will begin dynamically routing queries into different search experiences based on user intent, behavior, and context.
A straightforward navigational query might continue to return something resembling traditional search results. A complicated research question might trigger a conversational AI experience. A commercial query could produce an interactive shopping interface that allows users to compare products and complete purchases without leaving Google.
And these experiences may become increasingly personalized. Two people searching for the same thing might receive entirely different interfaces based on what Google knows about their preferences and previous interactions.
There's already evidence that Google is heading in this direction.
During its Q3 2026 earnings call, Google described its efforts to build a seamless experience across traditional Search, AI Overviews, and AI Mode. The company also emphasized that searches are becoming longer, more personalized, and increasingly multimodal.
Google's existing query fan-out technology also demonstrates how the company can break complex questions into multiple related searches and synthesize the results into a unified answer.
Of course, that doesn't guarantee Google will eliminate the recognizable traditional SERP. Familiar interfaces have enormous value, particularly for simple searches.
But I suspect we're heading toward a future where the search interface itself becomes another component of Google's ranking and retrieval systems, adapting to the individual and the query rather than forcing every search into a familiar format.
For SEOs, this will make understanding search intent even more important. It will also make traditional rank tracking increasingly difficult to interpret.
After all, what does it mean to rank number one when there is no longer one consistent set of search results?
2. Expect More Ads in AI Mode and AI Overviews
There's a fairly predictable pattern to how Google grows its Search advertising business.
First, it introduces new features and experiences that encourage people to search more frequently. Then, once those experiences gain adoption, Google finds additional opportunities to monetize them.
I don't expect AI Search to be any different.
During Alphabet's October 2026 earnings call, Google acknowledged that AI Mode was monetizing less effectively than traditional Search, which contributed to slower advertising revenue growth despite an increase in search queries.
According to Search Engine Land's reporting, Google's Search advertising revenue still grew 12% year over year, reaching $63.3 billion. However, the lower monetization of AI Mode was a drag on that growth.
Some might interpret this as evidence that AI Mode is a financial disappointment.
I see it differently.
Google is successfully getting people to use its AI-powered search experiences. Now it needs to figure out how to make those experiences as profitable as traditional Search.
And I suspect we already know what comes next.
More ads.
Google has already introduced advertising within and alongside AI-generated responses. Its advertising documentation explicitly describes advertisements appearing across AI Overviews and AI Mode, including follow-up questions within AI conversations.
This creates plenty of additional opportunities for monetization. Sponsored recommendations, commercial comparisons, and advertisements integrated into conversational workflows could all become more common.
There are limits, of course. Google cannot simply fill every AI-generated response with advertising without risking the usefulness of the product. There's also the possibility that advertisers won't see the same returns from conversational experiences that they see from traditional search advertising.
But Google has spent decades figuring out how to balance user experience with advertising revenue.
Its historical playbook is to build distribution, establish user habits, and then extract increasing commercial value from the experience.
I see no reason to believe AI Mode and AI Overviews will be exceptions.
AI Search isn't going anywhere because it's currently less profitable.
It's going to become more profitable because Google will find new ways to advertise within it.
3. Google Will Release Another Helpful Content-Style Update
I'm predicting another significant Google algorithm update in 2027 that will disproportionately affect websites publishing large amounts of low-value AI-generated content.
And when that happens, I fully expect the headlines to announce that Google is finally cracking down on AI content.
They'll be missing the point.
Google doesn't have a fundamental problem with AI-generated content. It has a problem with content that doesn't provide anything useful or original to the people reading it.
In fact, Google's existing spam policies explicitly identify scaled content abuse as publishing large amounts of unoriginal content primarily to manipulate search rankings, regardless of whether that content was created by humans or AI.
Its guidance on generative AI content makes a similar distinction, acknowledging that AI can be useful for research and organization while cautioning against publishing content without adding value.
The problem is that generative AI has made it incredibly easy to produce enormous amounts of content that looks useful without actually being useful.
Anyone can ask ChatGPT to create a content brief, expand it into an article, optimize the headings, and publish the finished product. The resulting content might be grammatically perfect, structurally sound, and technically optimized.
But if it simply summarizes information that's already available on a thousand other websites, what value does it actually add?
That's where I think Google will focus its next major quality-related update.
As AI-generated content becomes increasingly common, the ability to identify genuinely useful information will become even more important. Google has strong incentives to reward content that contributes something meaningful rather than simply repackaging existing material.
This doesn't mean AI-assisted content is inherently bad. It can be excellent when combined with original research, professional experience, unique data, and thoughtful editorial judgment.
But if your entire content strategy consists of feeding generic briefs into ChatGPT and publishing whatever comes out, I wouldn't feel particularly comfortable heading into 2027.
AI can dramatically increase your publishing capacity.
It doesn't automatically increase the value of what you publish.
4. Referral Traffic From Non-Google AI Systems Will Remain Tiny
For the last few years, we've heard repeated predictions that ChatGPT and other conversational AI systems are going to replace traditional search engines as major sources of website traffic.
I'm becoming increasingly skeptical.
