1 month with Google’s new ‘CC’ experiment: AI in the inbox

Forgive me, but I’m starting to get a teensy bit weary of The Future™.

Surely you can relate — right? Every week lately, it seems, we’re being bombarded with some new form of AI razzle-dazzle and being told how it’s a zillion times better than the last game-changing innovation and, like, totally gonna revolutionize the way we work.

And yet — well, I don’t know about you, but thus far, most of it hasn’t exactly blown my mind.

Generative AI at this point has been a lot of promise and a lot of “kinda-sorta helpful, maybe?” results. So far, most of the hype hasn’t quite matched the reality, and the high likelihood of a large-language-model-powered system confidently telling you something inaccurate — or even just missing important context around its certain-seeming assertions — is a pretty big liability to accept, particularly in a corporate environment (but honestly, even just in your own day-to-day doings).

So when Google quietly launched another new AI assistant ahead of the holidays, my skepticism meter immediately skyrocketed. But this one looked different. It looked intriguingly familiar, in fact. And it sounded like it might be an effective recipe for putting this rapidly developing technology to use in an environment that actually makes sense — and might lay the groundwork for the genuinely helpful on-demand assistant we’ve all been waiting to meet.

This latest AI-centric creation is called CC, and it’s described as an “AI productivity agent.” Unlike Google’s everywhere-you-look Gemini AI assistant, CC lives entirely in your inbox — within Gmail, specifically — and attempts to work proactively on your behalf in addition to serving as an on-demand info wrangler.

It’s an interestingly different approach and one that immediately reminded me of a short-lived Android-based system from over a decade ago — something that still stands out in my mind as the yet-to-be-matched gold standard for how proactive intelligence should work. And, luckily, I was able to get my hands on this new invention and spend the past several weeks seeing what it’s all about.

Getting to know Google CC

First things first: When you gain access to Google CC in Gmail — available as of now in an early access, waitlist-only setup, limited to individual (non-company-connected) Google accounts in the US and Canada with paid Google AI plans — nothing much changes.

There’s no grand entrance of a glimmering virtual genie, and no new icons or animations or sparkly colors appear anywhere in or around your inbox. You just get an email — and that’s it.

But that email might make you stand up, metaphorically speaking, and take notice — because, at least in my experience, it’s anything but generic.

My welcome email immediately showed how CC taps into your existing Google data to learn all sorts of facts about you, for better or maybe sometimes for worse, and then puts those pieces together to provide extremely personalized communications and suggestions for planning out your days. Specifically, within its third sentence, my CC agent casually mentioned a couple of very accurate but maybe slightly too personal things that it had gleaned from my past communications:

Google CC pulls in personal info to help inform its suggestions — but without human judgment, it doesn’t always know what’s appropriate or likely to be appreciated.
JR Raphael / Foundry

The line between helpful and creepy is a slippery one at times, particularly when it comes to an automated system seemingly knowing things about you on a personal level. For me, the flippant-feeling mention of my mother’s passing went a touch too far and felt like a slight violation, or at least an instance of a non-human algorithm not understanding context or being able to grasp the subtleties of communication.

As for where CC learns about you, it dips into your activity within Gmail, Google Calendar, and Google Drive for its background and then combines that with info from around the web to give you a daily briefing in your inbox each morning. That’s its main function, really, though it also has an on-demand element that we’ll explore in a moment.

But as far as the daily briefing goes, that’s where I can’t help but think of a little something called Google Now. Launched in the summer of 2012, Now was an ahead-of-its-time concept that brought a stream of proactive info onto your Android phone’s home screen — drawing from your location, your typical activity, and your Google account data to figure out exactly what info you might need when and then serve it up to you before you ever asked for it.

That’s almost exactly what CC attempts to do — though, ironically, without all the same data behind it and with a less prophetic-feeling result.

The daily CC briefings

More than anything (and outside of the unfortunate family death mention at the onset), my daily Google CC emails have consisted of:

Summaries and details about events on my calendar for the coming day

A list of tasks taken from incoming emails — things like bills to pay, documents to sign, proposals to review, and software hiccups to investigate

A series of reminders also collected from emails about things like upcoming events, retail offers, incoming order shipments, and impersonal PR pitches

The Google CC briefings hold a mix of information both relevant and awkwardly inconsequential.
JR Raphael / Foundry

Occasionally, the digests have picked up on something I might’ve otherwise overlooked — like a payment due on an insurance policy that had typically been on auto-pay (meaning I barely even glanced at the email that arrived about it) but had for some reason been switched off that status at the start of 2026. On other occasions, they’ve dug into emails and helpfully pulled together bulleted info about upcoming events that I hadn’t thought about in some time — for instance, showing the current costs and benefits of a tech protection subscription that was about to expire so I could think through if and when I wanted to renew it.

