Shopify Editions Winter 2026: Sidekick Pulse, apps built by Sidekick, SimGym and Rollouts

Episode recorded in Polish straight after the Shopify Editions Winter 2026 presentation (released 16 December 2025), so it describes the announcements as they stood then. As of October 2026, Rollouts (A/B tests of the store and checkout) and SimGym are on the feature list of Editions Spring 2026 (June 2026), which we cover in a separate episode; the teaser at the start of the video is not repeated here.

The "Renaissance" edition, with AI inside

Maciek: Did you watch the presentation?

Tomek: Of course! Like every six months, I had nothing else planned but to see what's new at Shopify. I think that's what we'll talk about today, because Shopify's announcements are always the moment when you realise all over again how much Shopify itself, and e-commerce in general, is changing and will keep changing. Shopify has always set the trends and I don't think that has changed. They're still far ahead in how they think about e-commerce and how it should take shape in the coming years. And I think this edition, provocatively, even historically, called the "Renaissance" edition…

Maciek: With AI built in.

Tomek: With AI built in, shows that something is being created anew, right?

Maciek: No doubt about it. It would be a truism to say that AI is now unlocking new reserves of creativity and possibilities. That says nothing. But the scale and pace of the changes Shopify presented at this year's Editions really are staggering. Some of the features they showed are really… I expected it to be about AI. I didn't expect the scale and the sweep.

A caveat: an announcement is not production

Maciek: It's definitely worth starting with a caveat about these Editions. They say they're announcing new features. For me, at this stage it's still mostly marketing hype around what they're working on. Some things have of course gone into production, but some only for certain merchants or in certain markets. Read: US only.

Tomek: Yes.

Maciek: So I think that caveat is worth making, and you have to approach all this with some reserve. We know that sometimes it goes like it did with variants: they announce a feature, and in practice it arrives a year and a bit later. That's how you have to read all of this. That doesn't change the fact that many of these features, and there are over 200 of them, are really interesting.

Sidekick Pulse: an assistant that comes up with ideas itself

Tomek: Yes, there are quite a few features and they cover different areas. But I think we agree that the most groundbreaking ones, the ones that signal big changes, are around what they've done with AI. Let me go straight to the first one. Some of you have probably already used Sidekick, treating it partly as an advisor and partly as a shortcut to features that were already in Shopify. We typed in queries that were supposed to give us results, reports, quick analytics, and we warmed to it with more or less trepidation. Whatever you say, until now it was a tool that took our queries and answered them. What Shopify is coming out with now, the direction this tool is evolving in, is meant to change that perception. Sidekick is meant to be a companion in our work that is not only our assistant, helping us carry out tasks we came up with, but also initiates some actions itself. I mean Sidekick Pulse. I remember your disclaimer, so we don't go out on a limb saying something already works fully and then it turns out the full functionality isn't quite there. But the idea is that this tool sits deep inside our e-commerce, our ecosystem, analyses what's happening in the store under the surface, in real time, and starts suggesting ideas and solutions to us, so sometimes it thinks for us about how we can optimise the business.

Are merchants actually using Sidekick?

Maciek: Surfacing insights. On Sidekick in general, my observation is that Shopify has of course built a lot of hype around it, but they quietly admit that adoption isn't yet at the scale they expected. I don't know what percentage of merchants use it, because I don't think they've shared that, but it's not the majority. That's one thing. And when we talk to our clients at the agency, I don't see people using it eagerly. I think that so far most of the use is in analytics, because it does handle that well. I also read yesterday that Anthropic announced that Shopify Sidekick uses their language model, Claude, and that's the architecture underneath, but that's a digression. Pulse is supposed to work by generating insights and suggestions on its own. That's not new on the tech market either, because there are probably at least a few tools already doing this: Analytics, Hotjar, Clarity, wherever there's data…

Tomek: Yes, but as you say, those tools are scattered. What Sidekick ultimately wants to become, and how Shopify recommends it, is an assistant that already has the data aggregated, so you can ask about it and get it in one place instead of being sent to different tools and having to put the results together yourself. I think that's the direction.

Maciek: Yes. There's also the question of the quality of these insights. What will they be like?

Tomek: I think what you mentioned, that Sidekick usage today isn't at the level Shopify expected, or even we expected when we thought about how useful it would be for clients, may come partly from the fact that we're not yet convinced about the quality of the suggestions. How meaningful are they really, do they give real pointers? Maybe people working with Sidekick today run into a fairly generic answer along the lines of: "OK, but I can see that in the report. That's not a revelation for me at all." Maybe that's why the approach to Sidekick is still very cautious for now.

Maciek: Maybe. I can't answer that. Take it not as a question but as a comment.

