Last updated: June 18, 2026
On June 17, 2026, Shopify announced the "Spring '26 Edition." It's a bundle of more than 150 updates, but the theme it raised was narrowed to one. Sell everything, everywhere, all at once. And at the center of that "everywhere," it placed the inside of AI chat.
When I saw this announcement, I wasn't surprised. If anything, the feeling was "it came, just as I expected." Why I'd been thinking that, I'll talk about in the second half. First, let me look as accurately as I can at what Shopify actually did.
The buyer sets up nothing
The most symbolic part is checkout. Shopify made it so that, from inside a Microsoft Copilot conversation, a user can buy a product directly and complete payment right there with Shop Pay. They say they'll soon extend this to shopping in Meta ads, too. What underpins all of this is the Universal Commerce Protocol (UCP), a shared language between agents that they built together with Google.
What matters here is that the person buying sets up nothing. They don't need to connect MCP (the Model Context Protocol, a common standard that connects AI to outside services) themselves. The buyer just talks to the AI they always use. Behind the scenes, the AI platform is connected to Shopify, and it finds the product, puts it in the cart, and pays. From the buyer's side, it's just "I talked in my usual chat, and I could buy." All the wiring sits below the surface.
The product-data side is automatic too. Shopify Catalog standardizes product information and distributes it to each AI channel, and they say it doubles the conversion rate in AI chat. Even merchants who don't use Shopify can, with the new Agentic plan, sync their products to Catalog and start selling through AI channels and the Shop app.
The real heart, easy to miss, is on the operations side
The flashy part is "buy in chat," but the place I paid the most attention to is somewhere else. It's that the work of the selling side — the people running the store — also moved into the conversation.
Spring '26 puts it this way. You can run your store inside Claude, ChatGPT, or Perplexity. From chat, you can add products, create collections, and even manage orders. The shop owner can move inventory, orders, and sales from the AI chat they always use, over MCP. Shopify also places an AI (Sidekick) inside its own dashboard, but at the same time it chose a two-track approach: letting an outside Claude or ChatGPT operate the store as well. The development foundation points the same way — they say they rebuilt Hydrogen, the base of the storefront, from the ground up with an agent-first lens.
Seen on a single sheet, it looks like this
Put in order, Shopify pulled every surface of buying and selling toward "from the conversation."
- The buyer, in their usual AI chat, buys and pays with no setup.
- The seller, from their usual AI chat, runs inventory, orders, and sales over MCP.
- The developer rebuilds the foundation itself, on the premise that agents use it from the start.
One of the largest commerce platforms in the world stepped halfway out of the race to "get the user to come to our screen," and moved to the side that gets called from any AI chat. This is, I think, the structural change I once wrote about in SaaS Is Moving from Chosen to Called, now arriving right in front of the consumer's wallet.
Why I wasn't surprised
Over this past year, I've watched the same move in other industries again and again. Accounting platforms such as freee and Money Forward put a hard, formal job — bookkeeping — inside the conversation. Watching that flow, I was convinced that FORMLOVA, which does form creation, operation, and analysis in chat, isn't mistaken. What remained was shopping. The most everyday thing people do — the act of paying money. Once that enters the conversation, the flow becomes decisive. Shopify came to take exactly that.
So for me, this isn't a threat. It's a tailwind. There are two reasons.
One is that the category becomes standard. The more the experience of buying in chat grows, the more the selling side gets chances to run a store over MCP. Then, running a B2B service over MCP — not from your own dashboard but from your own chat — becomes ordinary. This is the very premise of FORMLOVA that I've struggled so long to explain. The wall of awareness comes down from the other side.
The other is that the customers overlap. Commerce merchants running a store over MCP will, I think, surely want to hold their forms the same way. Inquiries, applications, surveys — they'll want to run all of it from the chat they always use. The more Shopify gets store owners used to MCP, the closer those people come to being FORMLOVA's prospects.
The operations side is the same as Shopify
Let me say this plainly. In the point of putting the selling side, the operating side, into the conversation, Shopify and FORMLOVA face exactly the same direction. Shopify's shop owner moves inventory and orders from their own Claude or ChatGPT. FORMLOVA's operator moves forms and responses from their own Claude or ChatGPT. We just handle different goods; the structure is the same. How FORMLOVA designs this operations layer, I've also laid out in the MCP Form Service Guide.
But I won't build my own AI
On one point alone, I make a different choice from Shopify. Shopify takes the two-track approach, but FORMLOVA holds no in-house AI chat. There's a single path: get called by the AI the user already uses.
I do it that way on purpose. There are two reasons.
