AI Order Processing for QuickBooks Desktop Enterprise: Order Demand Checking

The Message That Came Through

This arrived on our contact form from a manufacturer of commercial cleaning and janitorial products in central New Jersey, USA, which runs QuickBooks Desktop Enterprise:

"AI tool that will read customer POs and determine if any line item is being ordered at rate greater than that customer's historical demand"— A manufacturer of commercial cleaning and janitorial products in central New Jersey, USA

We recognised the problem immediately. It is not an unusual one for manufacturers with established customer accounts and predictable reorder patterns. The concern is not that any individual order is large — it is that a customer is on pace to take far more volume in a given period than they historically do, and that pattern only becomes visible when you look at their orders for the period together, not one at a time.

Reviewing purchase orders individually — which is what any manual order desk does — makes cumulative demand spikes easy to miss. The new order of 50 units looks fine. What you do not see, without checking, is that the same customer already ordered 70 units earlier this month against a baseline of 100. See also: Logistify AI.

Reviewing purchase orders One at a Time Hides the Pattern

When an order desk reviews a purchase order, they are looking at what is in front of them: the quantities on this order, from this customer, on this day. That is a reasonable thing to check. It does not catch what the order desk is not looking at: what this customer has already ordered this month, this week, or this quarter.

For commodity or high-velocity products, the volume a customer takes in a period matters more than the size of any single order. A customer who places five separate orders of 30 units in a month against a baseline of 100 is placing individually small orders. Their aggregate position of 150 units is a 50% overage — but each order, reviewed on its own, looked unremarkable.

This is the gap Demand Check is designed to close. It tracks cumulative demand by customer and product across the current period, compares it against the historical baseline, and flags the order that pushes the total over the allowed threshold — regardless of how routine the individual order size looks.

How Demand Check Works

When Demand Check is switched on, Logistify imports up to 18 months of sales orders from QuickBooks Desktop Enterprise. From them it works out a baseline for each customer and product: that customer's average total per period (day, week or month; weekly by default), across the periods in which they ordered that product.

When a new order is created or changed, Demand Check runs three steps:

• It adds up what the same customer has already ordered of that product earlier in the current period (day, week or month, depending on the setting), including orders already posted to QuickBooks Desktop Enterprise and leaving out cancelled ones.
• It adds the quantities on the new order to that running total to calculate the projected period demand.
• It compares the projected total against the customer's historical average for that period, plus the configured tolerance. If the projected total exceeds the allowance, the affected line items are flagged.

Flagged orders are held. A person is notified by email and in the app, and the order isn't sent to QuickBooks Desktop Enterprise until they approve it, giving a reason that's recorded on the order. If the order is edited after approval, the approval is cleared and the demand check runs again against the revised quantities.

The Calculation: Cumulative, Not Per-Order

The cumulative nature of the calculation is the part that matters most, and it is worth being precise about.

Say a customer's historical monthly average for Product A is 100 units. With a 15% tolerance, their monthly allowance is 115 units. Earlier in the month they ordered 70 units — that order posted to QuickBooks Desktop Enterprise without a flag, because 70 is well within 115. Now they submit a new order for 50 units.

Fifty units is not an unusual order for this customer. If you reviewed it in isolation, you would probably approve it without a second look. But Demand Check includes the earlier 70 in its calculation. The projected monthly total is now 120 — five units over the 115-unit allowance. The new order is flagged.

Logistify records the specific excess (5 units in this example) and suggests a quantity that would keep the customer within their allowance for the period: 45 units, since 70 are already ordered against the 115 allowed. The flag also records the customer's usual quantity, what they had already ordered this period, and the quantity on the new order.

What Happens When an Order Is Flagged

A flagged order does not fail silently and does not auto-post. It isn't sent to QuickBooks Desktop Enterprise, a warning naming the products that are over goes out by email and in the app, and the order's demand check records:

• The customer's usual quantity for the flagged product per period, and how many past periods it's based on.
• What this customer has already ordered of that product earlier in the current period.
• The quantity on the new order and the projected total if it posts as submitted.
• The allowance (historical average plus the configured tolerance) and the exact overage.
• A suggested adjusted quantity that keeps the customer within their allowance.

A person can approve the order as submitted, giving a reason, which is recorded. This is always an option, because Demand Check is a control, not a hard block on fulfillment decisions. They can edit the order to the suggested quantity, and the check runs again. Or they can contact the customer to verify the order before approving.

Until one of those happens, the order stays on hold. It does not appear in QuickBooks Desktop Enterprise.

How It Connects to QuickBooks Desktop Enterprise

Logistify connects to QuickBooks Desktop Enterprise through the QuickBooks Web Connector, which runs on the computer where QuickBooks is installed. Order, customer and item data passes through cloud services, including Logistify's, which keeps a copy of your customers, items and order history for matching and for the demand check. If your IT policy restricts external connections, review the setup with us before you start.

The 18-month history comes from your QuickBooks Desktop Enterprise sales orders. If your team invoices without creating sales orders, there's no history to compare against, so tell us before switching it on. When an order clears Demand Check, Logistify writes it into QuickBooks Desktop Enterprise as a sales order through the same Web Connector connection. Once a held order is approved, it is sent on too, so nobody enters it in QuickBooks by hand.

Demand Check works alongside Logistify's standard order intake features. Orders arriving by email (including PDF and spreadsheet attachments), EDI, WhatsApp, SMS or voicemail are read and extracted first, then run through Demand Check before they reach the QuickBooks Desktop Enterprise posting step. The demand check is not a separate workflow — it is a gate that every order passes through on its way to the ERP.

Configuring the Tolerance and Check Window

The 15% tolerance is the default. It is one setting for your whole company, not set per customer or per product. A business with tight allocation, such as regulated materials or limited-run components, may choose a narrower tolerance; one with seasonal peaks may choose a wider one to avoid flagging legitimate volume.

The check window is also one company-wide setting: daily, weekly (the default) or monthly. Pick the one that matches how your customers usually reorder.

Both settings are configured in Logistify without touching QuickBooks Desktop Enterprise. When either changes, open orders not yet sent to QuickBooks are checked again under the new setting.