Follow ZDNET: Add us as a preferred source on Google.
ZDNET’s key takeaways
- Gmail search doubles as a powerful AI command center.
- AI can quickly find contacts, history, and draft follow-ups.
- AI turns tedious inbox work into a rapid workflow process.
In this article, I’m going to show you how I used AI in Gmail to save myself hours of tedium and annoyance. In Gmail. No kidding. Really.
Gmail was announced on April 1, 2004. Back then, email services like Hotmail and Yahoo Mail provided 2MB or 4MB of storage. Google launched Gmail with 1GB of free storage capacity. Since that was 250 times the amount Yahoo offered and 500 times what Hotmail offered, the announcement was widely viewed as an April Fool’s joke. But Gmail was no joke.
Also: Inside Google’s vision to make Gmail your personal AI agent command center
While storage capacities have grown since then (my account has 2,048 GB), much of the Gmail interface remains almost identical to what was launched in 2004. Over the years, Google has released some gimmicks, like the ill-fated Inbox by Gmail offering from 2014 and the mostly useless AI Inbox launched this year.
Even though Google has been actively embedding AI into many of its properties, Gmail AI features mostly feel like poorly integrated patches on top of an aging interface. For example, Gmail plasters AI overviews at the top of messages, a feature that you can’t turn off. I’ve had many long conversations with folks who hate this intrusion into their email flow.
But last week, I had to work on a big, tedious project. I decided to try an AI hack with Gmail to see if I could save some time. What happened blew me away. I had one of those very rare moments where it’s 100% clear that I’m actually living in the future. It’s also fitting that I’m publishing this article on April 1, 22 years to the day since Gmail was first announced.
Also: Want to change your Gmail address? Now you can, without losing any data – here’s how
What Gmail did in this hack was personal to my email flow and project. To understand just how gobsmackingly insane its behavior was, you need to understand my project. Give me a few minutes to share the backstory, then I’ll show you what I did with Gmail’s built-in AI. I’ll also show you how you can make this hack work for you.
My project
About a year ago, I published an article reviewing the best AI web builders. This was a huge project, taking 70 hours overall. Much of that time was spent wrangling hosting providers who claimed to offer AI, sending and replying to 236 emails to acquire and set up test accounts on their services.
The article has turned out to be a good performer. My editor wants me to update the piece for this year. About a month ago, I sent out a set of emails to the five finalist companies I spotlighted last year. Over the past 30 days, I’ve interacted with most of these companies several times, creating a bunch of new email threads.
Also: Looking for an AI-powered website builder? Here’s your best option in 2025
The challenge with a project like this is that you send an email and you wait. So, after a spate of back-and-forth email messages, I went on to other work. Last week, I decided it was time to review the current correspondence and see if I was ready to start testing.
The normal way I have had to do this sort of thing in the past is to dig up the names and email addresses of the folks I’m corresponding with on the project, which is usually fairly easy because I keep the list in an Apple Note. But then I have to go through all the email threads for each person to figure out each company’s status.
This is time-consuming, tedious clerical work that, to be blunt, I truly dislike doing.
That’s when the lightbulb went off. Google’s always talking about its AI in Gmail. Maybe I could use this technology to find a better way.
Also: I tested ChatGPT vs. Claude to see which is better – and if it’s worth switching
I did, and it worked. I’ll walk you through what I did. I’m going to share screenshots of the real conversations with Gmail and the vendors, although I blurred some of those interactions because the details of my correspondents and their conversations aren’t meant to be public.
The Hail Mary prompt
American slang uses the phrase “Hail Mary pass” to describe a wildly desperate football pass, typically tried when all other saner options are unlikely to succeed. The expression goes back at least to the 1930s. It has been adopted to mean a desperate action done with little expectation of success when there are no apparent better alternatives.
I wrote the following prompt as a Hail Mary pass to get me out of digging through hours of email threads. I did not expect it to work:
Take a look at my emails in the last few months and let me know which of these vendors got back to me and provided test accounts and which have not. Give me status updates for each of the five vendors.
I appended the vendor names and email addresses to the prompt.
Also: 7 AI coding techniques I use to ship real, reliable products – fast
Now, here’s where it gets interesting. Instead of opening a sidebar or otherwise trying to find the AI interface, I just pasted the prompt right into the search bar at the top of Gmail.
What I got back was an AI overview that was genuinely very helpful:
Look at that. First, you can see that the prompt was just pasted into the search bar on top.
Second, in one place, I have the status of all five vendors, the logins, and even the coupon codes to set up what I need to update the article. GoDaddy was on hold, but I corresponded with the contact to get that test set up. Without digging through threads, I identified the vendor that had been unresponsive.
I mean, wow.
More follow-up magic with Gmail AI
My project was pretty well on track for four of the five vendors. But the fifth vendor was still off track. I didn’t know whether my contact had left the company or was just unresponsive. So, I decided that I’d reach out to everyone I knew at the vendor to see if I could get a response.
I put this prompt into Gmail’s search bar. I said, “What contacts do I have at [company] and what’s the date of their most recent contacts with me?” I’ve redacted the company name, but here’s what Gmail returned:
That output took like a minute. And I don’t mean “a minute” as slang for a long time. I mean less than 60 seconds.
Sure, Gmail has search. But scanning through all the emails, separating the marketing promotions from the company, and finding the contact information would have been an annoying process at best. Instead, I typed in one short prompt and got this result back instantly.
Also: How AI has suddenly become much more useful to open-source developers
See what I mean about living in the future? This is Gmail of all things.
Then I took the process one step further. I had Gmail dig up the original email I sent to the first contact and prepare a message for all the others. I prompted:
Okay, find the email I most recently sent to [name] about the new review opportunity. Prepare that to be sent to the [company] contacts I interacted with in 2025, saying I sent the original message on (insert date), haven’t heard back, and if they want to be included in this update, they’ll need to get back to me soon.
Gmail gave me back this construct in the sidebar:
I found that suggestion a little cramped, so I hit the Edit in Gmail button. Here’s what Gmail prepared for me:
The draft was exactly what I needed, so I just hit Send.
Gmail in 2026
By using AI prompts in Gmail, I cut hours from my work process on this project. All told, I spent about 10 minutes from start to finish. Most of that time was taken up by my awe at what Gmail just did with a few simple prompts.
Apart from this project, I always write my own emails. But I have to say that for processing and aggregating follow-ups on the status of a multi-threaded project, the AI in Gmail was simply amazing. I still have no plans to use the AI to write my daily emails. But if I have another project with a ton of moving parts driven by email, I will undoubtedly use it to save some time understanding and aggregating project status. If the next try works as well as this run did, it’s a win.
Also: What Google’s TurboQuant can and can’t do for AI’s spiraling cost
Obviously, your use of AI in Gmail will be different than mine, simply because you’re you and I’m me. That said, consider giving Gmail assignments for projects where you need to cull through a bunch of messages. When you do, provide the app with as much information as possible so it can have a better chance of success. Of course, don’t assume Gmail will get everything right. But definitely keep this approach in mind for tedious thread-scanning projects.
We now have a new superpower. It’s up to us to use it intelligently and responsibly. And that’s no joke.
You can follow my day-to-day project updates on social media. Be sure to subscribe to my weekly update newsletter, and follow me on Twitter/X at @DavidGewirtz, on Facebook at Facebook.com/DavidGewirtz, on Instagram at Instagram.com/DavidGewirtz, on Bluesky at @DavidGewirtz.com, and on YouTube at YouTube.com/DavidGewirtzTV.
