I tried the Google Pixel 9’s AI features. Here’s why you should use them

My mileage has been quite disenchanting so far, and it seems you should not buy into all those claims until Apple releases all the AI tricks that it has promised so far. That wait could extend well into 2025. On the other hand, you have the Google Pixel 9.

It’s a phone that is truly struggling with its AI identity. It wants you to buddy up with a chatty AI. It will put AI into your images in hopes of making Photoshop tricks look ancient. It wants to be your copilot at work stuff, too, such as writing mail responses or brainstorming a draft.

Aside from absolutely mundane stuff like setting up an alarm, the Pixel 9 wants to put AI into almost everything you want a phone for. Even screening and attending calls are not off the table here. AI everywhere. That’s the core premise.

But that doesn’t mean that AI deployment is necessarily meaningful. Or even accurate. I’ve been pushing the AI chops of the Pixel 9 ever since I got my hands on the phone, and if you’ve been wondering if the phone’s worth its AI salt, here are the AI tricks that just might win you over.

NOTE: The Pixel Screenshots app is exclusive to Pixel 9 phones. The remaining features are tied to Gemini Advanced, which is free for a year on Pixel 9 Pro devices and requires a Google One AI Premium subscription (starting at $20/month) with many other features in tow.

Gemini in Gmail

Unsurprisingly, Google is pushing Gemini in every corner of its product ecosystem. Gmail is one of the most ubiquitous, just after Google Search.

Google isn’t exactly the first name to experiment with AI for emails. Shortwave was among the first to take the plunge and continues to do an excellent job. Apple Intelligence has also pushed a summarization system and Smart Reply feature to mail, but it’s hit or miss.

Gemini offers the best integration of AI I’ve seen in any mail app so far. Though it exists separately as its own actionable icon at the top, right by the search bar, it can handle your queries more intuitively.

Let’s say the only thing you remember about an email was that it was about some keyboard, but no detail about the brand name or the sender. A query like “What was the email about the compact keyboard I got in July?” does the job just fine.

But that’s just the start. With a little bit of detail sprinkled into the prompt, you can write an email response without ever opening the email. You can say, “Send a reply to Jessica, telling her I like the idea of a bag with an embedded tracker.”

Gemini will not only craft a response but also pick up a handful of other details, like the product name, launch, features, etc., from the appropriate email and craft a well-written professional response. Alternatively, you can simply ask it to write an original email.

Going a step further, Gemini actually makes things even more convenient when handling an overflowing inbox. The summaries are mostly accurate and far better than Apple Intelligence or any other software I’ve used so far.

Moreover, these summaries work just fine for single emails and entire threads. However, I recommend tapping on the “View longer summary” option, as it does a fantastic job of picking up the overarching tone and the tiny details that matter. The “Suggest reply” system also does a tremendous job.

Then there’s the analysis part. I regularly receive scientific research papers in my box. Gemini lends a helping hand here as well. Gemini Advanced offers a context window worth a million tokens, which is good enough for a 1,500-page book.

I fed the AI a 60-page paper about the history of caste discrimination and asked it to summarize one particular aspect. Within about seven or eight seconds, Gemini parsed the whole PDF, found the relevant bits, and presented them in a neatly formatted article. It’s almost magical.

Gemini extensions

One of Gemini’s most underrated aspects is its integration with Google’s suite of Workspace tools, including Docs, Gmail, and Drive. The integration extends to YouTube, Google Flights, Maps, and Hotels.

There are two ways you can make use of extensions. You can use a shortcut specifier in your prompt, such as @Gmail (just like you would nudge a teammate in a group chat on apps like WhatsApp or Slack), to ensure that Gemini executes your query in that particular service.

If you’re inclined toward voice commands, you can mention the app where the AI should look for the relevant information. Even if you don’t, Gemini can pick up details or titles of a specified file in your Workspace portfolio and do what is needed.

For example, you can start with basic commands like “summarize the most recent email from Professor Lucien.” Or, you can dig deeper. I asked it to summarize an article in Google Docs, giving it only the title of the draft, and Gemini picked up all the core points and presented them in bullet list format.

Gemini also looked into my inbox and gathered information about my upcoming flight. The response also integrates quick shortcuts from different emails, each containing unique information such as a copy of the ticket, terminal details, and web check-in itinerary.