In October 2026, Marie Haynes published an analysis of twelve months of Google Analytics data from 100 websites representing more than 145 million sessions.
Her findings were revealing.
Traffic from ChatGPT amounted to just 0.62% of the volume those websites received from Google organic search. When she included all other LLM referral sources, that figure increased to only 0.73%.
Even more interesting, her analysis suggested that ChatGPT referrals had stopped growing exponentially and were actually declining across the sites she studied.
To be fair, her dataset also revealed that visitors from ChatGPT converted at a much higher rate than Google organic visitors: 13.8% compared with 3.02%.
That's significant. And it's a reminder that traffic volume alone doesn't determine the value of a channel.
But despite those higher conversion rates, Google still produced 27 times more conversions than all of the measured LLM sources combined.
Other research has reached similar conclusions. In a November 2026 analysis published by Search Engine Journal, Dan Taylor examined 108 million sessions across 57 websites and found that AI referrals accounted for approximately 0.66% of total traffic.
Now, there is an important distinction here.
Low referral traffic doesn't necessarily mean AI systems have little influence. People can research products, compare businesses, and develop purchase preferences inside ChatGPT without ever clicking a link. Traditional analytics tools aren't particularly good at measuring that influence.
However, that's different from claiming conversational AI will become a major referral traffic source.
In fact, I think these systems are increasingly designed to avoid sending users elsewhere. Their purpose is to answer questions and complete tasks, not necessarily generate website visits.
As these tools become more capable, I suspect that behavior will become even more pronounced.
For businesses, this means AI visibility may become an important part of brand strategy without becoming an especially meaningful source of direct website traffic.
And for SEOs deciding how to allocate limited resources, Google organic search will remain the much larger opportunity for the foreseeable future.
5. AEO Will Become More Popular Than GEO, While SEO Continues Doing All the Work
If there's one thing the SEO industry loves almost as much as algorithm updates, it's inventing new acronyms.
Over the past two years, we've watched the industry debate Generative Engine Optimization (GEO), Answer Engine Optimization (AEO), Large Language Model Optimization (LLMO), and several other variations.
My prediction is that AEO will become the preferred industry term in 2027, gradually overtaking GEO.
Why?
Because it's easier to understand.
Most business owners and marketing executives don't immediately know what a "generative engine" is. But they understand the concept of an answer engine.
And as AI products expand beyond generating text into answering questions, recommending products, and completing tasks, AEO becomes an increasingly convenient label for the work marketers want to sell.
There's already evidence of this terminological confusion.
In a 2026 analysis of AEO and GEO, EMARKETER noted that the industry has yet to establish consistent definitions for the various terms, and that the underlying optimization activities frequently overlap.
But here's the part I find particularly amusing.
Regardless of which acronym wins, the overwhelming majority of practical optimization work will continue to look remarkably similar to SEO.
Websites will still need to be crawlable. Content will still need to be understandable. Brands will still benefit from authoritative third-party references. Structured data, internal linking, information architecture, and technical performance will continue to matter.
Even Google's own AI optimization guidance argues that optimizing for generative search experiences is still fundamentally SEO.
That doesn't mean AI search requires no new thinking. Understanding how systems retrieve, synthesize, and cite information introduces legitimate strategic considerations. Measuring brand visibility inside conversational systems also presents new challenges.
But those considerations are extensions of search optimization, not replacements for it.
Unfortunately, extending an established discipline isn't nearly as exciting as announcing an entirely new one.
So I expect AEO to gain popularity, GEO to slowly lose some of its novelty, and a few enterprising marketers to come up with another acronym before the end of the year.
Meanwhile, the people doing the technical audits, fixing crawlability problems, improving content, and building meaningful brand authority will continue doing the work that actually matters.
They'll just have to learn what to call it this time.
The More Things Change, the More SEO Matters
I could be wrong about some or all of these predictions. That's the nature of trying to anticipate an industry changing as quickly as AI.
But there's a common thread connecting them.
The technology may change dramatically, but the underlying incentives driving search engines, publishers, advertisers, and investors are remarkably consistent.
Google wants to grow usage and advertising revenue. AI companies want to justify their valuations. Publishers want visibility and traffic. Marketers want to demonstrate measurable returns.
And businesses still need customers to find them, understand what they offer, and trust them enough to take action.
That's why I remain skeptical of anyone claiming that AI has made traditional SEO obsolete.
AI is changing how information is discovered and consumed. But it hasn't eliminated the need for valuable content, accessible websites, trusted sources, or a coherent marketing strategy.
If anything, those fundamentals are becoming more important as the systems connecting people with information grow more complicated.
My biggest prediction for 2027 is that the companies making the most progress in AI search won't necessarily be the ones chasing the newest tactics.
They'll be the ones that understand which parts of search are changing, which parts aren't, and where their efforts can produce measurable business value.
👤 Operator of Interest: Samantha Torres
Learn This:
Latent Space:
One more thing: AI is only as good as its operator, and if you are reading this newsletter, you’re better than most!
Till next time,
Joe Hall
PS: Let me know what you think of this issue, or anything else here: [email protected]


