The Sales Order That Was Almost Ready

At 9:12 a.m., Maya opens a sales order that is supposedly ready for production. "Supposedly" is doing some work there.

The customer wants 240 navy quarter-zips with an embroidered logo, and the in-hands date is next Friday. The order looks complete in the ERP, but Maya knows better than to trust how complete it looks.

First stop: artwork. The sales order says Approved, which would be excellent if the email thread agreed.

She finds the relevant message seventeen replies deep inside a thread titled RE: RE: RE: FINAL_v7_ACTUALFINAL.png. The customer wrote, "Looks good!" Two messages later, someone asked whether that meant approved for production, and there was no reply.

Maya adds "art?" to her mental list and moves on. It is 9:24.

Why does an SO readiness review turn into a scavenger hunt?

A sales order can be complete in one system while the information needed to release it lives across several others. Readiness is not one decision; it is a collection of small checks across artwork, supplier information, pricing, order details, customer communication, and dates.

Next: stock. The supplier site shows navy quarter-zips available, but Maya has been doing this long enough to know that "available" is a philosophical concept.

She checks the sizes. Small is fine, medium is fine, large is fine, and XL is seven short.

The sales rep's note says, "Supplier confirmed stock." Technically true, but not useful enough to release the order.

Maya adds "7 XL?" to the list. It is 9:39.

Next: decoration. The order says embroidery, which is a start, but it does not answer how many locations are being decorated.

The quote has one left-chest logo. The artwork folder contains that logo plus something called BACK_LOGO_USE_THIS.ai, and there is no decoration note explaining whether the back logo belongs to this order, an earlier version, or another dimension entirely.

Maya sends a message and gets no response. It is 10:03.

The SO readiness checklist

Before a promotional products sales order is released to production, the readiness review should confirm:

• Art approval: The correct proof version has been explicitly approved for production, because a "looks good" on an earlier version becomes a reprint once the decorator runs the wrong file.
• Stock availability: Every SKU is confirmed against live inventory by color and size, because "in stock" at quote time becomes a backorder at ship time when one size runs short.
• Decoration specifications: The method, locations, thread or ink colors, and artwork files are documented on the order, because the decorator produces exactly what the instructions say, not what the customer meant.
• Pricing vs. quote: Product pricing, decoration charges, setup fees, and freight match the approved quote, because a leftover setup fee or stale unit price turns into an invoice dispute after the goods have shipped.
• Dates: Production time, transit time, and the in-hands date still fit together, because an in-hands date entered at order time does not update itself when stock ships from two warehouses.
• Payment terms: Any required deposit, prepayment, or credit approval has cleared, because once blanks are ordered and decorated, the distributor owns that cost whether or not the customer pays.

If one of those is unresolved, the order is not actually ready. The checklist only works when every item has a clear source of truth.

Back to Maya, because pricing is next. The quarter-zips match the quote, and the embroidery charge matches too.

Then she gets to the setup charge. The quote says repeat logo, no new setup, while the sales order still has a setup fee.

It may be intentional, or it may be old data copied from the first version of the order. Either way, Maya now has another question to resolve before production.

It is 10:21, and now she looks at the date. The customer wants the order next Friday, production needs several working days, transit needs two, and the supplier stock is not all in one warehouse.

The in-hands date was entered before any of these facts joined the conversation. Maya stares at the calendar, and the calendar does not apologize.

Why has this process stayed manual for so long?

Because the checklist is easy, but gathering the evidence is not. An experienced coordinator can tell whether an order is ready, yet the slow part is finding the proof: the approval email, stock result, decoration notes, quote, production timing, and payment status.

By 11:00, Maya finally has the answer. The order is not ready, but not because six things are broken.

One thing is actually broken, and five other things had to be manually investigated to discover that. The XL stock shortage needs a decision, while the rest can be resolved from information that already exists somewhere.

What would the same readiness review look like with an AI agent?

Now replay the morning with the same sales order. Instead of immediately landing on Maya's desk, an agent starts gathering the evidence from the systems involved in the workflow.

It checks the order details, finds the relevant artwork and approval status, checks available supplier inventory, pulls the decoration details associated with the order, compares pricing with the approved quote, and gathers the relevant dates and payment conditions.

Trelium agents can work across email, attachments, connected business systems, supplier data, and proof or inventory workflows. When something is ambiguous or requires judgment, the workflow can route that exception back to a person instead of guessing.

Maya gets a readiness summary showing the state of each check. Art approval is confirmed, decoration is left chest only, pricing matches the approved quote, dates are within the configured requirements, and payment terms are satisfied.

Stock is the only exception. Seven XL units are unavailable, so that is the problem Maya actually needs to solve.

Now the job is one decision instead of six searches. The work has not disappeared, but the scavenger hunt has.

Does the AI agent make the final readiness decision?

It does not need to make every judgment call. A useful readiness workflow gathers the evidence, runs the repeatable checks, and surfaces missing or conflicting information so the responsible person can make the call.

Maya still has to decide what to do about seven missing XL quarter-zips. She might substitute another product, split the shipment, or ask the customer whether those units can arrive later.

That is operational judgment. Searching RE: RE: RE: FINAL_v7_ACTUALFINAL.png is not.

What is an SO readiness review?

An SO readiness review is the final verification that a sales order has everything required to move into production. For promotional products orders, that commonly includes artwork approval, stock, decoration instructions, pricing, dates, and payment requirements.

What should a sales order readiness checklist include?

At minimum, check art approval, stock availability, decoration specifications, pricing against the approved quote, production and in-hands dates, and any payment terms that must be satisfied before production. The exact rules may differ by distributor, but each check should have a clear source of truth and a defined exception path.

Can AI automate order readiness checks?

AI can automate much of the gathering and repeatable verification around an order readiness review. The useful model is not "AI approves everything"; it is "the agent assembles the readiness picture, clears the obvious checks, and gives a person the exceptions that actually require judgment."

By lunchtime, Maya is still the person making the call. She just gets to start with the problem instead of spending her morning proving that everything else was fine.

If you want to see what that kind of readiness workflow could look like across the systems your team already uses, see Trelium in action.