 |
|
|
|||||
Imagine an interface where you can quickly shift between talking, typing, clicking, and even drawing to instruct software, like moving around a whiteboard in a dynamic conversation, Carl Rivera tells me. An experience in which users are not presented with a barrage of nested menus, but with a blank canvas that invites creativity aided by an artificial intelligence that knows everything there is to know about online and brick-and-mortar retail and marketing. A fluid interface that adapts and anticipates your needs, automating tasks and recommending actions like the most brilliant partner you could dream of.Thats a dream in itself, but it isnt a fantasy; its Riveras future vision for Shopify. Rivera is the companys new Chief Design Officer and he believes that, in the very near future, the e-commerce platforms user experience is going to feel like sci-fi. Rivera joined Shopify through the 2018 acquisition of his startup TicTail. Right after that, he was key to launching Shop, the companys consumer-facing business. His new position directly responds to industry skepticism about designs relevance in an AI-driven landscape. In this time in which everyone is shifting to AI but almost nobody has a clear idea why, it makes sense that Shopifys founder Tobi Lutke thought he needed someone like Rivera helming that leading position.Were entering a new technological paradigm with AI, Rivera says, emphasizing that now, more than ever, it is strategic for Shopify to have a clear design vision about how to implement artificial intelligence in a truly empowering way for every company, from small retail shops to corporate giants. The company wants to reimagine its user experience, transforming it into a powerful tool for designers and business people that is easier to use and saves more time than ever before. Half of the people are talking about design being dead because the programs can design for you, he says. We take quite the opposite point of view at Shopify.Rivera believes that designs importance has only intensified with AI. He recalls how in strategic meetings participants, despite active discussion about the future, agreed but had different pictures in their head, different images of what that future they all discussed together looks like. This lack of shared visualization, Rivera explains, is precisely where designs power lies. Put a designer into that conversation and start prototyping live as the conversation goes by, you start visualizing these different points of your imagination. This process, he elaborates, enables participants to align and start to kind of come on to the same version of the future. He argues that when a designer begins to sketch ideas or build quick prototypes in real-time, abstract verbal agreements translate into concrete visual representations, forcing participants to confront and reconcile their differing mental models. [Image: Shopify]A Shopify UX revolution is brewingFor all of Shopifys history, merchants confronted your typical dashboard: nested menus, a labyrinth of left-hand navigation bars spanning orders, products, customers, and analytics, and module-specific toolbars for inventory or marketing features. This hierarchical menu systems approach has defined the digital age, from Word to Photoshop. It requires users to memorize pathways. The entire navigation of Shopify is revealed right when you come in, Rivera says, because thats how we have learned to navigate software.As Shopifys capabilities expandedadding banking services, global tax compliance, or AI analyticsthis static approach strained usability. It doesnt at all seem to me that that is how you must be navigating software in the future, Rivera says. Instead, he envisions interfaces where AI-driven progressive disclosure would surface tools contextually. Imagine a merchant discussing a winter campaign with Sidekick, he tells me, which prompts Shopify to generate just the relevant email templates and ad budget controlsbypassing the marketing modules full menu structure. The user experience will morph and adapt to the needs of the user at any given time, a sci-fi dream come true. Riveras redesigned point-of-sale interface previews this philosophy: visuals stripped of clutter respond to retail workers needing full proficiency within 15 minutes of the first day, as Rivera explains it. Future iterations could reduce that to seconds. Very, very complex software will look extremely easy, Rivera insists, because capabilities materialize only when essentialnavigation bars dissolving into contextual prompts, settings menus summoned by voice, dashboards generated from conversations. The blank canvas replaces the control panel.[Image: courtesy Shopify]His vision for this simplified interaction centers on Sidekick, Shopifys AI assistant. Previously, merchants might have asked Sidekick isolated questions like Whats my best selling product in Germany? Now, the system handles complex, multi-step requests. A merchant can prompt, Hey Sidekick, Im thinking about a campaign for next winter, I want to target my top markets and I want to be selling my top products and I think it should be email and social media. Sidekick then shows its chain of thoughts by pulling together disparate pieces of information and creating a linked series of tasks that the user can observe in real time.For instance, Sidekick would analyze sales data to identify top markets and top products, then leverage its understanding of marketing best practices to draft initial email copy and social media posts tailored to those products and regions. It might even suggest an optimal budget allocation based on historical campaign data, demonstrating why Rivera says its becoming this really capable and active business partner. Last year, Lutke described Sidekick as a Robin to merchants Batmansomething thats right there with you kind of working through the tasks. It aims to become what everyone calls today agentic AI but focused exclusively on your business and your industry.