The role of AI in revolutionising packaging: creativity, efficiency, and sustainability

AI in packaging might not seem like the most obvious fit, but its potential in this creative, technical, and complex industry is undeniable.

For those in the packaging industry, it’s easy to think of AI as something reserved for tech startups or industries far removed from the physical world. However, over the past few years, we’ve been exploring how AI can simplify processes, reduce waste, and even inspire creativity in packaging. Here are some of our observations so far:

‍

LLMs as magical duct tape

The phrase “LLMs as magical duct tape” comes from Alex Komoroske, whose insights resonate with our own experience. The idea is that Large Language Models (LLMs) are versatile tools with countless use cases, often applied in personal, unique ways.

In packaging, we’re already seeing LLMs leveraged in diverse contexts. Logistics agents use them for real-time translation, while senior managers train teams on Excel using free AI tools. Even outside of work, people are adopting LLMs for tasks like organising notes or studying. For example, an Uber driver shared how he uses ChatGPT in voice mode to prepare for exams between rides.

While these applications once felt like clever party tricks, recent advancements like the o1 preview have transformed LLMs into powerful tools for deliberate reasoning. They can now function as research assistants and analysts, capable of pausing, evaluating, and reasoning through decisions.

In the packaging manufacturing space, this means tackling tedious administrative tasks like order processing, estimating, or delivery coordination. Instead of buying rigid software solutions, companies could adopt AI agents capable of completing collections of tasks. This approach not only reduces bottlenecks but also frees up teams to focus on higher-value, human-to-human interactions.

While LLMs are not perfect and often require skilful prompting to deliver accurate results, their potential to drive efficiency and better outcomes for brands and consumers is undeniable.

‍

Fostering ‘Weird’ Creativity

Creativity often involves breaking through fatigue or mental blocks, and AI has become an incredible sandbox for sparking new ideas.

We recently spoke to a product designer who admitted that inspiration sometimes comes slowly. This challenge, whether at the start or middle of the creative process, is familiar across the industry. AI tools can help here by acting as brainstorming partners or offering fresh, unexpected perspectives.

For example, I often use AI in voice mode to brainstorm—it’s far more engaging than typing. Generative platforms like websim.ai provide another avenue, producing “weird” but thought-provoking visuals, such as a cardboard box made of crystals.

This concept inspired our work on Spring AI, a tool designed not to replace designers but to enhance their creative processes. By generating unconventional ideas, these tools expand the boundaries of what’s possible, making design more scalable and playful than ever before.

The more we engage with these AI-driven sandboxes, the more likely we are to uncover new ways to improve design and achieve better outcomes.

‍

Optimising for Win-Win Outcomes

Beyond LLMs, more traditional AI applications continue to demonstrate their transformative power. A great example is Google DeepMind’s AlphaFold, which has revolutionized protein structure prediction—a medical breakthrough unimaginable a decade ago.

In the packaging world, similar opportunities exist. AI can analyse data to make micro-decisions about structural design or material thickness, drastically impacting both costs and carbon footprints. When scaled to millions of units, these optimisations create significant benefits—less material usage often translates into lower costs and reduced environmental impact.

Manufacturers that adopt AI for such optimisations will gain a competitive edge while contributing to a more sustainable future by reducing waste in physical goods production. At Sourceful, we’re actively exploring these applications and would love to collaborate with others in the space.

‍