In the past few weeks, it\u2019s become clear that browsers are the latest entrant in the AI arms race. We\u2019re talking about things like Perplexity\u2019s Comet, ChatGPT Atlas, and even Chrome \u2014 browsers that natively embed chatbots into the internet experience. The pitch is to reorient how we browse, to move us away from the search engines that have reigned for the past three decades. The central idea is the same as we\u2019ve heard from all the other agents-all-the-way-down companies: AI will be just as good as you are at surfing the web. Possibly better.<\/p>\n
Right now, AI browsers come in two main flavors. There are your regular browsers that have an AI assistant stapled on in a collapsible window, such as Chrome with its Gemini features<\/a>, or Edge with Copilot Mode<\/a>. Then there are more specialized AI browsers, most notably ChatGPT Atlas<\/a>, Perplexity\u2019s Comet<\/a>, and The Browser Company\u2019s Dia<\/a>. This second category often supplants your search bar with AI and sometimes includes an \u201cagentic mode,\u201d in which the AI can complete more complex, browser-related tasks for you. Theoretically, that includes helping you book reservations or add items to a shopping cart.<\/p>\n While AI browsers share a similar approach, they each have varying takes on the ideal web surfing experience. Some make you pay for certain features, and of course, there are differences in the underlying models. But this isn\u2019t meant to be a ranking. For this piece, I\u2019m evaluating whether AI browsers can currently deliver a better internet. So I decided to focus on three main criteria:<\/p>\n When are AI browsers most useful? I\u2019m looking to see which, if any, browsing tasks become easier or faster by adding AI.How much prompt babying is needed? Theoretically, I shouldn\u2019t have to craft an overly specific prompt or answer a zillion follow-ups to get the result I\u2019m looking for. Google is good at figuring out what you meant to type \u2014 I expect the same from ChatGPT.If there is an agent, do I trust it to complete tasks for me? The whole point of AI agents is to let them do things for me. You need high confidence that the results are trustworthy.<\/p>\n For testing, I decided on a few ground rules. I kept it to five browsers: Chrome, Edge, Atlas, Comet, and Dia. There are more available, but this felt like a representative mix of both AI browser categories from a variety of players in the field. I focused on desktop apps, and tried to make settings as uniform as possible: I generally instructed the AI browsers to keep answers snappy, shared my location information where possible, enabled memory settings, and described myself as a \u201ctech journalist specializing in health and wearable tech.\u201d I also approached testing from a variety of AI skill levels. What would results look like if I was a complete AI newbie versus someone more adept at prompting? Lastly, if I tried one task in a browser, I gave it a go in all the browsers, down to the same exact prompt.<\/p>\n Ultimately, my question was not which AI browser you should use, but whether any of them are worth your time and energy. This was a journey to see whether any of them live up to the hype.<\/p>\n The short answer: they don\u2019t.<\/p>\n Stapling an AI assistant to a browser doesn\u2019t magically redefine how you interact with a chatbot. It\u2019s more like hanging out in person, rather than texting. You\u2019re having the same conversation, just in different formats, each with their own pros and cons. But no matter the browser, I kept running into the same fundamental problem: you have to think extra hard about how to craft the right prompt.<\/p>\n That\u2019s the opposite of how search, particularly Google, has evolved. At Google\u2019s peak, you could type in a string of misspelled words into the search bar, and somehow you\u2019d still get the right answer. AI models require a lot more prep and guidance.<\/p>\n At Google\u2019s peak, you could type in a string of misspelled words into the search bar, and somehow you\u2019d still get the right answer.