Our team was on the ground at Shopify's DotDev in Toronto this year, and the clearest takeaway was that the operational friction of running an ecommerce store is having its rough edges sanded off. Shopify is committing further to being the platform you build on, for merchants and developers alike, and leaning on AI to make that possible.

Here's what stood out.

Sidekick on every store, and now into apps

Sidekick is on every store, and it's just been expanded to reach into apps. Developers can build app extension points that tell Sidekick what functionality it's allowed to access.

To be precise about how that works, Sidekick can only reach into apps that have built those extensions. So as app developers build these out over time, Sidekick becomes a progressively more capable coordinator across a store's whole stack, not just a bolt-on assistant. This is the major leap forward from the event. It matters most for larger, more complex stores, where keeping track of everything is the real operational problem.

A great example of this in practice came from Eric Panofsky, Founder & CEO of US-based agency SLTWTR, who demoed building a back-in-stock notification system, the kind of thing that would previously have needed a third-party app, entirely with Shopify primitives. The upside is that powerful functionality is composed natively but the catch is that merchants sometimes push back with "why does this flow exist?" because a composed system is harder to explain than a single app with a name and a settings page.

Their answer was to build Sidekick skills that group the composed functionality together, so merchants can just ask what the system does instead of having to remember how the pieces fit. Because it all lives inside Shopify and Sidekick, the system effectively self-documents. Sidekick sees when things change, so there's no separate documentation to maintain.

It's a good illustration of where AI actually delivers value right now, not flashy standalone features, but making complex, composed systems usable and self-maintaining.

UCP and why your product data matters

Shopify walked through the thinking behind UCP (Universal Commerce Protocol), built as an extensible standard designed to adapt as new commerce experiences emerge, including ones we can't yet imagine.

The practical link for merchants is what actually gets published through it. Shopify Catalog syndicates your products to AI channels with title, description, category, options, images, price, and availability, structured so agents can parse them. The quality of what gets published depends entirely on the quality of the product data you enter.

Category is one of the fields worth getting right, and it's one of three, alongside title and description, where you can explicitly map a data source in the admin, under the Agentic sales channel. That's useful if your product data actually lives in metafields or metaobjects rather than the standard fields.

Where those fields are thin or empty, Shopify's AI fills the gap. That's useful, but shouldn't be relied on. Shopify marks these as inferred fields, and flags that their accuracy varies depending on the underlying product data. If you leave a field blank, the AI's best guess is what agents will act on. So populate as much as you can, product descriptions especially should be a priority.

Winning in generative search

Shopify also gave a first-party view on generative search, which was more useful than the usual outside speculation.

Two things stood out:

1. The SEO work you're already doing still counts, and still contributes to generative search results.

2. Off-site discussion: Reddit and other online communities are actively drawn on by the frontier AI labs when they generate product recommendations. (OpenAI's data deal with Reddit is a concrete example, with ChatGPT product suggestions often citing Reddit as a source.) Conversations about your products happening off your site are now a real ranking signal, not just word of mouth.

Other platform updates worth knowing

- Vercel partnership: should make deploying Hydrogen (Shopify's headless framework) smoother and more directly integrated.

- Liquid investment: Shopify continues to invest in its templating language, with new features on the way.

- App store speed-up: faster review processes have led to something of an app explosion, a much wider range of offerings now available.

- Collections 2026: deeper rule-based product selection, variant-specific collections, Sidekick assistance for building collections, and a new visual drag-and-drop builder.

What this means for merchants

AI is quietly taking over a lot of the plumbing... discovery, transacting, documentation, even parts of merchandising and shifting the real work upstream to data quality and off-site reputation. The merchants and agencies who win from here will be the ones who treat their product data and their off-site reputation as core infrastructure, not an afterthought.

If you’re looking to make sense of these changes, that’s exactly where we help. Reach out to us at hello@convertdigital.com.au