[Image: courtesy Shopify]Though agentic AI still has a long way to go before its truly useful, Rivera says Sidekicks strength is that it operates as an agentic AI constrained to the context of Shopify. Unlike a general purpose model, Sidekick performs actions within the Shopify admin on behalf of the merchant, and its context comes from Shopifys proprietary data, not the open internet. And thats a lot of data: an internal dataset, derived from millions of merchants. It allows Shopify to come up with the best business strategy to cater to each individual one of them and offer a marketing strategy. We understand more. We have a better output of how businesses run their companies than any other company in the world. For instance, a generic AI might offer general marketing advice, but Shopifys AI can tailor recommendations based on actual e-commerce success patterns within its specific ecosystem, understanding granular details like how certain sneaker designs sell when it rains in Berlin, for example.[Image: courtesy Shopify]Rivera also talked to me about Shopifys recently redesigned point-of-sale (POS) system as a proof of concept for this design philosophy. The overhaul focused on making the interface intuitive for hourly workers, who need to understand entire software and how it works, you know, within 15 minutes of the first day at the job. This presented a distinct design challenge compared to designing for the Shopify admin, where users are typically more accustomed to complex software. The solution involved simplifying complexity while maintaining the extensibility and specific configurations required for various retail environments. He points to the common complexities of a POS systemmanaging inventory, applying discounts, handling returns, processing multiple payment typesand how the new design strips away extraneous elements to make these operations intuitive for a novice user. It looks gorgeous, Rivera says enthusiastically. The future of Shopifys designersThe future of design work within Shopify, Rivera explains, will see designers shift from crafting individual interfaces to building foundational systems. He predicts that engineers will be able to vibe code interfaces by describing them. This means an engineer or product manager could generate a draft UI by simply putting it into words, for example: I want to create an analytics page that shows me my email campaigns for Germany over this time period. The AI would then generate an initial experience that looks pretty good. Designers, he clarifies, will then spend their efforts perfecting these reusable components by focusing on building great design systems that that that AI can pull from. [Image: courtesy Shopify]He tells me that they will spend much more time honing the details. If you create a perfect form, if you create a perfect animation from one page to the next, its going to be able to be reused over and over and over again for every consumer thats vibe coding a Shopify app or creating a Shopify experience in our team or amongst our partners, he says. This demands meticulous system building, as Rivera emphasizes that to create really great AI outputs, you need really amazing AI inputs. This approach means designers will spend more time on fewer things, and dedicate their expertise to high-impact foundational elements that AI can then propagate across millions of stores.[Image: courtesy Shopify]The roadmap for this evolution will unfold through incremental updates. Sidekicks current capabilities will expand into multimodal interactions. He envisions an interface where users can effortlessly shift between modalities. Its not just point and click. Its not just chat. Its not just voice. Its all of them at once. He suggests an environment similar to a meeting with a whiteboard, where one can explain this to you using voice, then draw a rectangle and have an arrow, and then pull up my computer and show you something that is there on that side or whatever. This means a fluid, open canvas where users can chat, talk, click, or draw to instruct the software. Rivera anticipates a notable level of AI personalization in just a year. Drawing a parallel to hiring a business partner, Rivera envisions a future where merchants can participate in a hiring process and choose which business partner you want to put into your admin and call it Sidekick. This implies distinct AI agents with varying approaches, akin to how different human partners bring unique qualities to a business, perhaps a data-driven analyst persona or a more creative marketer persona. But while that seems sudden, Rivera tells me that the revolution wont come all at once. Progress will be cumulative. I like the expression overnight success, 10 years in the making,' he says. He believes the change already started a year ago, when Shopify first launched Sidekick, and will continue with updaes like a voice version. Eventually, he predicts, it will feel like the sudden market dominance of the iPhone, transforming silly apps that like cant do anything into an intelligent interconnected system. By the end of this year, well have made a ton more progress, he says. And by the end of next year, well be pretty science fiction-like.
