AI & Artwork. A View From Us About The Future

AI won’t replace artwork. It just changes the work that needs to be done.

AI is beginning to reshape one of the least visible, but most operationally critical parts of the design process: technical artwork.

This is perhaps not the glamorous end of our industry and it’s certainly not where brands are created or invented. It’s where packaging becomes real and hours of work gets translated into hundreds of SKUs often having to involve multiple languages, perplexing regulations, and increasingly inventive print and production requirements. It is here where precision matters most and where errors are frankly unthinkable.

AI is just another tool in the artworkers tool kit. A tool for reducing the friction in complex, repeatable work and the impact we are seeing at TBN is already tangible.

From ‘Craft’ to ‘System’

Artwork has always been a craft discipline built on experience and attention to detail. The work is structured, but the process has largely been manual and heavily reliant on human checking, version control and a filing system that doesn’t say ‘artwork V1’ (if you know, you know). That kind of workflow and model really does not bode well for the future.

As ranges expand and as clients demand more variants the volume of artwork has grown exponentially. A single product range can now generate hundreds of individual files. Each one needing to be accurate, compliant and consistent. This is where AI works really well. Think of it as an organiser of complexity as opposed to an out and out threat.

AI systems are particularly effective in environments where the rules are clear, the data is structured, and tasks are repetitive. Packaging artwork fits that description almost perfectly. As we look at the work that goes through our studio what we are witnessing and designing our systems for is that artwork is no longer a series of individual files but a connected, data-driven system.

Removing friction, not adding magic

There is a tendency to overstate AI’s role in the creative industries. In volume artwork, the reality is more grounded and arguably more useful. AI is not replacing artworkers. It is removing the parts of the job that are the most time-consuming, most error-prone, and least valuable (time wise).

Take compliance. Food packaging artwork is governed by complicated and constantly evolving regulation. Ingredient lists, nutritional tables, allergen declarations, barcodes, legal copy, front of pack compliance… There is no margin for error in any of these things.

Traditionally, this has required multiple rounds of manual checking. Slow, meticulous work, often repeated across dozens of SKUs. AI helps change this part of the process. It can scan artwork in real time, flag for missing or incorrect elements, and validate the artwork created against a set of predefined rules. AI doesn’t get tired, it does not overlook inconsistencies, it just applies the same logic every time. However, it does hallucinate! That’s why we need people to oversee the output and make sure that every nuance is covered, and the systems we make manages to learn every peculiarity of the brands we work for.

Scaling the unscalable

If compliance is one pressure point, versioning is another. Food packaging rarely exists as a single, fixed design. It exists as a system of variations. Flavours, formats, languages, promotions with each variation introducing an element of complexity.

Historically, this has been managed through duplication and manual updates. Change one element, then replicate it across every relevant and or connected file. Check each one individually and fingers crossed nothing is missed.

AI helps reframe this by understanding artwork as structured data rather than static design. AI systems can identify what needs to change and propagate those changes automatically. Update an allergen statement once, and it can be applied consistently across an entire product range. Adjust a legal address, and every instance is updated in parallel.

Just as importantly, AI can compare versions and detect unintended differences, those small, easy to miss errors that often slip through manual processes. For studios handling large volumes of packaging, this is transformative. It turns a fundamentally linear process into a scalable one.

What AI doesn’t do

For all its strengths, AI has clear limitations. It does not understand ‘brand’ in any meaningful sense. It does not make nuanced typographic decisions. It does not interpret creative intent or navigate client dynamics.

It cannot resolve ambiguity where rules are unclear, or where judgement is required.

In other words, it cannot replace the human layer of the process and is unlikely to in the foreseeable future. What it does instead is create time and space.

By automating some of the technical and repetitive aspects of artwork production, it allows artworkers to focus on the areas where judgement still matters. It shifts the artworkers’ role from execution to oversight. From ‘doing the work’ to ‘managing the system’ that does the work. That is a hugely meaningful change, but it is not one that will see jobs disappear.

What happens next

The current wave of AI in artwork is still relatively contained focussing on validation, automation, and workflow management. The next phase will go further.

We are likely to see tighter integration between artwork systems and live regulatory databases, reducing the lag between rule changes and implementation. More sophisticated template generation, where master artworks are built with automation in mind from the outset. Greater connectivity between design, approval, and print production, creating more seamless end-to-end production workflows.

In that context, the role of the artworker evolves again from operator to controller, to orchestrator. Never removed from the process (because knowledge is power) but repositioned within it.

The bottom line

AI is not transforming packaging artwork by making it more creative. It is transforming it by making it more reliable. It is removing friction from complex workflows and is helping in reducing the likelihood of error. It allows systems to scale in ways that manual processes could possibly never do.

For all studios including us, the opportunity is not to replace people, but to redesign the processes around them. We will use AI to handle what machines are good at and perhaps most importantly we will preserve human judgement where it still matters most.

In technical artwork, that balance is not just desirable. It is essential.

Words by James Acton