<\/p>\n Take the universal torture of sorting your emails. On any given day, I want to know what my most important emails are, and which ones I need to respond to ASAP. The first few times I took a crack at this task, I asked the various browsers to summarize my emails. (I know \u201csummarize my emails\u201d isn\u2019t a stellar prompt, but it\u2019s often a default suggestion. And defaults exist because they should be generally helpful.) All I got were literal descriptions of the emails in my inbox. In my personal inbox, it said I had one thread in my primary folder, listed the subject, summarized the preview, and then stated it was \u201cdated Nov 20th, and was marked not starred or important.\u201d<\/p>\n I tried refining my ask. Instead of \u201csummarize,\u201d I prompted AI to \u201cidentify important emails based on urgency.\u201d In my work inbox, that generated a list of not-important, not-urgent email threads because the models have no idea what I actually find important. I wanted reader feedback, pitches from trusted contacts, or threads I\u2019d forgotten to reply to. What I got instead were irrelevant pitches, mostly for health scams.<\/p>\n I made zero headway until Comet suggested the prompt \u201cfind important unanswered emails.\u201d The top four emails that surfaced were littered with important keywords for a tech journalist \u2014 Urgent! Embargo! Exclusive for The Verge! All had multiple follow-up requests. You can see why Comet would think they mattered, but after reviewing them, all were emails I didn\u2019t need to read at all, much less reply to. AI had fallen for the oldest trick in the book: conflating keywords with truth.<\/p>\n I was ready to write off the experiment when I noticed Comet\u2019s AI had buried the lede. Nestled three-quarters of the way into its long-winded summary was a bullet point labeled \u201cPersonalized requests\/follow-ups.\u201d It highlighted two emails: one from a CEO addressing feedback I\u2019d made in a recent product write-up, and another from a reader with a tip relevant to my beat. Neither was \u201curgent\u201d but both merited a closer look.<\/p>\n I tried Comet\u2019s \u201cfind important unanswered emails\u201d prompt in the other AI browsers. They all highlighted other previously skipped, keyword-stuffed pitches. None flagged the two emails I was interested in. So I tried even harder:<\/p>\n Find unanswered emails in which I had previously responded with interest or feature personalized requests\/feedback. Then, evaluate which ones I should respond to based on timeliness and keywords such as \u201cembargo\u201d featuring dates in the next two weeks. Ignore emails with multiple follow-ups to which I have not responded. <\/p>\n This went slightly better with Comet and Dia. Both surfaced multiple relevant email threads, but only one ultimately required a response. Copilot in Edge highlighted one relevant thread and five junk pitches. Gemini in Chrome was a dud: It surfaced only a Black Friday marketing email.<\/p>\n In Atlas, ChatGPT merely replied, \u201cIt looks like Gmail successfully returned the unread message IDs, but the actual content for those messages didn\u2019t come back \u2014 the batch read returned empty, which means the Gmail API didn\u2019t provide the email bodies this round.\u201d It proceeded to ask two long-winded follow-ups. At this point, my options were to refine my prompt further or give up.<\/p>\n Emails were mostly a failure, but there were some daily tasks where AI browsers were alright. I had to search through a 48-page legal document for a family matter, and while CMD-F is tried and true, the legalese made my brain melt. So I loaded the document in a tab and prompted the AI browsers to list all the relevant pages and sections, with an accompanying summary. All the browsers returned the same pages, with slightly different summaries. I still had to do my own reading, but it got me to a useful jumping-off point faster.<\/p>\n These browsers can also work well for internet search \u2014 provided you\u2019re patient enough to reprogram 20 years of Google muscle memory. Where AI search works best is answering questions about the exact site you\u2019re on. While pondering a phone upgrade, I asked the bots to compile various iPhone specs and size dimensions into a table from Apple\u2019s website and the wider web. That was much more convenient and helpful than flipping through multiple tabs. At the end of the process, I was much more confident about which iPhone I was upgrading to.<\/p>\n I was much more successful whenever I shifted my mindset to \u201chow can AI help me interact with this page?