Category:
E-Commerce
CosMcs, the glimmering, retro, space-agey concept restaurant from McDonalds, is no more. In 2023, McDonalds announced the spin-offbilled as the next frontier for the fast-food chain to test its most otherworldly specialty beverage ideasto a deluge of marketing fanfare. CosMcs was a drive-through-only concept with a pared-down menu of neon-colored drinks and a few snack items. The first CosMcs restaurants opened with lines around the block before the sun was even up. Now, less than two years later, McDonalds is jettisoning the stores back into the ether. According to a press release published late last week, McDonalds plans to shut down all five of its CosMcs locations (one in Illinois and four in Texas) in late June, as well as delete the restaurants associated app. In the coming months, CosMcs-inspired flavors will be landing in hundreds of U.S. McDonalds locations as part of a wider beverage test. The announcement comes in the wake of McDonalds first-quarter 2025 financial report on May 1, which revealed that the chains sales dropped at the beginning of the year, marking its second consecutive quarter of declines. Experts say there are a few main reasons why CosMcs didnt work out as a stand-alone conceptbut that doesnt necessarily mean the spin-off was a failure for McDonalds. [Photo: Stacey Wescott/Chicago Tribune/Tribune News Service/Getty Images] Bubble tea, energy drinks, functional soda, oh my! From the beginning, it was fairly clear what McDonalds hoped to gain from CosMcs: an entry point into the speciality beverage category (dominated by players like Starbucks, Dutch Bros., and Dunkin) thats been on the rise in recent years. As Gen Z has become increasingly interested in beverages like bubble tea, functional soda, and colorful energy drinks, other quick-service restaurants (QSRs) have moved to catch up. In 2024, Starbucks experimented with adding bubble tea to its menu; Dunkin introduced an energy drink lineup; and even Taco Bell opened its own beverage-only spin-off called Live Más Café. Meanwhile, McDonalds beverage offerings have remained largely limited to its soda machines and McCafé coffee menu (which, interestingly, also originated as an Australian spin-off concept). CosMcs was McDonalds answer to this gap in its offeringsa space to, as the restaurant put it at the time, perform a limited test of otherworldly beverage creations at a safe distance from its main restaurants. Within CosMcs blue-and-yellow beverage test kitchen, the chain was free to trial-run concepts like Tropical Spiceade and Island Pick-Me-Up Punch to a smaller audience of consumers. On the companys first-quarter 2025 earnings report, CEO Chris Kempczinski called this strategy quarantining the complexity in a stand-alone concept. [Photo: Matt Schwerin for The Washington Post/Getty Images] According to Matt Michaluk, executive creative director at the branding agency JKR, CosMcs made sense as a viable innovation for McDonalds. With an increasing share of occasions within QSR now solely drinks-only missions, and the diversification of menus by the big coffee chains, this should be a competitive yet fertile ground for growth, Michaluk says. In spite of that promise, he says, there are three reasons CosMcs fizzled out as a stand-alone: brand contradiction, absence of experience, and decline of hype. To start, Michaluk notes, CosMcs was shaped around a pseudo-nostalgic play on historic McDonalds brand characters, like the oft-forgotten 80s alien CosMc. But the spin-offs menu failed to align with that conceit. Further, the pilot format’s focus on drive-through architecture takes away from the overall brand experience, leaving consumers overwhelmingly underwhelmed. And, to cap it off, he says, Innovations and pilots work best when theyre new, exciting, and highly salient. McDonalds seemingly didnt invest in sufficient marketing efforts to support CosMcs. Hence, the hype died far too quickly. Within weeks of launch, there was nothing more to talk aboutnothing new, nothing to get people to come back. [Photo: Matt Schwerin for The Washington Post/Getty Images] Why CosMcs hasn’t failed yet Michaluks assessment might seem like a fairly bleak one, but Danny Klein, editorial director of the trade publication QSR, says the failure of CosMcs as a stand-alone doesnt necessarily equate to a failure for McDonalds business. From its inception, Klein says, McDonalds likely viewed CosMcs as more of a test run for a potential beverage expansion on its main store menus than a restaurant in its own right. Now that CosMcs recipes are rolling out across stores in the U.S., it appears that the initial experiment was a success. Hundreds of locations are going to start testing [CosMcs beverages], and I think from the general McDonald’s system standpoint, a beverage extension is what they all wanted, Klein says. I don’t think its a failure. People are going to say that because it was such a big deal, and then it just flamed out into the universe. But in my opinion, it was always a marketing test with the potential to be something else, and that just didn’t materialize. In addition to broadening the availability of CosMcs beverages, McDonalds also announced last week that it would create a new beverage category team dedicated to gaining share in the space. As Kempczinski told investors in early May, There’s a lot of growth that we see in beverages, and the profitability of beverages is very attractive, adding, frankly, we think there’s more that we can be doing to capture our fair share of that. Ultimately, Klein says, the true test of CosMcs will be whether the average McDonalds customer is interested in supplementing their Big Mac and fries with a Sour Cherry Energy Burstor if they choose to stick with a plain old Coke.
