AI inventory management agents, built on the stock data you already have.

Pick by job. The products we read on 11 Oct 2026 each lead with one of the three. For inventory planning on Shopify, Prediko lists plans from $49 a month and Inventory Planner quotes on request. For stock and orders across many sales channels, Cin7 and Linnworks are whole systems. Customer messages are a separate job, covered on our ecommerce AI agents page. FactoryJet builds agents for the work that crosses systems, such as stock in an ERP and orders in a store, with a person approving anything that spends money.

Going by the vendors' own pages, read on 11 Oct 2026: Linnworks lists a Rules Engine with unlimited order routing rules, plus Spotlight AI, which finds repetitive manual work to automate. Cin7 lists an AI assistant, AI forecasting and document recognition that turns a customer's purchase order into a sales order. Shopify's help pages describe Sidekick analyzing stock balances and creating transfers. When routing rules span an ERP, a warehouse and a store, FactoryJet builds the agent around your own rules. Customer service agents are on our ecommerce AI agents page.

It is software that reads your stock, sales and purchase records and does the repeated thinking for you. It tells the buyer when an item reaches its reorder point, drafts the purchase order, keeps the store and the ERP showing the same count, and lists stock that has stopped selling. Some products add a demand forecast. People still approve orders, count the shelves and decide what to stop selling.

This page covers five jobs. Low-stock alerts that say why an item was flagged. Draft purchase orders for a person to approve. Stock figures kept matched between a store, a warehouse system and an ERP. Slow movers flagged for a decision. Stock questions from your own staff answered from live records. Demand forecasting is the next step, and it works best once those five have put clean history in one place.

It is a system that updates stock counts and raises reorders without someone typing each change. The ERPs we read each do a version of it. Odoo creates a request for quotation when forecasted stock falls below a minimum. ERPNext raises a Material Request at a reorder level. NetSuite lists items below their reorder point on its Order Items page with a suggested quantity. An AI agent works across several systems at once and explains each flag in plain words.

There is no neutral ranking. On 11 Oct 2026 the first page of US Google results for "ai inventory management" was eight guides and lists from software and services firms. Shortlist by where your stock is kept. A single Shopify store can start with a planning app such as Prediko. A business on an ERP can look at StockTrim or Netstock, which sit on top of one. If stock lives in several systems and the rules are your own, have agents built. Ask any vendor for a trial on your own data.

It is the idea that a small share of your items brings in most of the value. NetSuite's guide to ABC analysis says the method identifies the 20% of goods that deliver about 80% of the value. In practice it tells you where to look first. Count and reorder those items with the most care, and let simple minimum and maximum rules handle the long tail.

It sorts stock into three classes by value. NetSuite's guide puts class A at 10% to 20% of items and 70% to 80% of annual consumption value, class B at 30% of items and 15% to 20% of value, and class C at 50% of items and 5% of value. An agent can re-run the sort every week and flag items that have moved class, which is one way to spot a slow mover early.

It depends on your trade, so compare with your own past before anyone else's. For scale, the Census Bureau put the total US business inventories to sales ratio at 1.30 at the end of July 2026, against 1.37 a year earlier. That means businesses held about 1.3 months of sales in stock. Your own figure by product line, tracked month by month, is the more useful number.

Multiply the supplier's lead time by your average daily sales, then add safety stock. Oracle's NetSuite manual gives that sum and an example, an item with a three-week lead time selling 25 a week, which needs a reorder point of at least 75. ERPNext documents the same formula. NetSuite's defaults are a 14-day lead time and 7 days of safety stock when an item has none set, so check them before trusting a suggested order.

Demand forecasting estimates future sales so you can order before the need shows up. The simple version averages past sales. NetSuite's manual shows one, where three months of 72, 83 and 94 units average 83 a month, which it divides by 30 for a daily rate. Better forecasts add season, promotions and supplier lead times. FactoryJet designs and builds a forecast on your own history as the step after the agents, and you test it on months you have already lived through.

These are list prices from the makers' own pricing pages, read on 11 Oct 2026. Prediko starts at $49 a month for the smallest Shopify stores. StockTrim showed $49 a month for under $100K of revenue and $298 a month for $3M to $5M, billed annually. inFlow runs from $129 to $879 a month. Cin7 Core runs from $349 to $1,199 a month. Netstock starts at $900 a month. Katana lists a free plan and custom pricing above it. Inventory Planner and Linnworks quote on request.

A few, with limits. On 11 Oct 2026 Katana's pricing page listed a Free plan at $0 for up to 30 SKUs, a SKU being one stock item. Prediko and StockTrim each listed a 14-day free trial. Your ERP may already include reorder rules. Odoo and ERPNext both raise a draft order or a request from levels you set. Free plans suit a small catalog. Past that, price the paid tier before you move your data in.

Yes, for a small catalog in one location with one person keeping the sheet. It stops working when two people edit at once, when stock sits in several places, or when the store sells while the sheet is closed. The sums on this page, such as reorder point and days of cover, are easy to set up in a spreadsheet. The next step up is usually the reorder rule inside the system you already own, and an agent comes after that.

