Will the Next Version of Google Chrome Change What Websites Are?
Google Chrome’s share of the browser market is (as of August 2025) 69% (Statcounter). That’s the highest it has ever been.
That means that when Google does something different with Chrome, it impacts more than two-thirds of the web-using global population.
So, when Google announces “the biggest upgrade to Chrome in its history”, it’s worth taking notice.
The new update announced last week focuses on the inclusion of a feature called Gemini in Chrome.
Chances are you can’t currently access it; it’s only available to English language users located in the US who are subscribed to the Google AI Pro and Google AI Ultra plans.
It will, however, roll out to the rest of us eventually, starting with Google Workspace business users, “in the coming weeks”.
What Gemini in Chrome does first
Right now, Gemini in Chrome will be all about information retrieval and summary. It can take the information from a complex web page and summarise it for you.
It can also take information from across multiple open tabs and consolidate them into a single summary for you. The example given is when planning a trip, Gemini in Chrome will pull together an itinerary from all of your open tabs.
A more mundane, but arguably highly useful feature is that it will scour your web history to find a site you are looking for. For example, when you find the perfect company to make your child’s birthday cake but can’t remember the name, you can ask Gemini in Chrome, “Who was the baker who created the life-size Elsa cake?”
It will also let you interact with the content of a page you are viewing. Things such as jumping to a specific point in a YouTube video, or asking a question about the article you’re reading.
It’s all very useful stuff, but it’s what is to come that has the potential to change the way we use the web.
Becoming an agent
“In the coming months, we’ll be introducing agentic capabilities to Gemini in Chrome,” says the announcement.
“These will let Gemini in Chrome handle those tedious tasks that take up so much of your time, like booking a haircut or ordering your weekly groceries.”
This is the really interesting part of Google’s announcement.
What it’s suggesting is that Chrome will, at your request, visit websites and complete actions for you.
It’s this that I think has the potential to change the way websites are built.
Take the haircut example; if Chrome is going to be able to book a haircut for you, it will need one of these two scenarios to be true:
It can essentially visit any website, figure out how to make a booking (could be a form, an email address, phone number, etc), and then book
Or it uses a standardised approach; something like a feature built into Google Business, or a standard form implementation on the website that it knows exactly how to use
I don’t doubt that eventually, AI agents will be able to visit any website, figure out what it needs to do, and then take action, however, will the very fact that this will become a thing incentivise further standardisation across the web?
Standardising the web
Right now, a website can be built using an almost infinite number of approaches. You could create a picture in Microsoft Paint, host it online, and call it a website.
You could equally build an entirely bespoke system that works just for you and your organisation.
Because websites are built for humans, emotional impact is, whether you like it or not, very much ‘a thing’.
Gemini in Chrome, though, presumably won’t have an emotion-layer.
If it’s booking you a haircut, it’ll be more interested in the booking process than the hairstylist.
Does this mean that hair stylists will need to standardise their booking procedure, or risk upsetting Gemini in Chrome’s (presumably) methodical approach?
I think yes. I think that, regardless of how good AI agents get at navigating the minestrone soup that is the web, the winners will be those who pave the way for agents.
A WordPress site held together with a patchwork of badly maintained third-party plugins will potentially be at a big disadvantage when compared to a ‘headless’ site built by the most efficient team on earth.
What could happen?
I’m predicting the future here, so there’s a good chance I’ll be entirely wrong, but I think we’ll see the following things begin to happen.
1. More companies will converge on standardised technology
AI agents will surely favour efficiency, and the key to efficiency online is standardisation. I think certain elements of the web will become increasingly standardised. Methods of making contact and ways of presenting data will all benefit AI agents if they follow some pre-agreed ground rules.
2. One web for the agents, one for the humans
I also think we may begin to see ‘agent versions’ of websites. In the same way that a website can be served in different languages in different locations, I think we may start to see the creation of agent versions of sites. These sites will be able to deliver user journeys optimised for an agent, making it simple for bots to interface with a business.
A lot has to happen before any of this becomes reality. Not least do people need to make use of Chrome’s new agentic possibilities. Right now, web users just aren’t used to handing over actions to a bot and letting it run its course.
I think the Gemini in Chrome announcement marks a pretty significant milestone in the journey of the web. It is a step in making AI agents available to the masses and over time, could drive real behavioural change (if you’re interested in that concept, check out Pat’s recent article on the Law of Behavioural Shift).