Update, October 9: Three days in, we're ahead of our own plan. Ten features are live, eight of them shipped today. So we're changing one rule: instead of holding finished features back to post one a day, every feature now goes on the changelog the moment it's live. We've also added something the original lineup didn't have: getting each store's digital twin ready for the AI agents that shoppers now send to shop for them. The updated lineup is below.
Short answer: From October 7 to October 20, we ship one new feature of the sqrly Shopify app every day: fourteen features in fourteen days, each one live for merchants and posted on our changelog the day it lands. We're doing it because our first pilot shops are about to install sqrly, and the fastest way to learn what they need is to put real improvements in front of them every day. The pace comes from a few rules: features small enough to finish in a day, a staging store where every change gets checked before it reaches a real shop, leveraging AI coding where possible, and a changelog entry as the definition of done.
Why ship every day?
sqrly gives a Shopify store a digital twin: an AI sales agent that knows the catalog, live stock and the merchant's own knowledge, and answers shoppers on the storefront. The first version works end to end, and our first pilot shops are about to install it.
That's the moment when speed matters most. A pilot shop's first two weeks decide whether sqrly becomes part of how they sell or an app they forget about. If a merchant tells us on Monday that something is missing, we'd like them to see it on Tuesday.
There's a second reason. Small teams usually move slower than they could, not because the work is hard but because each change waits for something: a review, a test, a deploy, a write-up. Shipping every day for two weeks forces us to find every one of those waits and remove it. Whatever survives Shiptober becomes how we work.
How we keep the pace without breaking anything
Fast is easy if you don't mind breaking things. Our pilot shops will have real customers talking to sqrly, so we set a few rules first.
1. One feature, one day. Every feature is scoped to fit a single day, from first line of code to live in production. If it doesn't fit, we cut it into a smaller piece that does. "Thumbs up and down on answers" is a day. "Analytics" is not.
2. Nothing reaches a real shop untested. We work on two branches. Every change goes to a staging app on our own test store first, where we check it the way a merchant and a shopper would. Only then does it move to production and the pilot shops. Fixes for real problems can jump the queue; features can't.
3. The changelog entry is the definition of done. A feature isn't shipped until a merchant can read what it does, in plain language, on sqrly.ai/changelog. Writing the entry is part of the work, not an afterthought, and it's often where we notice that a feature is harder to explain than it should be.
4. Leverage AI coding where possible. We use AI coding tools wherever they make us faster: scaffolding, routine changes, tests. That moves our time to the parts that decide whether a feature is any good: what to build, how it should feel for a merchant, checking it on the test store, and saying no when it isn't ready for a real shop. Every change is reviewed before it ships.
5. Merchants pick what's next. We start with a lineup, but it isn't sacred. If a pilot shop asks for something that fits in a day, it takes the next free slot.
What we're shipping
Here's the starting lineup. All of it is about one idea: a digital twin that gets better at selling for your store every week, with less work from you.
- Shipped October 7: your twin starts with what the web already knows. Right after install, sqrly searches the public web for what customers, reviewers and the press say about the shop, and offers it to the merchant to accept, edit or dismiss. Read the changelog entry.
- Shipped October 8: your twin speaks your shoppers' language. Shoppers ask in German, Spanish or Japanese; the twin answers in kind, using knowledge you wrote once in your own language. Read the changelog entry.
- Shipped October 9: fix an answer from the transcript. See an answer you'd phrase differently? Correct it right in the conversation, and the twin learns it for next time. Read the changelog entry.
- Shipped October 9: this week at a glance. A weekly summary on your home screen: conversations, top questions, products added to cart. Read the changelog entry.
- Shipped October 9: knowledge tidy-up. The twin spots duplicate entries and offers to merge them. Read the changelog entry.
- Shipped October 9: your discounts, known to the twin. Active Shopify discounts become offers the twin can mention and apply at checkout, with nothing to set up twice. Read the changelog entry.
- Shipped October 9 (new): your digital twin, ready for AI shopping agents. We sent an AI agent to shop our demo store through its digital twin, then fixed everything that tripped it up. Read the changelog entry.
- Shipped October 9 (new): "Ask our expert". The shopper's agent and the store's twin, working it out together, starts here: an "Ask our expert" page on the store's own domain that any agent can use. Read the changelog entry.
- Shipped October 9 (new): AI agents can consult your digital twin directly. Site tools next to Shopify's own let a visiting agent ask the twin and get back the exact products to add to the cart. Read the changelog entry.
- Shipped October 9: was this helpful? Shoppers rate answers with a thumbs up or down, so you know where the twin shines and where it needs you. Read the changelog entry.
- Contradiction check. "Your return policy says 30 days, a note says 14." The twin flags knowledge that disagrees with itself before a shopper hears both.
- A five-minute interview. The twin asks you a handful of questions tailored to your catalog, so it knows what makes your shop different on day one.
- Conversation starters. On a product page, the twin suggests the questions shoppers actually ask about that kind of product.
- Never lose a shopper to "I don't know." When the twin can't answer, the shopper can leave their email, and you see it next to the open question.
- Usage and limits. How many conversations your twin handled this month, and how much room is left.
- The sales funnel. From conversation to product card to add to cart to checkout, in one view.
- Twin health. One view of how ready your twin is, with the single next thing that would make it better.
We'll post every feature on the changelog the day it ships, and we'll write up what we learned on October 21, including the days that didn't go to plan.
What we'll measure
Feature count is the headline, but it isn't the point. We're tracking three things:
- Idea to live: how long it takes from deciding to build something to a merchant being able to use it. The goal is under a day.
- Problems reaching pilot shops: how many issues get past our test store. The goal is zero.
- What merchants use: which of the fourteen features pilot shops actually use in their first weeks.
If we hit the first goal and miss the second, Shiptober failed. Speed only counts if shops can trust what we ship.
Follow along, or join in
Follow along on the changelog. New entries go up as features ship, through October 20.
And if you run a Shopify store where good advice sells (gear where fit, compatibility or use case decides the sale), we're still taking a few pilot shops. Join now and the features you ask for during Shiptober could ship within the same two weeks. Email florian@sqrly.ai with your store's URL and what you sell.