Tomek: As a comment.

Sidekick builds apps in the admin

Maciek: What I find more interesting, though, is what's happening in Sidekick in the area that's already available and that we can use. The killer feature, in my view, is the ability to create apps in Sidekick that show up in the admin. Until now Sidekick returned simple information and instructions. Now it can create a mini app that will do some mini task for you.

Tomek: Give an example.

Maciek: On the page they gave the example of a tool that tells you which products you should include in your next restock.

Tomek: OK.

Maciek: That's a good case. We also checked with the guys in development whether it can connect the front end with the back end, for example whether it could build us an app for Omnibus. For now these are only interfaces strictly for back-end things. But I think it's strong, it's cool, and I can hear the shallow breathing of Shopify agencies as another part of the market slips away from them. That's a separate story altogether, because you can see the direction of handing as much power and control over the store as possible to merchants. It unlocks possibilities, but you know: with great power comes great responsibility, and you can actually break a lot. I think the game isn't lost.

Tomek: That's right. I also think that up to a point you're purely a beneficiary of these changes. You feel they really make your everyday work easier. But once you cross the line where everything starts to depend only on you, then, as you say, risk and responsibility come in. Second thing: if a solution has a certain level of complexity, it's no longer so obvious…

Maciek: I think it can be a straight road to taking on technical debt.

Tomek: I agree. And with more complex solutions there's often a big risk and responsibility for business continuity: we interrupt something that was working properly, and the responsibility lies with whoever did it. Playing with this on your own mostly brings benefits up to a point, and we see them, but at some point it may scare us a bit.

Maciek: It's vibe coding in its purest form, for merchants. You can click an app together just by typing.

Sidekick in the theme, in Shopify Flow and in segments

Tomek: I also went through what was announced about how Sidekick itself is developing. You said that after checking with the developers we don't yet see a connection between the back end and the front end. So we can treat these separately: on the one hand, Sidekick lets you build custom apps that run on the back end and apply back-end logic to return certain solutions to us. On the other hand, we can now say with confidence what's happening on the front end itself. Sidekick also lets us propose changes strictly in the theme, for example describe: "I need a section with recommended products" or "I need a section where I show customer reviews together with the products they bought". Sidekick can create such a section directly in our theme. We have control over it and we don't need a developer layer for Sidekick to physically write components into the theme for us.

Maciek: You can see it can do more and more. The second feature I find pretty cool is Sidekick's integration with Shopify Flow. Until now Sidekick understood Flow, but if you wanted to create an automation, a workflow, it kept giving you written instructions on how to click it together. Now it will be able to prepare it for you and present it for review and publishing. The same with segments, which we'll probably talk about in the next episode: it can propose them too. You can describe a segment in Polish or in English, better in Polish, because it understands Polish better, and it will create it without any clicking. That's a nice thing.

SimGym: tests on AI bots instead of customers

Maciek: The next thing is, I think, the most interesting feature right after custom apps. If you watched the demo where Harley and Tobi talked about it, I had the impression this was the part they were most excited about. They worked on it for more than a year and called it SimGym. To explain what it is, we can compare it to a gym, a training hall where we prepare for a competition. It gives us a virtual environment for testing our ideas, assumptions, business hypotheses and changes to our front end, in controlled conditions that don't expose you to business risk. The test isn't run on people but on traffic from AI bots, let's call it that, which have learned to behave like your users based on the ocean of data Shopify has access to. Depending on how the experiment goes, you can then roll it out to production. That links to the next feature, Rollouts.

Tomek: Yes, Rollouts.

Maciek: Sorry, just to close the SimGym topic, which I find super interesting: my bet is that it will be paid. I think it was said during the presentation that the cost of running this feature is absurd.

Tomek: I agree. The way I understand it, I see two scenarios for using it. First: you're a big business with huge traffic on your site. When you want to introduce a new solution, what you mainly face is the risk that you'll launch something that converts worse or doesn't do its job the way you assumed. You'll take a real loss because your revenue will drop. Second scenario: you're a business that doesn't have enough traffic. Even if you put a new feature into testing, you need a very long time to gather a big enough group of people to test it, and you'll only draw conclusions after a month. Then you use SimGym like an incubator: in a safe environment, without bringing anything out into the open, you get very reliable feedback you can base decisions on. I also expect it won't be a free feature, because I can imagine how much infrastructure it takes to run it at the scale Shopify has today.

Maciek: It will definitely need some credits.