It can also handle queries across entertainment platforms such as YouTube. But do keep in mind that it won’t get granular work across another Workspace app, even if it’s something as basic as adding a track to a specific playlist on YouTube Music.

For now, Gemini Extensions aims to quickly give users information they have saved across Google’s ecosystem services. From spreadsheets to PDFs saved in your Google Drive, the AI assistant will pull up all the requisite information.

Once it has access to the requisition file, you can choose how it is presented — including as a summary, in bullet points, or have it spoken to your ears. All the options are on the table. Parsing for the right files(s), especially without the right Workspace shortcut, can often take a few seconds, so it’s not perfect.

But the sheer potential here is huge. Imagine third-party apps opening up to Gemini Extensions and creating actionable shortcuts. Google has already laid the foundations, and it’s only a matter of time before we see eager developers lap up the idea.

I just hope that when it eventually becomes mainstream, the process will be transparent, and data security will not be an afterthought. That’s a big if right now, so here’s to hoping.

Pixel Screenshots

Screenshots are the new bookmarks — or so I tell myself. For a more scientific peek into their utility, Christopher Moore wrote a fantastic research paper called “Screenshots as Virtual Photography” in the Advancing Digital Humanities journal. In it, he argues that they are directly linked to the dynamic live sensory experience of digital consumers.

For humans sharing screenshots online, they also play a crucial role in fostering social connections and curating the online cultural experiences you indulge in. But let’s stick to the less scientific, and more regular, stuff. We often take screenshots as some sort of memory diary. It can be anything from a ticket or professionally valuable information to just a fun meme.

We partake in the chore of screenshotting because it’s convenient. Yet, we hardly ever go back to that “screenshots” folder. Part of that reluctance has to do with the fact that it’s tedious, especially when you can’t recall what month or year the screenshot was taken or how deep you have to scroll to find the one you’re looking for.

The Pixel 9 brings a feature that essentially acts as the wise keeper of your screenshots. Eerily wise, I’d say. The secret sauce here is Pixel Screenshots, an app powered by the on-device Gemini Nano. Think of it as an OCR tool with AI superpowers, one that exists more like a guardian of memories and the knowledge hidden in your screenshots.

It just works, and in more ways than the app’s name suggests. I’ve got a few thousand screenshots stored on my phone dating at least five years in the past. Yeah, I’m a hoarder, and cripplingly afraid of losing something valuable if I clean up the screenshots folder. But all that hoarding now has a purpose.

With the Pixel screenshot app, I was able to pull all kinds of information. From abbreviations on a game controller setup page to describing a game character flexing their gigantic sword in a cut scene, the Pixel Screenshots app does a darn good job at absorbing all kinds of information and then surfacing it.

It’s a fantastic tool for not just finding the serious kind of information, like checking the due for my March electricity bill, or the Starbucks coupon that I lost to the washing machine. For photos, with humans and pets in them, it works fairly well.

The app processes the natural language inputs, or even short descriptors, in real time as it parses through the gallery of relevant screenshots until it lands on the exact match. For a person who often screenshots articles in the Discover feed or social media feed for future reading, Pixel Screenshots is a convenient tool that acts more like a super-smart memory bank.

It’s worth revisiting, time and again, owing to its accuracy. I’ve rarely seen an AI tool being this precise at its output as the Pixel Screenshots app. It’s a sleeper hit, one without any functional caveats or dramatic pitfalls. The icing on the cake is that as you land on the desired screenshot, you can add a calendar entry for it, and slot it in a new or pre-made folder.

There’s a lot of AI promise

I’ve never been a fan of the generative AI hype, and for good reason. For the most part, it seems fun, but in a rather useless fashion. Pixel Studio text-to-image generator is a great example. Apple’s attempts at a similar tool are nothing more than a flex.

But putting AI where we can get valuable work done is difficult. Gemini, in whatever little ways it has managed to influence a few core products, has done fantastic. I never thought I would say this, but I think the best place to put AI in your workflow is the email inbox.

Gemini is really good at this task. The Pixel Screenshots app, on the other hand, is not just a wizard for your memories but also a reliable route to finding information that you may have saved as a screengrab and forgotten.

Extensions, though limited in their current form, have immense scope for transforming how an AI assistant should interact with other services that are part of your daily phone habits. Overall, I think Gemini makes the Pixel 9 worth buying on its own, as it truly redefines how you interact with your phone. That’s a leap worth looking forward to and a testament to just how good Google’s AI tools really are.