\u201d<\/p>\n Whenever I went into a task asking AI to do things for me, I\u2019d end up frustrated. I was much more successful whenever I shifted my mindset to \u201chow can AI help me interact with this page?\u201d For example, I was trying to parse a clinical study, and hit a particularly technical paragraph written in dense medical jargon. Asking the models to summarize and explain some concepts I was iffy on in plain English was helpful.<\/p>\n Summarizing or compiling data like this was the most convenient part of using AI browsers. All the browsers do this fairly well, and it\u2019s a useful thing to have at your fingertips \u2014 it\u2019s not without the occasional back-and-forth, but overall, I needed less time and fewer tabs to get to a point where I\u2019d take over the heavy lifting of getting something done online. I\u2019m always in favor of fewer tabs while browsing.<\/p>\n We already know that AI is good at summarizing and compiling, though. Complex queries are where these browsers are meant to shine. Here, too, one must grease both elbows and wrestle AI into submission.<\/p>\n Ahead of the Stranger Things season 5 premiere, I was chatting with a colleague about watching an 18-minute YouTube recap video<\/a>. They were separately working on another AI project and asked if AI browser assistants could turn videos into downloadable, .txt transcript files. So I tried prompting: can you rip a transcript of this YouTube video?<\/p>\n 1\/2Note how Dia says it will export the full transcript. Screenshot: Dia<\/p>\n Copilot said no, on account of the video\u2019s copyright. (Never mind that most YouTube videos already have transcripts, right there on the page. This should not be a hard problem.) What I could get was a summary or an outline of the video\u2019s content. Comet ripped an accurate transcript for the first 25 seconds before stating that the \u201ctranscript continues for Seasons 1-4 with detailed plot and character recaps.\u201d Dia gave a time-stamped transcript, but only for the first 15 minutes. Atlas and Chrome were the only two to give full transcripts. As in, an extremely long, line-by-line transcript right into the chat window.<\/p>\n Next, I asked each browser\u2019s AI if they could turn that transcript into a downloadable .txt file with timestamps. Only Atlas completed the task. The rest said generating a downloadable file wasn\u2019t in the cards, but I could copy-paste the plain text into a file myself.<\/p>\n So much for \u201cjust telling the AI what I want.\u201d<\/p>\n After several detours, I returned to my original task: figuring out which pair of New Balances to buy, and finding the best deal possible.<\/p>\n When I say I wanted a pair of New Balances, it\u2019s because I\u2019ve spent about three months researching. I look at social media, I ask friends, I read up on the history of various brands before ultimately choosing one. Then I\u2019ll spend a few hours on that brand\u2019s website whittling down my options until I have about three. Afterwards, I\u2019ll try to find a deal online. It\u2019s a long, arduous process prone to human error. Hence, why it\u2019s been two years since I set out to find a pair of durable, stylish, and comfy walking shoes and I\u2019ve yet to find one.<\/p>\n With AI browsers, the research part was \u201ceasy.\u201d In a nutshell, I had to give it highly specific research prompts. That meant telling it that I\u2019m: flat-footed, more comfortable in wide shoes, looking for a lifestyle sneaker and therefore no running shoes, looking for something that can easily handle 15,000 to 20,000 steps per day, interested in a versatile color but prefer a neutral white, wanting something that works with athleisure and elevated street wear, and not looking to spend more than $120 (but would prefer to stay under $100).<\/p>\n Because I didn\u2019t specify \u201cclosest store near me\u201d in my original prompt, I\u2019m being asked to pick which store. Screenshot: ChatGPT Atlas<\/p>\n What ensued were multiple back-and-forths where the browsers both did and didn\u2019t listen to my needs. The longer the response, the more likely I\u2019d get contradictory advice \u2014 here\u2019s a $200 ultra-performance running shoe with a carbon plate as your top rec, but at the very bottom, here\u2019s an $85 model that has more of a lifestyle feel in the completely wrong color. Rinse and repeat. After roughly 20 rounds across five browsers, I arrived at the New Balance 530.