But just as often, the system has shown me info that’s fallen somewhere between irrelevant and just plain puzzling. More than anything, it’s painfully apparent the technology is lacking any real perspective on what matters to me and why — so it’ll frequently surface spammy-seeming things that I’d otherwise ignore and present them as if they’re critical knowledge for my day, whether that’s calling out an automated “business loan offer” that landed in my inbox at some point or flagging some unimportant software setup as a top-of-mind, very important item just because a random company whose app I don’t even use sent me a message about it.

As an individual person using this with my own independent inbox (one that combines business and personal emails but that isn’t connected to any broader corporate entity) — and as someone who’s especially on alert and viewing this with a critical eye — it’s easy enough for me to separate the actually relevant info from the context-lacking randomness. But it’s also easy to imagine this presenting some very serious challenges if it were to eventually make its way into a more enterprise-level environment, as noted in a Computerworld analysis published around the time of the feature’s launch:

“Analysts warn that the very compression that makes these tools attractive can introduce subtle risks. Summaries inevitably strip away nuance, and prioritization logic encodes assumptions that may not align with how decisions are actually made inside an organization.”

Looking at examples like my misinterpreted business loan reminder or inapplicable software setup nudge, I can envision this exact sort of quandary if the same sorts of scenarios were to arise in a more complex company context — with someone seeing an important-seeming reminder about something they aren’t familiar with and don’t immediately recognize as irrelevant and then taking steps to follow through on that suggestion.

Even so, I’ve found myself surprised by how much I look forward to opening each morning’s CC email. At this point in my time with the service, it’s proven to be interesting to see what it surfaces and to get a tidy day-ahead summary in a place where I’m already starting my morning, anyway — even if I know a fair portion of what it tells me will be of little to no real consequence.

And that’s still just half of the Google CC experience.

Google CC on demand

The other part of using CC — and one I haven’t found myself tapping into often  — is the ability to interact with CC and reply to any of its emails — or even just email the service separate from any incoming missive — to ask it questions and make custom requests. But my limited use of those features isn’t for a lack of trying.

A few days into my Google CC adventure, I emailed CC to ask exactly what it could do on that front and how I could best take advantage of its interactive potential, and within a matter of seconds, the system responded with a list of easy-to-follow suggestions — even personalizing its text to pull in specific names and scenarios it had found in my emails:

In addition to its proactive emails, Google CC works as an on-demand assistant in your inbox — if you want to engage with it in that way.
JR Raphael / Foundry

I tried some of those suggestions, and they absolutely did work as promised (well, mostly — asking it to draft a reply or new email on my behalf ended up feeling like more work than just writing the thing myself, given how long it took to type out the prompt and then rework CC’s suggestion so it actually made sense and sounded like me). But I couldn’t help but ponder when and why I would actually want to interact with an agent over email in that way.

First, there’s the fact that it just feels weird to be chatting with an AI agent by sending and receiving regular ol’ emails. We’ve been conditioned to expect these systems to show up when we hit a specific button or open up a panel of some sort, and having the system be positioned as another run-of-the-mill contact in your inbox is somehow slightly odd (even having interacted with Gemini in a text messaging app before). Maybe it’s the lack of immediacy followed by the sudden appearance of a human-like reply, but for whatever reason, it seems more unnatural to correspond back and forth with a bot in this way.

Then there’s the fact that Gemini itself is already thoroughly integrated into Gmail and available to perform those same on-demand functions right there in my inbox — and that’s to say nothing of the same system’s integration into Chrome, at the browser level, mere pixels above the inbox view. It comes across as strangely redundant, though it does also seem in line with Google’s goal to stick this sort of AI technology everywhere imaginable.

AI everywhere: the overwhelming redundancy of CC’s on-demand abilities and the same sorts of suggestions from Gemini all throughout Gmail — and beyond.
JR Raphael / Foundry

On a related note, just this month, Google previewed a new AI-in-Gmail function called AI Inbox — which relies on, yes, AI to organize your emails and offer up at-a-glance summaries of your (allegedly) most pressing messages and the tasks within ’em. It doesn’t incorporate info from Calendar and Drive, like CC does, nor does it have the more on-demand elements that CC pulls into the equation. But it feels kind of like where CC might ultimately be headed, if and when the two converge.

Still, it’s important to remember that CC is an “early Labs experiment,” so some of these rough edges and awkward juxtapositions are probably to be expected. As of now, we don’t even know if or when it’ll graduate into a full-fledged, widely available feature or anything that’s offered on the more company-centric Google Workspace side of the spectrum.

For the moment, CC presents a potpourri of potential that feels equal parts promising and problematic. The big question is how it’ll evolve and what it’ll look like in its final form (assuming, of course, that it doesn’t get brushed aside and abandoned somewhere along the way). That’s the question that’ll determine if CC ends up being a place where this type of artificial intelligence feels appropriate and advantageous or just another area where it’s being forced onto us with questionable practical value.

And that, so far, is a question that even CC itself can’t effectively answer. Believe me: I tried.1 month with Google’s new ‘CC’ experiment: AI in the inbox – ComputerworldRead More