What does an Inventory FTE own?
An Inventory FTE on ftes.ai owns inventory planning for an online store as a job function: knowing what will sell, when stock will run out, what to reorder, from whom, in what quantity, and what that costs in cash. It is accountable for one outcome, stock that matches demand with every purchase order approved with its cash impact, and it is reviewed on targets such as no stockout on a top SKU and cash in dead stock below a ceiling.
It does not own supplier negotiation, payment terms, or the decision to add or retire a product. The owner decides those. The Inventory FTE prepares the evidence, proposes the orders, and reports what it measured, inside a cash ceiling per month it cannot exceed even in proposals.
How does the forecast work, and how is confidence shown?
The Inventory FTE forecasts demand per SKU from the store's own sales history, adjusting for seasonality where the history shows it, for promotions the owner has scheduled, and for trends in recent velocity. Each forecast carries a confidence label derived from how much history the SKU has and how stable its sales have been: a SKU with two years of steady sales is high confidence; a SKU launched three weeks ago is low confidence and is labelled as such on every screen it appears on.
Confidence is not decoration. A low-confidence forecast produces a smaller, more frequent order proposal and a note asking the owner for any knowledge the data lacks, such as a planned campaign. The FTE never presents a guess as a number.
How are reorder points and safety stock set?
A reorder point is the stock level at which an order must be placed so that new stock arrives before the current stock runs out. The Inventory FTE calculates it per SKU from forecast demand over the supplier's lead time, plus safety stock that covers the variance in both demand and lead time. Lead times come from the store's own purchase order history per supplier, with the variance measured, not from a single default the store typed in once.
The owner sets the service level the safety stock protects, which is a business decision about how much stockout risk to accept on which SKUs. Top sellers usually get a higher service level and more safety stock; slow SKUs get less. Each reorder point shows its inputs so the owner can challenge them.
What does a purchase order proposal contain?
A purchase order proposal in the Approvals tab contains the SKU or SKUs, the supplier, the quantity and why that quantity, the supplier's lead time with its measured variance, the unit and landed cost, the cash the order ties up and when it falls due, the days of cover it provides, and the forecast with its confidence that drove it. Minimum order quantities and price breaks from the supplier record are applied and shown.
The approver the owner named approves, edits the quantity, or rejects with a note. Orders above the threshold the owner set need a second approver. Nothing is sent to a supplier until that happens; the Inventory FTE cannot place an order at any autonomy level.
How are stockouts and dead stock surfaced?
Every morning the Inventory FTE lists the SKUs at risk of stockout within a horizon the owner set, usually fourteen days, ranked by the revenue the store would lose, with the purchase order that would prevent each. It also lists dead and slow stock, SKUs whose cover exceeds a threshold at current velocity, ranked by the cash trapped, with a proposal for each: a markdown for the Conversion or Retention FTE to run on approval, a bundle, or a return to the supplier where terms allow.
Where the store also employs a Marketplace FTE, the stockout list is visible to it before it proposes advertising a SKU that is about to run out, which is how the roles work as a department rather than a set of tools.
What does it report, and how honestly?
The Inventory FTE reports four things it measures: stockouts that actually happened on which SKUs and for how many days, cash in inventory against the ceiling and the share of it in dead stock, forecast accuracy per SKU comparing last month's forecast to what sold, and purchase orders proposed, approved, edited and rejected. Forecast accuracy is reported plainly, including the SKUs it got wrong and why, because a forecasting role that hides its misses cannot be trusted with the next order.
It does not report a stockout prevented, because a prevented stockout is a counterfactual. It reports the stockouts that happened and the cover it holds today.
Which tools does it work in, and why is onboarding custom?
The Inventory FTE reads sales and stock from the store platform, currently Shopify, WooCommerce and BigCommerce, or from an inventory or ERP system such as NetSuite, Cin7 or Linnworks, and reads supplier lead times from the store's purchase order history or a supplier sheet. Onboarding is custom: the supplier records, lead times, minimum order quantities, landed costs, service levels and the cash ceiling are agreed with the ftes.ai team before the role starts, because inventory proposals built on wrong costs or lead times are confidently wrong.
Credentials are encrypted at rest and revocable, the work is reviewed in the ftes.ai workspace, and the role draws tokens from the organization wallet with a hard stop at zero.