Rollouts: A/B tests in the admin

Tomek: Yes. A related topic, maybe not going as far, because what you described is closely tied to new technologies. The feature that in my view will be a great help and will hopefully appear soon is called Rollouts. It's a tool for running A/B tests of the solutions we want to use, quickly and efficiently. A very simple example: we're redesigning the collection page or the product page, so a specific element of our site. In this tool we set up two versions of that view. And just like in the old Shopify Launchpad app, we can schedule for how long and for which group…

Maciek: Only Launchpad didn't have A/B tests, I think. It let you publish things in bulk…

Tomek: At a set time.

Maciek: …so, for example, at a given moment everything you prepared for Black Friday goes live in the store. Codes appear, a new banner.

Tomek: Some price cuts.

Maciek: Exactly.

Tomek: When I mentioned Launchpad, I meant the event calendar more than what you can do with it. I completely agree with you. Here you set when version A and when version B should be published, so you can roughly plan on the calendar how many days a given view will be shown to visitors. The second thing: you choose what share of traffic goes to version A and what to version B, so the results are as comparable as possible. After the test you get ready results from the app, a summary and a comparison of both versions. When you decide which one to finally put into production, you're not relying only on intuition, a hunch or what you imagine, you simply compare the numbers. Most simply: what the conversion rate looks like, if we're testing a product page, for example.

Maciek: I think it's great. It's actually one of the nicer features. I thought you'd start doing a Matthew McConaughey from "The Wolf of Wall Street". Come on.

Tinker: a playground for AI tools

Tomek: No, nothing like that. As we said, there are a bit more than 150 features. 150-plus. And as you said, a good part of them concerns markets that are still a bit exotic for us, which, as usual with Shopify, means the market across the ocean: the United States and Canada. We won't go into how those apps are developing here. I'd draw attention to one more tool Shopify has created. As I understand it, it's almost a separate element from Shopify itself, because I can't imagine it being in the Shopify admin, though let's not rule it out. It's an app they called Tinker. It's a kind of aggregator of available AI solutions that can be useful to an e-commerce owner. Today we come across loads of ready-made apps that help you create better images for your store, video or animations. Loads of tools meant to speed up your work. What Shopify noticed above all is that there's a huge number of them and they're scattered, so it's hard to find your way around and hard to keep up with what's coming out. So they wrote an app that's unified on the front end, under one brand. You choose what type of feature you're looking for, and behind it the external tools are plugged in, so you can test how they work in this app and decide whether a given app would be attractive for you and whether you could use it in your store. First, you have at hand what's going on in AI: the most popular products but also the new ones, so you stay up to date. Second, in a pretty nice, friendly setting, you can test them safely and only then make decisions, instead of jumping from one solution to another because you heard about something new and want to check it out. That also takes a lot of time.

Maciek: Yes. I think the idea is for it to be free exploration, a playground for these tools, and along the way maybe something will come out of it that one day turns into a product, because you decide it looks nice and is interesting. You send the materials you generate to friends, someone tells you: "Great, you should sell this", and somewhere it turns into a business. I see it a bit as a creativity incubator.

Tomek: Creativity for people who maybe aren't merchants yet.

Shopify experiments: scale versus quality

Maciek: Or are. That was probably the best line of the whole presentation, when Tobi said he deeply believes the future depends on what creative and smart people do in their spare time, at weekends. I think that hits the nail on the head. What will this app be? Probably an experiment, like the Shop app was meant to be, and that one, as it turned out, took off. Maybe this one will take off, maybe not. Shopify has a lot of ideas, and I think that's an interesting summary: they have many ideas and various investments, and of course some don't work out. They invested in a huge shipping company, I think for two billion dollars, and then sold it for 10–15% of the value. There are many attempts and experiments. You can see that the quantity and scale don't quite go hand in hand with quality. If you look at the apps in the Shopify App Store that are "built by Shopify", they don't have phenomenal ratings. I think the problem is that they may be doing too many things at once.

Tomek: On the other hand, I wonder whether from the leader…

Maciek: That's kind of expected.

Tomek: …that's exactly what you expect. Of course.

Maciek: I think for them it's like the funds that invest in 100 startups: if one takes off, you make money. I think that's the approach. Finally, it's worth mentioning that the website they made for this edition is beautiful. It's probably the most beautiful thing I've seen online this year. Absolutely fantastic. If someone hasn't had time, it's worth spending a few minutes scrolling through it. An absolutely beautiful thing.

Wrap-up: the proof is in production

Tomek: One last sentence: let's not get attached to what we saw in the presentation, because we'll only get excited about it once it works in production.

Maciek: That's the moment of truth.

Tomek: Sure. Right, Maciek, what do you say? I think we've got it.

Maciek: That's it. Thanks a lot.

Tomek: Thanks very much. Talk soon, see you.

Maciek: Right on.