<\/p>\n The 530 had also made it to my short list when I\u2019d done the process manually. But while I was faster at narrowing down New Balance models on my own, AI had provided reasoning behind each choice. Things like, this model has extra cushioning for durability or the silhouette in that model would work with multiple outfits. My picks were mostly based on vibes.<\/p>\n Enter phase two: finding a deal. I asked all five browsers to find me the lowest price on a pair of New Balance 530s in all white, white-and-silver, or white-and-pink, in a women\u2019s 8.5 (25cm), that\u2019s in-stock in my zip code. If there was an agentic mode, I asked the AI to put it in my cart.<\/p>\n This was the third attempt to close the pop-up. Screenshot: ChatGPT Atlas<\/p>\n Cue several more back-and-forths, in which I got differing results. Dia, Comet, Chrome, and Edge found the same local Foot Locker, but selected different colors. Atlas was able to finally put the right pair in my cart, but not without checking in several times to make sure I really wanted to. It also tried to override my preference for pickup and switch to delivery. Once, I watched Atlas spend a minute trying to close a pop-up just to get back to shopping.<\/p>\n I ran the full experiment several times, and each time, I was sure the browsers were finding the best price on a given day. However, I became less and less confident that these were the shoes I really wanted. Especially when Atlas threw in the New Balance 574 Core as an alternative, because they were \u201cone of NB\u2019s most iconic everyday silhouettes\u201d and were a versatile, androgynous shoe. (ChatGPT knows I love unisex styles.)<\/p>\n If I\u2019m judging these browsers on the premise that AI could be better than you at surfing the web, that simply isn\u2019t the case. At no point would I have considered the experience \u201chands-off.\u201d But more broadly, my whole AI browser experience reinforced that I spend a lot of time doing things for AI so that it can sometimes do things for me. I\u2019m changing the way I think, the way I word questions, and the way I search and digest information. It\u2019s less about how AI fits into my life and more about how I can adapt what I do naturally to accommodate its growing presence.<\/p>\n A good experience with these browsers assumes a lot of things. So does googling, but after 20 years, it requires much less mental effort than the best of what AI browsers currently offer. With AI browsers, you\u2019ve got to be fairly adept at prompting. You\u2019ve got to understand the strengths of chatbots \u2014 and be patient enough to work around their weaknesses. Or, at the very least, you have to be willing to learn. This is true for many people. But I\u2019m not confident that anyone who downloads an AI browser will find the learning curve worth it.<\/p>\n AI can sometimes be useful, but it\u2019s always a lot of work. And I still need new shoes. I\u2019ve decided to just visit a New Balance store in person.<\/p>\n Follow topics and authors from this story to see more like this in your personalized homepage feed and to receive email updates.Victoria SongClose Senior Reviewer, Wearable Tech<\/p>\n Posts from this author will be added to your daily email digest and your homepage feed.<\/p>\n FollowFollow<\/p>\n See All by Victoria Song<\/a><\/p>\n AICloseAI<\/p>\n Posts from this topic will be added to your daily email digest and your homepage feed.<\/p>\n FollowFollow<\/p>\n See All AI<\/a><\/p>\n ReviewsCloseReviews<\/p>\n Posts from this topic will be added to your daily email digest and your homepage feed.<\/p>\n FollowFollow<\/p>\n![\"<em>Note](\"https:\/\/www.newsbeep.com\/ie\/wp-content\/uploads\/2025\/12\/Screenshot-2025-11-24-at-5.30.24\u202fPM.png\"\/)
<\/a>![\"<em>Semantics](\"https:\/\/www.newsbeep.com\/ie\/wp-content\/uploads\/2025\/12\/Screenshot-2025-11-24-at-5.31.54\u202fPM.png\"\/)
<\/a>PreviousNext<\/p>\n![\"Screenshot](\"https:\/\/www.newsbeep.com\/ie\/wp-content\/uploads\/2025\/12\/Screenshot-2025-11-24-at-5.13.11\u202fPM.png\"\/)
<\/a><\/p>\n![\"ChatGPT](\"https:\/\/www.newsbeep.com\/ie\/wp-content\/uploads\/2025\/12\/Screenshot-2025-11-13-at-11.21.26\u202fAM.png\"\/)
<\/a><\/p>\n![\"Victoria](\"https:\/\/www.newsbeep.com\/ie\/wp-content\/uploads\/2025\/12\/VICTORIA_SONG.0.jpg\"\/)
Victoria Song<\/p>\n