Personal Intelligence in AI Mode: What It Means for Search Results

Maybe you’ve seen the headlines and wonder if Google results are about to look different for every user. The short answer: inside AI Mode, for some people, yes. With Personal Intelligence in AI Mode, Google can now read a user’s Gmail and Google Photos, with their permission. It then uses what it finds to shape the answer they see.

Right now this only applies to a small, paying, opt-in group in the US. But in my view it’s the clearest sign yet of where AI search is going. It changes how you should think about visibility, rank tracking and reporting.

Below I cover what Google actually announced, who gets it, how the privacy controls work, and what it means in practice for businesses and SEO.

What Google Announced on 22 January 2026

On 22 January 2026, Robby Stein, VP of Product for Google Search, announced Personal Intelligence in AI Mode on Google’s official blog. It lets AI Mode (the chat-style part of Google Search) use a person’s own Google apps when it builds an answer.

At launch, two apps can be connected:

• Gmail, which shows up under “Workspace” in the settings, even though Workspace business accounts can’t use the feature.
• Google Photos.

AI Mode runs on Gemini 3 for these answers. Google calls it a Labs experiment, so it is officially a test, not a finished product.

The idea isn’t brand new. Google showed “personal context” for AI Mode at I/O in May 2025. The week before this launch, on 14 January 2026, it also brought Personal Intelligence to the Gemini app. The AI Mode version takes the same idea and puts it inside Search.

What It Looks Like in Practice

Google’s examples are mostly about shopping and travel. That tells you where Google sees the value:

• Shopping: someone asks about trainers and gets a suggestion based on a brand they bought recently. AI Mode found that brand in an order confirmation email.
• Trip planning: AI Mode uses a hotel booking in Gmail and past holiday photos to suggest family-friendly activities near where the family is staying.
• Weather-aware buying: a flight confirmation shows a trip to Chicago in March. So a question about winter coats returns windproof options in the brands and styles the person already likes.

While it works, the screen can show messages like “Checking Connected Apps” and “Applying Personal Intelligence”. Personal sources appear as inline citations with Gmail or Google Photos icons. There is also an option to run the answer again without personalisation.

Who Can Use Personal Intelligence in AI Mode

Access is narrow. A user must meet all of these conditions:

• Pay for Google AI Pro or AI Ultra.
• Be in the United States and search in English.
• Use a personal Google account. Workspace business, enterprise and education accounts are left out.
• Opt in by connecting Gmail, Google Photos or both.

Google said the rollout would happen over the next few days. Eligible subscribers may see an invitation inside AI Mode, or they can turn it on themselves in settings.

So if your customers are mostly outside the US, or most of your audience uses free Google accounts, almost none of them will see this yet. That’s why I treat it as an early signal to prepare for. It’s not a reason to rebuild your strategy this week.

How to Turn It on and Off

If you qualify and want to test it yourself (and I think anyone working in SEO should), follow the path Google gave:

1. Open Google Search while signed in to an eligible personal account, then tap your profile picture.
2. Select Search personalization.
3. Choose Connected Content Apps.
4. Connect Workspace (this covers Gmail), Google Photos, or both.
5. Run a few AI Mode searches you already know well. Compare them with the “without personalisation” version of the answer.
6. To stop, go back to the same settings and disconnect the apps. Google says you can do this at any time.

Testing it on your own account is the quickest way to see how much a personal answer differs from a standard one in your industry.

Privacy Controls: What Google Says It Does and Doesn’t Do

Privacy is the main question most people have, and it’s a fair one. Here is what Google has said:

• It is opt-in. Nothing is connected unless the user chooses to connect it.
• It can be switched off. Connections can be turned on or off at any time.
• No direct training on inboxes or photo libraries. Google says AI Mode does not train directly on your Gmail inbox or Google Photos library. Training is limited to things like the prompts you type into AI Mode and the answers it gives.
• Feedback options. If a personal answer is wrong, you can correct it with a follow-up question or give it a thumbs down.

Google is also open about the feature’s limits. It says the system can wrongly link unrelated topics or misread context. For example, it could treat a gift you bought once as a personal taste, or read an old booking as a future trip. Keep that in mind when you see odd answers in testing.

For businesses, the key point is that users stay in control. People who care about privacy will leave it off. So personal answers will only ever reach part of your audience.

What Personal Intelligence in AI Mode Means for Search Results

Search results have always changed a bit by location, language and device. This is a much bigger step. Two people asking AI Mode the same question can now get answers built partly from their own emails and photos. Nobody else can see those.

Results Become Less Uniform

Picture two people in the same city asking “best running shoes for flat feet”. One has years of order emails from one brand. The other has holiday photos from trail races. Their answers may suggest different products, mention different shops and cite different web pages.

The idea of “the” result for a search was already weak in AI Mode. Now it gets weaker. There is no single answer you can check or screenshot that stands for everyone.

Personal Context Can Push Web Sources Down

Some questions may be answered mostly or fully from personal data. Think of “when does my flight leave?” or “what did I order from that hardware shop last spring?” Those searches never needed your website. But now they happen inside Search instead of Gmail.

For broader questions, personal context works like a filter on top of web information. AI Mode still needs product details, reviews, opening hours and local information from the open web. The personal layer decides which of those details matter to this one user.

Commercial Queries Are Most Affected

Google’s own examples are about shopping, travel and planning. Those are the searches where purchase history, bookings and past trips matter most. They are also the searches most businesses care about. Pure information searches, like how a tax rule works, gain less from someone’s inbox.

What It Means for Businesses and SEO

None of this replaces the basics. It does change where you should put extra effort.

1. Being in the Customer’s Own Context Now Matters

AI Mode uses order confirmations and booking emails to understand what people like. So businesses that already appear in a customer’s inbox have an edge. A brand someone has bought from is a brand AI Mode can see.