Category:
E-Commerce
People like to say that change happens gradually, then all at once. That pattern seems to be holding with respect to AI in search, and we may be at the beginning of the “all at once” part now that Google has officially launched AI Mode, which turns internet searches into conversations where you get answers instead of links. The point of AI Mode is for Google to act as an assistant to help you accomplish what you were trying to do with the search in the first place. Need to book a flight, find a sushi restaurant nearby, or grab a statistic that supports the email pitch you’re authoring? AI Mode will simply find what you need and even complete the action for you in many cases. And those cases will continue to expand: The company showed a future shopping capability where Google completes checkout for you without ever needing to leave the search page. Potential for Disruption The potential disruption to industries is staggering, not just for the media but also for marketing, e-commercethe whole web, really. For now, however, it remains mostly potential. AI Mode primarily lives as a button on the Google homepage and one of the tabs on results pages (alongside tabs for News, Photos, Videos, etc.). Users need to deliberately engage with it. And the omnibox in Chrome, where most Google searches occur, still defaults to regular search. {"blockType":"creator-network-promo","data":{"mediaUrl":"https:\/\/images.fastcompany.com\/image\/upload\/f_webp,q_auto,c_fit\/wp-cms-2\/2025\/03\/mediacopilot-logo-ss.png","headline":"Media CoPilot","description":"Want more about how AI is changing media? Never miss an update from Pete Pachal by signing up for Media CoPilot. To learn more visit mediacopilot.substack.com","substackDomain":"https:\/\/mediacopilot.substack.com\/","colorTheme":"blue","redirectUrl":""}} So despite the hype and panic emanating from Google’s I/O conference over AI Mode, Google isn’t going all-in just yet, and with good reason: Its existing business model depends greatly on the search results page. AI Mode can display ads, too, but it’s going to take time for the product to mature as a business. There’s also the simple fact that it costs Google more to serve up an AI answer versus a search pageit needs to move slowly in order to keep from tanking its own profits. The undeniable rise of AI search Make no mistake, though: The AI Mode summary will be the new battleground for attention. It’s fundamentally more engaging than even Google’s AI Overviews that appear at the top of search results pages. Whereas Overviews are a kind of “extra” to the list of links, AI Mode effectively creates a bubble around your Google experience, one that you deliberately enter and stay within. It’s designed to “fan out” from your initial query, turning search into something that’s more like a collaboration with Google on a task that search is just one part of. While that may sound like work compared to just getting served a search query, you have to remember: Once you had those results, you had to do the workthe navigating to sites, judging which were credible, and then manually absorbing information, filtering the irrelevant stuff. Now AI Mode does most of that work for you, greatly reducing the friction or “cognitive load” involved in getting information. I see this all the time in my own experience: Over the Memorial Day weekend, I ended up looping in AI assistants for several different projectshanging outdoor lights, what those metal ring-thingies are called, and how to optimize my cooking methods for pork ribs versus beef ribs. In all those interactions, no list of links was required, and in many cases, I got the information verbally, reducing friction even more. I’m a sample size of one, but studies suggest I’m far from alone. A recent study revealed 17%, or one in five consumers, now rely on AI answers more than traditional search. Referrals from generative AI to websites surged over 1,200% between July 2024 and February of this year, according to Adobe research. The AI search wave is real. When knowledge goes flat AI search experiences are more convenient, but it comes at a cost. If the AI summary is the new place for information brokers to conquer, there’s less land to fight over. Summaries simply can’t meaningfully cite dozens of sources in a curt answer. Moreover, if one or two sources change, the effect on the summary will be minimal. If an AI answer gets a new site fueling it, it’s still an averaged, homogenized consensus built from several sources. You don’t have the unusual link suddenly gaining prominence on a results page, inviting users to go down a rabbit hole. An AI summary is made to pave over those holes. This tendency toward singular, concise answers may have the inadvertent effect of flattening knowledge diversity. Mainstream perspectives will get prioritized, and niche or contrarian voices will have a tougher time standing out. Signal generators This shift puts a burden on journalists and media organizations to act not just as content creators but also as distinctive signal generators in a noisy, compressed ecosystem. In a world where AI systems synthesize information from thousands of sources, what gets retained are the most statistically common patternsnot necessarily the most insightful or original voices. That’s why it’s going to be essential for media sites to be able to do both: structure content to acknowledge and align with the mainstream view, but also provide original and unique perspectives that will offer real value for those who go deeper. It’s an updated version of a diverse content strategy, only in the AI world it can mean serving those ingredients in new ways: possibly by remixing content into formats recognizable by a multimodal AI that cares just as much about sound and video as it does about text. One thing’s for sure: AI answers are here to stay, and “winning” them is going to be the game to master. What’s unclear is what will be harder: influencing readers through what the summary says, or getting them to click through the AI ode bubble so you can influence them yourself. Let the games begin. {"blockType":"creator-network-promo","data":{"mediaUrl":"https:\/\/images.fastcompany.com\/image\/upload\/f_webp,q_auto,c_fit\/wp-cms-2\/2025\/03\/mediacopilot-logo-ss.png","headline":"Media CoPilot","description":"Want more about how AI is changing media? Never miss an update from Pete Pachal by signing up for Media CoPilot. To learn more visit mediacopilot.substack.com","substackDomain":"https:\/\/mediacopilot.substack.com\/","colorTheme":"blue","redirectUrl":""}}
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||