Printing Has the AI Tools. Is the Workflow Ready?

Most print businesses now have access to AI. It is built into preflight tools, colour management, estimating systems, scheduling software and maintenance platforms. The harder question is whether the business underneath is ready to use it.

Writing for WhatTheyThink, Pat McGrew and Ryan McAbee point to Deloitte's 2026 Global Leadership Technology Study. More than 80% of senior IT leaders said they were confident their organisations could deploy and govern AI at scale. Yet 75% of the same leaders said their operating models and business processes needed major changes within 12 to 18 months before AI would actually deliver value. Their summary is blunt: the tools are not the problem, the way work actually gets done is.

That is exactly what we see in print. AI added to a workflow full of manual handoffs, repeated data entry and undocumented knowledge does not fix those things. It just runs into them faster.

Why print is particularly exposed

A typical commercial or wide-format job touches a lot of systems and a lot of people between enquiry and invoice. In many businesses the route looks something like this:

• The order arrives by email, portal or phone, and someone types the details into the MIS.
• Artwork arrives separately and is matched to the job by hand.
• Preflight results are read by an operator and passed back to the customer in an email.
• Imposition, colour and finishing decisions rely on what one or two experienced people know.
• Production status lives on the production floor, not in the system, so the office chases updates by walking out to the press.

Put an AI estimating assistant or an AI scheduling tool on top of that and it can only work with the data it is given. If job specifications are re-keyed with errors, if status is out of date, or if the real process differs from the documented one, the AI produces confident answers built on bad inputs.

Five signs your workflow is not ready for AI yet

1. Job data is typed more than once. If the same specification is entered in a web-to-print system, an MIS and a prepress tool, each copy is a chance for a mismatch.
2. Nobody can draw the real process. Ask three people how a reprint is handled and you get three answers.
3. Exceptions go through one person. When a file fails, the fix depends on who is in that day.
4. The MIS does not reflect the floor. Job status is updated at the end of the day, if at all.
5. There is no baseline. You cannot say how long a job waits in prepress, how many files fail preflight or why reprints happen.

None of these are AI problems. They are workflow problems, and they need fixing whether or not AI is part of the plan.

What to fix first

One entry point for job data

Job details should be captured once, at the point of order, and passed through the rest of the chain automatically. That usually means connecting the storefront or order intake to the MIS, and the MIS to production. We cover the mechanics in how to connect your MIS to your print production workflow and web-to-print integration.

Rules-based automation before AI

Most of the repetitive decisions in prepress are rules, not judgement: check the file against a standard, fix what can be fixed safely, route it to the right hot folder, impose it, report back. Tools such as Enfocus Switch and callas pdfToolbox handle these reliably, and every rule you automate is a rule that is now written down. That is the foundation AI needs.

Measure before and after

Record touches per job, time in each stage, preflight failure rates and reprint reasons. Without a baseline you cannot tell whether any tool, AI or otherwise, has made a difference. Our guide to print workflow automation ROI shows how to turn those numbers into a business case.

Where AI genuinely fits

Once data flows cleanly and the routine is automated, AI has something useful to work on. In our experience it works best in three roles:

• Suggest: proposing an estimate, a schedule change or a fix, which a person approves.
• Prepare: drafting scripts, configuration or customer replies that a person checks before use.
• Act with review: only where the action is reversible and the result is checked automatically.

Anything that commits to a customer or cannot be undone, such as plating, cutting or a price, should stay with a person. We look at the prepress side of this in more detail in AI in print: necessary evil or genuine game-changer?

The order matters

AI will become a normal part of print production. The businesses that get value from it will be the ones that fixed their workflow first: a single source of job data, rules-based automation for the routine and clear measures of what is happening. The ones that buy the tool first will mostly find out where their process is broken.

If you want an honest view of where your workflow stands, a workflow audit maps it end to end, or you can get in touch directly.

Source: Printing Companies Have the AI Tools. What About the Workflows? by Pat McGrew and Ryan McAbee (WhatTheyThink), citing Deloitte's 2026 Global Leadership Technology Study.