Store data pulled live 2 connectors + 1 export · 312 ASINs · 9 disagree
Seller Central · ordersLive Shopify · returnsLive search-terms.csvExport
E-commerce store reviewAn AI agent team that reads your store data
An AI agent team that reads your store data
and writes Monday's to-do list.
Connect the store through its own connector and OfficeWorker reads it live each week: Seller Central for orders, FBA inventory and finance, Shopify Admin for orders, inventory and returns, eBay or Etsy for listings and orders. Whatever a connector does not expose, such as Amazon's advertising search term report, comes in as the export you already download. Commander lines the sources up by ASIN, finds the SKUs where the numbers disagree, and hands you a short action list with the evidence and the denominator behind every line. Nothing is ordered, paused or repriced without you.
- Start with
- A connected Seller Central, Shopify Admin, eBay or Etsy store, or this week's exports if you would rather not connect yet, plus your cost and fee assumptions
- Finish with
- An action list for the week, the aligned workbook behind it, and a negative keyword list you can paste in
OfficeWorker reads the connected Seller Central and Shopify stores live, fills the gap with the advertising export, and flags where the sources disagree by ASIN. Commander then turns the outliers into a short action list with owners and evidence.
Follow along
Run this week's review on the connected store
Copy any of these, paste it into the Orkas input box, and start trying it.
Weekly store review
Read this week's orders, FBA inventory and financial data from my connected Seller Central store and the returns and refunds from my connected Shopify store, then add search-terms.csv from the exports folder. Align everything by ASIN. Find the SKUs whose return rate is above 1.5× the category rate, the search terms with spend and no orders over 14 days, and any listing where ad spend is carrying the sales. Give me a short action list for this week with the evidence and the denominator behind each line. Do not place orders, change bids or write anything back to the stores.
Restock and slow movers
Look at the last 60 days of sales velocity and the current inventory in my connected store. Tell me which SKUs will run out before the next shipment lands and which ones have been sitting too long and should be discounted. Show the daily rate behind each call, and do not change any listing or place orders.
Why the bad reviews
Pull the recent reviews and return reasons for the three SKUs with the lowest ratings. Group the complaints by root cause, and separate the ones that are real product problems from the ones caused by the listing setting the wrong expectation. For each group, tell me what to change and where.
Price against competitors
Compare our prices and delivery terms against the competing listings for our five best sellers. Show where we are losing on price, where we are losing on delivery time, and where we could raise the price without losing position. Mark anything you could not read directly from the page as unverified.
Where the ad spend goes
Review the last 30 days of ad spend by campaign and search term. Find the terms with spend and no orders, the campaigns where ads are only carrying sales that would happen anyway, and the terms that deserve more budget. Give me the numbers behind each line, and do not change any bids.
What you receive
An action list you can act on Monday, with the numbers still attached
Every line names the SKU, the evidence and the denominator, so you can disagree with a recommendation on a specific ground instead of a feeling.
Weekly action list
Restock, pause, rewrite or replace: five to eight lines, each with the SKU and the number that triggered it.
weekly-actions.mdAligned store workbook
Live connector reads and your exports joined by ASIN, with the disagreements between sources highlighted for review.
store-review.xlsxNegative keyword list
Search terms that spent and never converted over the window, ready to paste into the campaign.
negatives.csvDetailed process
Connect the store or drop the exports, align by ASIN, find the outliers, review and schedule
Connect the store, or drop the exports
Seller Central, Shopify Admin, eBay or Etsy through an application you register yourself. Anything a connector does not expose comes in as the report you already download.
Pull and align by ASIN
The connector reads orders, inventory and returns as they stand right now; the exports fill the gaps. Four layouts become one table, with your cost and fee assumptions applied and shown.
Find the outliers
Returns above the category rate, terms with spend and no orders, listings where ads carry the sales.
Review, then put it on a schedule
Accept the lines you agree with; they become to-do items. Schedule the review for Monday to read data available through your connectors. Refresh any required exports, such as Amazon advertising reports, before each run.
Start this work
Download Orkas — freeTry this use case in Orkas
Free, open source, and runs on your machine.
FAQ
Store review questions
What sellers ask about connectors, spend authority and where the exports go.
What does connecting a store actually involve?
Configure the credentials required by each platform; they stay encrypted on your machine. Available data depends on the granted permissions and requested operations, and customer operations can include personal data. Check the access requested when connecting. Seller Central needs a Professional plan, an approved developer profile and a private SP-API app you self-authorise. Shopify is a Dev Dashboard app with the Admin API scopes you choose, a native connector rather than an MCP server you host. eBay is a developer app with official OAuth, and Etsy needs an app approved for Seller API Access. It is not install-and-go, which is why exports work from day one.
What can a connector read, and what still comes from exports?
Seller Central reads catalog, listings, non-personal order status, FBA inventory and financial data. Shopify Admin reads orders, customers, inventory, fulfilment, returns and refunds. eBay reads items, offers, orders, fulfilment and refunds; Etsy reads the shop, listings, inventory and orders. Amazon's advertising search term report is a different API and still comes in as the export, and so does anything a connector does not expose.
Will it change my bids, place a restock order or write to the store?
Not in this review. The connectors can write, but this scenario only reads: final sourcing, stocking, ad spend and pricing decisions stay yours by design, and the merchandising skills do not create campaigns or set budgets. Anything irreversible asks before it runs, and the output is a list of changes for you to make.
Can it run every week on its own?
Yes. Schedule the review to use the data available through your connected store and project files. Before each run, refresh exports that the connector cannot retrieve, such as Amazon advertising search term reports. Keep the reporting period consistent so weeks are comparable. Review the proposed to-do items in the project; the review does not make store changes by itself.
Where do my store data and credentials end up?
Connector credentials are encrypted on your machine. Files remain in your folder, but content read by the agent can be sent to your selected model provider. Remove unnecessary names, order numbers and addresses from exports before giving them to the agent. The review skill asks the model to omit personal identifiers from its output; this does not guarantee redaction before the model receives the file content.
Ready when you are
Start the next task in Orkas
Free, open source, and runs on your machine.