Start with what you own. Shopify has purchase orders and transfers in the admin, QuickBooks Online Plus and Advanced have low stock alerts, and Odoo, NetSuite and ERPNext have reorder rules. If that falls short, a planning product adds a forecast, and a full system such as Cin7, inFlow or Linnworks takes over from the spreadsheet. The eight products on this page are grouped by the kind of buyer each suits, with list prices from their own pages.

When you sell through one store or one ERP, your reorder rules are a minimum and a maximum, and you want a forecast this month. Prediko and StockTrim each list a 14-day trial, so you can test on your own data before paying. Built agents earn their place when stock lives in several systems, when your rules come from supplier terms, or when purchase orders must land in your ERP in a set format.

No. FactoryJet is a services company. We design, build and support AI agents on the ERP and the store you already run, and you own what we build. There is no FactoryJet inventory app to subscribe to. If a product you can rent this week would cover your case, we name it on the scoping call, and the eight on this page are a fair place to start looking.

Shopify's help page for Stocky carries a notice that the app will no longer be available after August 31st, 2026, and tells merchants to move to Shopify's own inventory management features. Stocky tracked stock levels, forecast what you needed and suggested which products to order. Shopify's admin has purchase orders and inventory transfers. If you relied on Stocky's reorder suggestions, the choices are a planning app such as Prediko or an agent built on your own rules.

Yes, with the Advanced Inventory Management feature. Oracle says NetSuite uses past purchases and sales to calculate a reorder point and a preferred stock level for each item, and lists items below their reorder point on the Order Items page with a suggested quantity. The defaults are a 14-day lead time, 7 days of safety stock and a 6-month look back, recalculated weekly. An agent adds the alert, the reasoning and the draft purchase order for your approver.

Each rule holds a minimum and a maximum for a product at a location. When forecasted stock drops below the minimum, Odoo creates a request for quotation for a bought product, using the vendor at the top of the product's list. The trigger is automatic by default, and the scheduler runs once a day by default. Set the trigger to manual and the item appears on a replenishment report for a person to order. Odoo also rounds to a pack size you set.

Yes. SAP's Service Layer, the API for SAP Business One, lists PurchaseRequests, PurchaseOrders and Drafts, so an agent can raise a request or leave a draft for approval. In ERPNext the natural place is a Material Request with its purpose set to Purchase. FactoryJet has worked on both systems, on jobs that included purchase and sales orders. In our builds the agent drafts and a person approves.

It means keeping one true stock count while you sell in several places, such as your own store, a marketplace and a wholesale desk. The hard part is which number to share. Shopify's developer pages show an item with 101 on hand, 29 committed to orders and 72 available. Share 101 and you can sell stock that is already promised. Linnworks and Cin7 are built for this job. An agent is the route when one of your systems is custom.

It depends on how many systems the agents read and write and how many rules they follow. Layer3 Labs' 2026 guide to agency pricing gives $5,000 to $15,000 as typical for a single-workflow build and $15,000 to $50,000 for a project of several workflows. ProductCrafters, a development firm, titles its 2026 breakdown of custom AI agent builds $5K to $180K+. The count of systems and rules is what moves a project along those ranges. FactoryJet quotes a fixed price in writing after a short scoping call.

We do. FactoryJet quotes the build and the ongoing support as fixed prices in writing after a short scoping call. After launch, the team that built the agents keeps managing the servers, the AI models, the API connections and the maintenance, so your operations team does not have to. That includes re-testing when Shopify, your ERP or the AI model changes. The service is described on our AI agent monitoring and support page.

Two to four weeks is usual for a pilot on one narrow workflow, which for inventory means low-stock alerts on one system going to one buyer. A production build, with permissions, logging, approvals and monitoring, usually takes six to twelve weeks. Stock sync between a store, a warehouse system and an ERP sits at the longer end. At first a person checks every alert and draft, so you see results on live stock before the agents are trusted with more.

Yes. Before a contract, we show working software on your own data. For inventory that means an export of your stock, open orders and supplier list, run through a first version of the agents. You check its alerts and draft purchase orders against what your buyer did in the same week. If a reorder rule in your ERP or a rented product would do the job as well, that is when we tell you.

No. In our builds the agent drafts and a person approves. The draft sits in your ERP in its normal place, such as a request for quotation in Odoo or a purchase order pending approval in NetSuite, with the reasoning attached. Your buyer changes it, approves it or deletes it. The agent's sign-in can read stock and create drafts. It cannot approve an order or send one to a supplier.

Yes. It is the step after the five jobs on this page, because a forecast needs clean stock, sales and lead-time history in one place, and the agents produce that. We design and build the forecast on your own records and wire it into the reorder point, so alerts move with the season. No accuracy figure is quoted here. You judge it on months you have already lived through, and your buyer can override any number.

You do. The agents are built for you, and the code, the written rules and the test data are yours. They can run in your own cloud account. We still look after them day to day. If you later change ERP, bring the work in-house or move to another firm, everything goes with you.

It stays in your store and your ERP. To write an alert or answer a staff question, the agent sends the records it needs to an AI model. Before the build we name that provider, and every other system that touches your data, in writing, so your IT lead can review the list. Each agent signs in with a user of its own, limited to the records its job needs.