24 AI systems for an online store. All of them start with knowledge about your store.
runsitself maps how an ecommerce or DTC brand can use AI, layer by layer. Click a system to see what it does, what it depends on and whether I have already released it. You are not buying an AI tool. You build your store's context, and the systems read it and feed it back.
Level 1
A chat project with your store knowledge file. A start for everyone, nothing to install.
Level 2
Cowork on files and folders. The same tasks, on hundreds of products at once.
Level 3
Claude Code or Codex connected to your store, ads and email. The system runs on its own.
All 24 systems as a list
The same map in plain text. Each system shows its status, level and what it depends on.
Core
Every system reads from here. Four of them write back.
- F1 Store context. A knowledge file about the brand, products, customers and rules. What not to say, which voice to write in, what is true about each product.
Foundation
The data and connections everything else stands on.
- F2 Product data and feed. Complete, consistent attributes, specs and images. Without them every description and every ad is guessing.
- F3 Connections. Store, marketplaces, email, ads and analytics available to AI in one place.
- F4 AI costs and security. What it costs per month, who has access and which customer data may be sent to a model.
Acquisition
Where people come from.
- P1 Meta ads assistant. Sets up and reviews campaigns using knowledge of your store, not generic advice.
- P2 Competitor ads radar. Which competitor ads run for a long time, so they are probably profitable.
- P3 Creatives and content. Video stills, graphics and copy for ads and social media in your brand voice. See the kadry repo
- P4 Visibility in Google and AI. Articles and pages that ChatGPT cites and Google finds.
Conversion
What convinces them to buy.
- K1 Product descriptions in bulk. Hundreds of descriptions at once, without invented specs. Facts come only from the feed.
- K2 Product page and cart audit. What blocks the purchase on the product page and in the cart.
- K3 Prices and promotions. Competitor prices, the promotion calendar and margin in one view.
- K4 Marketplaces. Offers and listings on Amazon and other marketplaces from the same data as your store.
Retention
What brings them back.
- R1 Email flows. Welcome, abandoned cart and post-purchase emails.
- R2 Campaigns and send calendar. What to send and when, all year round, from Black Friday to peak season.
- R3 Reviews and surveys. Customer words turned into rules the AI writes by.
- R4 Win-back. Customers who stopped buying and how to win them back.
Operations
The back office the customer never sees.
- O1 Customer service. Draft replies from the knowledge base and order status. A person only approves them.
- O2 Returns and complaints. Return reasons turned into lessons about the product and its description.
- O3 Shipping costs and invoices. Carrier and supplier invoices checked against the contract, line by line.
- O4 Stock and reordering. An alert before stock runs out and a suggestion of how much to order.
Brain
Decisions on data, not on gut feeling.
- M1 Weekly report. Sales, ads and email on one page, every Monday.
- M2 Alerts. A sales drop, an out-of-stock item, an ad cost spike. Before you notice it yourself.
- M3 Board debate. Four roles argue about your decision before you make it. See the debata repo
- M4 Market radar. Demand, seasonality and competitor moves in one overview.