That makes your transactional emails (order and booking emails) part of your search footprint, in a way they never were before. A few practical points:

• Use your real brand name the same way in sender names, subject lines and email bodies. That makes the link between the email and your business clear.
• Make order and booking confirmations specific: product names, models, sizes, dates and locations. Avoid vague “Your order #48213 is confirmed” messages with everything else hidden in images.
• Keep product names the same in emails and on your website, so the details match.

Google has not said which email signals AI Mode relies on. So treat these as good housekeeping, not a ranking formula. I’d do this anyway, because clear, text-based confirmations are better for customers.

2. Repeat Customers and Loyalty Become Search Assets

The examples all point the same way: past purchases and trips shape future suggestions. Keeping customers used to be a job for email marketing and CRM teams. Now it also affects how visible you are in AI search.

A local café that people book often, or a clothing brand with loyal buyers, may get suggested to those same people again. That’s a good reason to make sure the first experience is good enough to earn the second.

3. Brand Preference Signals Outside Google Still Count

Personal Intelligence is one of several ways Google lets users’ own choices shape what they see. Google has also been growing preferred sources, which lets people pick the sites they want to see more of in Top Stories. If you publish news or regular content, my guide to getting readers to choose your site as a preferred source covers that side of personalisation.

The common thread: a user’s relationship with your brand is becoming a visibility signal on its own.

4. Your Web Content Still Has to Be Easy to Use

Personal context narrows the choices, but the web supplies the facts. Say AI Mode decides a user wants windproof coats from a brand they already buy. It still needs clear product pages to describe and cite.

So the usual work still matters:

• Product pages with specific details (materials, sizes, uses, prices) written in plain text.
• Correct, up-to-date information about stock, opening hours and locations.
• Pages that answer the follow-up questions a real buyer would ask.

Personalisation may reward the brands people already know. It does not reward thin pages from those brands.

5. New Brands Face a Harder Start

There is a downside. If AI Mode leans on what a user already buys, a business the user has never dealt with has nothing in their history to work with.

For newer or smaller brands, other ways to get found matter more: reviews, local listings, mentions and recommendations on other sites, and paid or social channels that get you that first sale. Once someone buys, you are in their context. If I ran a new brand, this is where I’d put my extra effort.

The Measurement Problem

This is the part I’d flag first to anyone I work with. Personal answers are hard to see, and even harder to measure.

Rank Trackers Can’t See Personalised Answers

Rank tracking and AI visibility tools run searches from clean, signed-out or neutral accounts. They don’t have your customers’ inboxes or photo libraries. Whatever they report for AI Mode is, at best, the non-personal version of the answer.

For users with Personal Intelligence turned on, the answer they see may name different brands or cite different pages. Treat any AI Mode visibility report as a baseline (a starting point), not the full picture.

Search Console Won’t Separate It Out

Google reports AI Mode clicks and impressions inside the Search Console Performance report as part of Web search. There’s no separate AI Mode filter. Google has said nothing about splitting personal and non-personal AI Mode traffic, and no tools for this were announced.

In practice, you’ll see the combined effect in your totals. But you won’t be able to point to Personal Intelligence as the cause. The newer views in Search Console help you spot trends sooner. My explainer on Search Console’s weekly and monthly views shows how to use them for this kind of tracking.

How to Measure Sensibly Anyway

1. Track branded demand. If personalisation favours familiar brands, branded search impressions and direct traffic are useful signs of whether people are getting to know you.
2. Watch commercial landing pages separately. Group product, category and booking pages in your reports. Those are where personal context is most likely to change things.
3. Test by hand on eligible accounts. Run your key searches in AI Mode with personalisation on and off. Write down which brands and pages appear. It’s a small sample, but it’s real.
4. Ask customers. Add “Google AI Mode” as an option in “How did you hear about us?” forms. It isn’t perfect, but it captures what analytics can’t.
5. Be careful with single-search conclusions. One personal answer tells you about one person. Look for patterns across many searches and weeks.

How This Fits the Wider Shift in AI Search

Personal Intelligence arrives in a busy month for AI search. OpenAI has just said it will test ads in ChatGPT, which I covered in what we know so far about ChatGPT ads. The AI assistants are competing on who knows the user best and who can make money from that knowledge.

Google has a natural edge here. Few companies hold a person’s email, photos and search history in one account. Using that data inside Search, even as an opt-in test, is a strong sign of where Google wants AI Mode to go.

Google hasn’t said if or when Personal Intelligence will reach free users, other countries or Workspace accounts. I wouldn’t plan around any timeline until Google gives one.

What to Do Now

You don’t need to panic, and you don’t need to rebuild your strategy. Here’s what I’d do over the next few months:

• If you sell to US shoppers, check your transactional emails for clear branding and specific product or booking details.
• Test AI Mode yourself on an eligible account, if you can, for your top commercial searches.
• Add a note to your reports that AI Mode visibility figures are non-personal baselines.
• Invest in keeping customers and getting reviews. These now support search visibility as well as revenue.
• Keep your product, service and location pages factual and complete, because personal answers still rely on them.

Conclusion

For now, Personal Intelligence in AI Mode is a limited, opt-in Labs experiment: US, English, AI Pro and Ultra subscribers, personal accounts only. But it moves AI Mode from answers shaped by the web to answers shaped by the web and the user’s own life.

For businesses, that means less uniform results, more weight on existing customer relationships, and reports that can only ever show part of the picture. In my opinion, strong pages, clear emails and loyal customers are the best way to prepare.

If you’re planning the year ahead, fold these points into your wider strategy using my practical 2026 SEO plan for small businesses.