One is that I don't want to split the experience. In the end, if you open a dashboard and talk to the bot inside it, there's no point doing it in chat. People will lean on their usual chat more and more from here. So many things finish there, and yet you deliberately move to a separate screen and ask a separate bot the same things. Frankly, I think that's nonsense. In fact, I'm that way myself.
The other is cost. If I hold an AI's API inside my own screen, that expense lands on me. Every time a new model comes out, there's a switching cost too. Conversely, if I use the AI on the user's client side, that cost doesn't arise. So FORMLOVA can keep its pricing low. And when the model the user uses gets smarter, FORMLOVA's experience gets better without my doing anything. I treasure this structure.
Here is the real point — conversation doesn't "erase" the form
So, does everything dissolve into conversation? I don't think so. This is what I most want to convey in this article.
Look once more at Shopify's "buy in chat." When you become able to buy inside a conversation, the form doesn't disappear. The payment is carried by Shop Pay — a structured finalization surface that properly receives the address and the payment. In fact, Shopify states explicitly that the address-format validation rules apply to agent-based experiences too. Even when an agent moves shopping forward in conversation, the final, settled data is caught by a surface like a form.
Conversation didn't replace the form. It called the form, and wrapped it.
Once you see this, you see that FORMLOVA's design isn't going against the trend — it's riding the same principle.
So the "answering person" is best left as a form
FORMLOVA is a service for the people who make and run forms. It isn't for the people who answer them. And I believe the experience of the person answering a form can't become chat.
There are two reasons. One is that the form is best at capturing settled data. A name, a preferred date, a consent — receiving them on a structured surface is faster, more accurate, and easier to handle in the work that follows than collecting them loosely in conversation. It's the exact same reason payment for shopping stays on Shop Pay.
The other is that the direction of delivery is reversed. Chat is a private place the person themselves opens on their own timing. I can't slip a form into someone else's chat from my side. A form is something you send, and the other person opens. So the universal surface you can deliver to a respondent ends up being a link — that is, a form.
And keeping it as a form isn't a step backward into the past, away from the age of agents. It's the opposite. A human, and a browser-operating AI agent, can both fill in the same form. The form is the lowest-common-multiple surface that both people and agents can use.
To be honest, putting the answering experience itself inside the conversation isn't technically impossible. But it would require clearing several external conditions I can't move on my own — for instance, being integrated into a large platform. Even if those someday lined up, what to do beyond that is something I can't picture now. So for now, the answer is clear. The operating side belongs to chat; the answering side belongs to the form. This boundary wasn't decided by trend. It's the natural answer drawn from the nature of settled data and the structure of delivery.
As the one being called, what to hand to the conversation and what to keep
SaaS went from being chosen to being called. In an age where conversation becomes the entrance to everything, what's truly being asked, I think, isn't to dissolve everything into conversation. It's what you hand to the conversation and what you keep on a structured surface — and where you draw that line.
Shopify handed the discovery, the consultation, and the ordering of shopping to the conversation, and kept the finalization surface of payment on Shop Pay. FORMLOVA hands the creation, operation, and analysis of forms to the conversation, and keeps the finalization surface of the answer on the form. What we're doing is the same judgment.
It isn't only the SaaS industry that's shaking. The ground under me, the one who builds it, is shaking just the same. But on this boundary, I'm not shaking. Rather than dissolve everything into conversation, use the conversation and the finalization surface each for what it's good at. That, I think, is what it means to make forms in this era.
Related articles
- SaaS Is Moving from Chosen to Called — Built-in AI, MCP, and the Shifting Ground Under Software
- The Philosophy Behind FORMLOVA — finishing work as an extension of the AI chat you already use
- Why FORMLOVA Kept Its Dashboard
- Most Form Tools Stop at Creation — FORMLOVA Starts After You Publish
- Why Form Operations Need an MCP Layer
FORMLOVA is currently available from MCP-capable AI clients such as Claude, ChatGPT, Gemini CLI, Cursor, and Windsurf.
Sources
- Shopify: Selling everything, everywhere, all at once: The Spring '26 Edition
- Shopify: Agentic commerce for every developer: The Spring '26 Edition
- Shopify Dev Docs: Agentic commerce
- Google: New tech and tools for retailers to succeed in an agentic shopping era
Disclosure and Verification
This is a FORMLOVA owned-blog article. The author is the developer of FORMLOVA. The descriptions in the body of Shopify's announcement (Spring '26 Edition, the Universal Commerce Protocol, Shop Pay, Catalog, Sidekick, Hydrogen, and so on), Google's agentic commerce announcement, and the mentions of freee and Money Forward were written after checking primary information publicly available as of June 18, 2026. Each company's announcements and figures may be updated, and expressions vary, so if you rely on them for important decisions, please confirm with each company's official announcements. Pricing, features, and limits are subject to change, so please check each service's official pages for the latest information.


