The personalized Internet is here
For now, personalization is really only in Gemini, but it will soon migrate globally to AI Mode, which then goes to AI Overviews and all of search.
This week’s newsletter is sponsored by North Star Inbound and Semrush Enterprise
If you will be at Google I/O this year, let me know if you are there and would want to grab coffee
Paid subscribers can download my personalized search workbook
At Google I/O last year, Google announced a product they called context-aware search, a description that could be game-changing: a search engine that answers questions based on who you are rather than just what you typed into the box.
This is a capability Google has been trying to build in some form for most of its existence as a company, and every previous attempt didn’t live up to the hype because the underlying technology wasn’t good enough to make personalized search work in a way that improved searcher experiences.
[Sponsored by Semrush Enterprise]
Only 22.5% of teams have fully integrated SEO and AI search workflows. Is yours one of them?
Most brands are already losing in AI search without knowing it. And most marketing leaders can’t pinpoint where.
Semrush published two new playbooks to solve this, giving marketers the foundation to understand the landscape, secure internal buy-in, and start building toward execution.
Brands that treat SEO and AI visibility as separate workstreams are building on a fault line. The ones pulling ahead are making visibility an enterprise operating model.
AI finally closed that gap, and when I first heard the announcement at I/O last year, my first reaction was that if Google could actually ship this as described, the SEO industry would need to be rebuilt very quickly around a completely different set of assumptions. For now, personalization is only in Gemini, but it will soon migrate globally to AI Mode, then to AI Overviews, and eventually to all of search. The future is now.
AI is the tool personalization has been waiting for
Every personalization attempt before this one was built on the same foundation: Google would look at what you’d searched before, layer in some browsing history from Chrome, try to infer something about you from those signals, and then adjust results. The result of that attempt was not meaningful enough, as it became a glorified “near me” with only a few parameters added.
A user searching for “Giants” (for non-US readers, this is a local sports team) in NYC would get different results from someone searching for the same term in California. A user searching for credit cards in the US would get different results than someone in Canada. Since traditional search was built on keywords, it could only append more words to the query lookup, rather than append entire objects as AI does.
Personalization is built on a series of flags
What Google built this time is architecturally different in a way that matters because AI is driven by questions already loaded with context. The personalization adds even more context on steroids. It runs on the same infrastructure that powered Google Alerts, which the team internally called the notifications system, and the logic is much simpler and more powerful than anything they’d tried before. Every search you run sets a flag that signals to the system you are interested in that topic. As those flags accumulate over time, every future search gets answered in the context of all the interests you’ve already demonstrated.
In real time, Google continuously builds a personal history of you through those notifications, and that history becomes the background context for every answer the system generates. Product recommendations will be based on your needs, budgets, and past purchases. Unlike traditional SEO results, this is not a small adjustment to the ranking order; it is foundational to your experience and will ultimately lead to a different answer for a different person.
Your data is everywhere in Google
This is your entire life on Google, not just your searches. Google pulls in your calendar invites to understand who you meet with and what those conversations might have been about, which it then combines with Gmail signals, Maps history, and every behavioral trace you’ve left across Google’s products over the years. Then it uses AI to fill in any remaining gaps, constructing a fairly complete picture of your professional and personal life. Android users will have their call history, contacts, and more added to the context.
I did a search last week where Gemini told me to consult my CPA to confirm its recommendation and then listed my CPA by name, getting it right, based on a single meeting invite in my calendar. It didn’t ask me who my accountant was, nor did it prompt me to connect an account. It just knew and incorporated that knowledge into the answer, just as a genuinely well-informed assistant would. There will be hallucinations and mistakes embarrassing enough to make news, but the trajectory of this technology is toward greater accuracy over time, not less.
The product launched as a beta inside Gemini first, then rolled out to AI Mode in January across nearly every market where AI Mode is live, and it is not yet inside AI Overviews, which is the experience most SEOs still think of as the main search event.
AI mode is coming
Eventually, Google will announce that AI Mode becomes the default search experience, not an opt-in or a toggle or a lab feature, but what you get when you search. That announcement would make everything that follows inevitable, and we are on the cusp of a huge change.
Even a “near me” query for dinner suggestions won’t produce a ranked list of nearby restaurants anymore. Instead, it will be shaped by your DoorDash orders that showed up in Gmail, the restaurants where you’ve had calendar meetings, your Maps navigation history, and every restaurant you’ve ever clicked on in search, all weighted simultaneously with the magic dust of AI. Two people in the same neighborhood searching for the same phrase at the same time will get different results.
The implication for SEO and AEO is that keywords are now genuinely secondary in a way they’ve never been before. Words in a query were always a proxy for what we actually cared about: reaching a user at the moment they needed something specific. The keyword was just the mechanism we had available to narrow that down.
That mechanism is becoming obsolete. Optimizing for the keyword without accounting for the person who typed it is like that annoying car salesperson trying to sell you financing before you even settle on a car that you want to buy.
All LLMs will follow Google’s lead
I expect the competitors to follow Google here because the experience gap will become impossible to ignore once enough users spend time with genuinely personalized AI search. An LLM giving a generic answer about mortgage options to someone whose credit score and financial history Google has already factored in is a product that will feel noticeably worse, and users will feel that difference even if they can’t articulate why.
Every LLM will build its own personalization layers. Still, none have what Google has: two decades of behavioral history across billions of users, spanning search, email, calendar, maps, and browser activity. There may eventually be a mechanism for users to export their Google history and bring it to a competing model, some data portability layer that makes personalization transferable, but for now, Google has the advantage.
What you need to know
A business can now have zero visibility in any keyword-tracking tool and still drive significant revenue from AI-powered search because the system routes relevant users to them based on contextual fit rather than keyword match. The inverse is true too: you can hold strong positions for high-volume keywords and see no traffic because the user doesn’t need you. Tracking the keyword position tells you nothing about either scenario.
Paid subscribers can download my personalized search workbook
Instead, you need to know whether the right people are finding you and what they do when they get there, which means the metrics that matter are revenue from search, qualified pipeline from search, and LTV from search, not impressions or positions or organic clicks as an aggregate number. It’s time for measurement to catch up and not use proxies like clicks and rankings.
Users need to be thought of as people with a history of decisions, not as a query. What matters the most is whether you understand your customer at the level of specificity that the offline world has always operated at, and Google has caught up. This is what SEO should have always been about, but now Google is making it a requirement.
[Sponsored by North Star Inbound ]
What happens when you combine best-in-class SEO with conversion optimization?
BigRentz’s traffic increased by 186% (85k), added 1950 conversions in 12 months.
Self Financial’s traffic increased by 50k/month, and added 685 new customers.
Lastly, Secure Data added 1968 phone calls.
North Star Inbound’s SEO strategies earn leads, conversions, and revenue..
Book a call for a free content audit and 10% off any engagement.
Intrigued? [Learn more here]
Every person is different. Someone will search for latest fashion news. Others for investment advice. How are you supposed to fill this template? Producing another hypothetical "Marketing Mary" portrait will not be helpful, will it?