AI for print-on-demand and decorated apparel Find the margin your system isn’t showing you.

An AI intelligence layer that works with the system you already run. It warns you before money is lost, and tells you what to do about it.

Bring us one problem → See what it finds

Loss-making orders 38% up from 22%flagged the month it crossed 30%

Before you startJob 48213

120 × black tee · DTG

6.1%expected spoilage · shop average 1.9%

Do this first: raise pre-treat to 36 g and pull 8 spare blanks.

Illustrative data

The industry right now

Costs are rising faster than prices.

73%of decorators saw costs rise. Only half raised prices.

70%saw profitability stay flat or fall.

55%name tariff costs and supply disruption as a top risk.

Your system records what happened. It doesn’t show which order, which job or which vendor is eating the margin, or what will go wrong next.

Source: PRINTING United Alliance, State of the Decorated Apparel Industry, Winter and Summer 2025.

Today’s options

Every option works, up to a point.

Option	What it does well	Where it falls short
Spreadsheets and gut feel	Free and familiar	Margin leaks stay invisible until they hurt
Stay on current tools	Good at orders and job tracking	Records what happened, with little analysis
Move to an all-in-one AI platform	Strong AI across operations	Means migrating onto one vendor’s stack
Hire an analyst or build BI	Tailored to your business	Slow, after the fact, and depends on a few people

So what would the right fix look like?

The perfect world

If you designed the fix yourself, it would tick these boxes.

Keep the system you already run. No migration.
Find margin leaks in specific, named places.
Warn you before a problem, not after it.
Work on your own data in weeks, not months.
Keep your data yours, safe and under your control.
One flat, predictable cost.

Not another report

Most tools stop at what happened. Torno tells you what will happen, and what to do.

What happened

Reporting

768 blanks were spoiled last month.

Why it happened

Analysis

Pre-treat was too light on Press 4, night shift.

What will happen

Prediction

Job 48213 has a 6.1% spoilage risk before it starts.

What to do now

Action

Raise pre-treat to 36 g and pull 8 spare blanks first.

It reads every record, every day, with the full history in view, so it doesn’t overlook what a busy team would.

Where most shops start

Three places your margin leaks, found before they cost you.

01 · Cost-to-serve

Know which orders make money, not just which clients.

What it findsThe true margin on every order after setup time, run size, reprints, returns, freight and payment terms.
What it predictsWhen the share of loss-making orders starts to climb, months before it shows in the P&L.
What you doSet pricing rules by run size, batch small jobs, and take loss-making orders as a conscious choice.

Data we use: order history, job time records, reprints and returns, freight invoices.Industry context: decorators’ operating costs rose 6.8% on average while prices rose 5.3% (PRINTING United Alliance, Winter 2025).

Cost-to-serve by orderIllustrative

Orders analysed4,812

Loss-making orders38% up from 22%

Margin lost on them≈ $61,400

Worst segmentRuns under 25

Share of orders losing moneyby month

alert at 30%

flagged

38%

Margin by run sizeafter setup

1–24

−6.2%

25–99

+4.1%

100–499

+12.8%

500+

+18.5%

Recommendation

Runs under 25 units lose 6.2% after setup time. Add a $1.10-per-unit setup charge below 25 units, or batch them with same-artwork orders.

Impact≈ $38,000 a year

back in margin, with the loss-making share flagged the month it starts rising.

02 · Spoilage

Stop spoilage before the press starts.

What it findsEvery spoiled unit, and why: blank, vendor batch, print method, press, shift and artwork.
What it predictsThe spoilage risk of each job before it starts, from how similar jobs went before.
What you doA check for the operator at job start, and spare blanks pulled so a reprint never waits on a new order.

Data we use: purchase orders, job records, shipments and any misprint logs you keep.Industry context: 1–3% spoilage is accepted as normal. At 2%, a shop printing 1,000 garments a day writes off about 5,000 a year.

Spoilage · before the job startsIllustrative

Blanks bought (30 days)31,420

Spoiled768 2.4%

Cost of spoilage≈ $4,600

Jobs flagged before start126

Why units were spoiledtraced per unit

Pre-treat too light

298 · 39%

Blank defect

143 · 19%

Misregistration

117 · 15%

Artwork error

92 · 12%

Handling / other

118 · 15%

Before you startJob 48213

120 × black tee, DTG front

6.1%expected spoilage · shop average 1.9%

Blank: Gildan 5000 Black, Supplier D batch 2291
Past jobs: 14 similar, 11 with pre-treat faults

Do this first: raise pre-treat to 36 g and run a test print. Pull 8 spare blanks now.

