Personal Intelligence Didn’t Arrive Overnight — It Compounded

Last month, Google published a piece about “Personal Intelligence”. The idea of an AI assistant that becomes more useful when it’s allowed to connect to your context (your emails, photos, searches and online behaviours).

That’s the headline. But the bigger story isn’t Google, or any single platform.

It’s that personalisation is becoming the default layer across everyday life, and it’s going to feel normal far quicker than most people expect.

When I say “personal intelligence”, I mean this: AI that doesn’t just answer questions, but can act with your context. Not a generic assistant, but one that becomes more helpful the more it understands your world.

The moment it clicked for me

Working in digital marketing, my ear is closer to the ground than most. I’ve watched personalisation evolve from “nice-to-have” targeting to invisible infrastructure behind almost everything we touch.

Back when ChatGPT first launched at the end of 2022, a friend mentioned it to me in passing. I genuinely didn’t understand what it was capable of. At first, my exposure was mostly around writing support — helpful, sure, but not exactly world-shifting.

Then something changed.

AI didn’t just get better. It got baked in. Around March 2025, it became harder to find a mainstream tool that didn’t have an AI layer creeping into the default experience. Not as a novelty, but as a built-in assistant that was “getting to know you”.

That’s the trick of compounding. It looks incremental until it suddenly looks inevitable.

Why it feels like “all of a sudden”

Personalisation has been developing in the background for years. Not in a single big leap, but through thousands of smaller ones:

• More data captured by default
• More of your life stored in the cloud
• Better predictions from better models
• Faster chips, cheaper processing
• Products designed to reduce friction (and quietly increase context)

Compounding doesn’t announce itself. It just keeps stacking small gains until the result feels like it arrived overnight.

So when someone says, “This came out of nowhere,” what they usually mean is: I’ve only just started noticing the results.

Google’s “Personal Intelligence” framing is simply one example of where this is heading: AI that becomes more useful when it can connect across your own sources, not just the open web.

The new normal: your life as context

Here’s where it gets real. This isn’t limited to “AI chat”. It’s weaving into ordinary moments.

1) The doctor’s office (and the end of forgettable conversations)

Consultation note-taking is an obvious one. Increasingly, clinics are using systems that transcribe and summarise appointments, so what was once the doctor’s notes becomes the discussion itself, recorded and searchable. That’s great for continuity, but it changes how conversations feel. There is less room for nuance, and any error can linger.

Real-world moment: you walk out of the appointment and the summary hits your inbox before you’ve made it back to the car.

2) Your calendar and inbox become an active assistant

When AI sits over your email and calendar, it stops being a tool you “go and use” and becomes something that nudges your day. It can summarise threads, flag clashes, and draft replies. Helpful, but not perfect. It often gets you 80% of the way there, then you still need to correct the context.

Real-world moment: it suggests a reply that’s nearly right, but misses the one detail that matters, so you rewrite it anyway.

3) Shopping recommendations shift from “suggested” to “pre-emptive”

Shopping personalisation has been around for years, but it’s getting sharper. It’s no longer just “people also bought”. It’s recommendations tuned to your patterns, timing and habits. The convenience is real, and so is the influence. It doesn’t just respond to desire; it can shape it.

Real-world moment: you’re just browsing and it prompts a “restock” because it’s learnt your cycle. Helpful, or a little too close. Sometimes both.

This is our new reality 

Business meetings, professional consults, and even internal discussions are increasingly recorded, transcribed, and searchable.

And it’s not just your phone or your laptop anymore. The direction of travel is “always-on assistants”: audio capture paired with agents that can take actions based on verbal cues. That’s convenient for some people, and genuinely creepy for others. Both reactions make sense.

Which brings us to the part we can’t skip.

Privacy, trust, and the line between helpful and invasive

I’m optimistic about this space, but I’m not naïve about it. There’s something slightly unsettling about watching the world shift from “searching for answers” to “systems predicting what you need”.

Personalisation only works if people trust the trade. If the user feels like the product is doing things to them rather than for them, it breaks.

Google’s approach (at least in how it’s being presented) highlights what I think are the non-negotiables if this is going to be sustainable: opt-in by default, clear controls, and transparency around what’s connected and when. They also state their model doesn’t train directly on personal sources like Gmail or Photos, and that people can switch personalisation on or off.

Whether you’re a platform, a brand, or a business using AI internally, the same principles apply:

• If you can’t explain it, you shouldn’t deploy it.
• If a person can’t control it, they won’t trust it.
• If it’s “creepy helpful”, it’s not helpful.

And we’ll all need to sharpen our discernment. Not in a fear-driven way, but in a grown-up way. Like many of us, having had to learn how to use the internet safely, we now need AI literacy so the vulnerable and uninformed aren’t the easiest targets for scams and privacy breaches.

A simple default I’m personally leaning into: turn personalisation off until you understand the settings. Before you switch it on, ask yourself one umbrella question: What am I gaining, what am I handing over, and can I turn it off again without losing the basics?

What this means for businesses (and why foundations suddenly matter more)

Because AI doesn’t magically fix messy inputs. It scales them.

One real-world example: if your lifecycle stages are wrong in your CRM, your automation can send the right message to the wrong person at exactly the wrong time, like a “Still deciding?” nurture email to someone who’s already purchased, or a discount offer to a long-term client who should be receiving a loyalty experience. It doesn’t just look sloppy; it damages trust.

The opportunity is huge, but the entry price is boring: clean data, clear definitions, connected systems, and reporting you can actually trust.

The punchline

AI personalisation isn’t coming. It’s already woven into the way we communicate, buy, plan, learn, and get care, and it’s only getting more intimate from here.

The question isn’t whether we’ll use it.

It’s whether we’ll use it with enough intention to stay in control and whether we’ll earn the trust required to use it on behalf of others.

Convenience is coming. Agency is a choice.