Done, start jobSkip

Impact2.4% → 1.6%

spoilage by fixing the top cause first, about $12,000 a year at 1,000 garments a day.

03 · Delivery drift

See an order drifting off schedule while there’s time to save it.

What it findsWhich orders are slipping, and why: late blanks, a full press queue, unapproved artwork or a short PO.
What it predictsThe chance each order misses its ship date, updated as the day changes.
What you doA specific fix for each: another supplier or press, an approval to chase, a PO to top up.

Data we use: purchase orders, receiving dates, the production schedule and ship dates.Industry context: 55% of decorators name tariff costs and supply disruption as a top risk for 2025 (PRINTING United Alliance, Summer 2025).

Delivery · orders drifting off scheduleIllustrative

Orders due this week612

Drifting off schedule14

Warning given3.2 days early

On-time rate96.1% target 98%

Orders at risk, with the cause and the fixupdated hourly

Order	Ships	Risk	Why it’s drifting	Do this
48190	Thu	82%	Blanks from Supplier D running 3.2 days late	Switch to Supplier A (in stock)
48207	Fri	64%	DTG queue over capacity on Thursday	Move to Press 2, Wednesday night
48213	Fri	57%	Artwork still waiting for approval	Chase approval today
48221	Mon	41%	Purchase order 36 blanks short	Top up on the open PO

ImpactNo rush freight

on 11 of 14 drifting orders, because the warning came while rerouting was still cheap.

And more

These three are where most shops start, not our limits.

Production planning

Sequence jobs and steps in the order that wastes the least time and movement.

Capacity and bottlenecks

See which press or step will jam next week, before it does.

Labor on the floor

Spot time spent moving and searching instead of producing.

Artwork and prepress

Catch artwork problems before they reach a press.

Purchasing and inventory

Buy blanks against predicted demand, not last month’s rush.

Returns and reprints

Trace each return back to the job, press or vendor behind it.

We add them one at a time, as each earns its place, on the same layer and the same subscription model.

How it plugs in

No rip-and-replace. It reads what you already have.

1
Connect
Read-only access to your system.
2
Map
Understand your data, and what’s missing.
3
First findings
On one workflow, on your own data.
4
Review
Go through the results with you.
5
Expand
Add the next module only if it earns it.

Where data is missing, we tell you what to capture and what it will be worth, extend the module that should record it, or build the module if it doesn’t exist.

A flat annual subscription per module, scoped with you.

We build it and prove it on your data first. The subscription starts only once it works. No one-time project fee.

Your data stays yours

Built on security, governance and compliance.

Security

Read-only access by default.
Your data stays in your environment.
Works with your existing access controls.

Governance

Never shared with other clients.
Never used to train anyone else’s models.
A full log of what the layer read and suggested.

Compliance

Bring your own LLM: connect the model account your organization already approves.
You decide which data the layer can see.

Your competitors never see your numbers, and you never have to trust a black box with them.

Case study · Stakes Manufacturing

Modernizing a print-on-demand apparel manufacturer without stopping production.

Torno’s hybrid Pod gave us onshore alignment with offshore velocity, exactly what we needed. Tim ShodaChief Technology Officer, Stakes Manufacturing

0 pausesProduction kept running through a full ERP modernization.

13+ monthsEmbedded with the Stakes team.

Module by moduleOrder intake, purchasing, receiving and warehouse.

AI layerShows where spoilage is happening across the company.

Trusted by teams at Crateful

Questions you haven’t asked yet

The concerns we hear before anyone says them.

Is this another system to learn?

No. It works inside what you already run.

Isn’t this just reporting?

No. It predicts problems before a job or order goes wrong, and tells you what to do.

We don’t have enough data.

We start with what you have, tell you what to capture next, and extend a module if it doesn’t record it.

Is our data safe?

It stays in your environment, is never shared or used to train other models, and you can use your own LLM.

Is this a big project?

There’s no one-time fee. It’s a flat annual subscription per module, and it starts once it works.

Why not an all-in-one AI platform?

That’s a fair choice if you want to consolidate. If you want to keep your system, this adds the intelligence without a move.

Our analysts already do this.

After the fact. This reads every record every day and warns before the job, for less than an analyst costs.

Start here

Let’s pick the one problem to start with.

Which costs you most? Pick one, or bring your own.

Book a 30-minute call → Topic: Spoilage & waste

1
Pick the problemThe one costing you most.
2
A demo on your dataWe build it on your own records, read-only.
3
Pay once it worksThe subscription starts only then.