AI in packaging: where will it have the biggest impact?

Artificial intelligence (AI) is moving into practical use across the packaging industry, from design and manufacturing to procurement, supply chains and recycling. But its biggest impact may not come from its most visible application: generating packaging concepts.

The greater opportunity lies in using AI to optimise the interconnected decisions that determine how packaging performs – from material selection and package geometry to manufacturing, logistics, cost and environmental performance.

The result could be a shift from designing an individual package to optimising a wider packaging system.

That change is already under way. A 2025 McKinsey survey of 110 senior packaging leaders found that 82% said their function had launched, was developing or was considering generative AI solutions.

The functions with the highest reported use of machine learning or generative AI were commercial excellence, procurement, and supply chain and logistics.

The finding is significant because it shows that AI in packaging is extending beyond design into the commercial and operational decisions that determine how packaging businesses compete.

How AI is changing packaging design

Packaging design is one of AI’s most visible applications.

Generative AI can rapidly produce and visualise different design concepts, while other AI tools can analyse technical requirements and identify combinations of materials, dimensions and structural features capable of meeting specified performance targets.

McKinsey has identified applications including rapid idea generation, visualisation, consumer testing and the optimisation of parameters such as strength, materials and manufacturability.

Speed, however, may not be the technology’s most important contribution.

Packaging development requires designers and engineers to balance competing requirements. A package must protect its contents, run effectively on production equipment, appeal to consumers and fit efficiently into storage and transport systems. Increasingly, it must also address recyclability, material reduction and other sustainability requirements.

AI can analyse more of these variables simultaneously.

A design system could compare package geometries and material structures against targets for weight, strength, cost, production efficiency and logistics performance. Instead of manually evaluating a relatively small number of alternatives, engineers could use AI to explore a much larger design space before applying their technical judgement to the strongest options.

That changes the role of AI from simply generating designs to helping optimise packaging systems.

AI and sustainable packaging

Sustainable packaging could be one of the biggest beneficiaries of this capability because environmental decisions rarely involve a single variable.

Packaging companies face pressure to reduce material consumption, improve recyclability and lower environmental impacts without compromising product protection or commercial performance.

Assessments become more complicated when companies operate across markets with different regulations, collection systems and recycling infrastructure.

A 2025 systematic review of 48 studies on AI-driven green packaging found that applications were concentrated in areas including process optimisation, smart packaging monitoring, computer vision and waste reduction.

It also identified material innovation and circular-economy integration as important areas for future development, while highlighting cost, technical complexity and regulatory uncertainty as barriers to wider adoption.

AI can help companies compare packaging alternatives against multiple environmental and technical objectives rather than optimising a single metric.

Lightweighting, for example, does not automatically make a package more sustainable. Reducing material can be counterproductive if it compromises product protection, increases product or food waste, reduces recyclability or performs poorly in existing recycling systems.

AI can analyse these trade-offs, but it cannot determine the most sustainable solution without reliable data and clearly defined objectives. Results will depend on factors including the environmental metrics selected, system boundaries, assumptions about end-of-life treatment and the quality of the underlying data.

Its value is therefore less about producing a definitive sustainability answer and more about enabling packaging specialists to evaluate complex choices more quickly and systematically.

AI in packaging manufacturing and quality control

Manufacturing could be one of the areas where AI’s financial impact is easiest to measure.

Packaging production generates large volumes of operational data. Equipment can record information about speed, temperature, pressure, energy consumption, material behaviour, production output and quality. Machine-learning systems can analyse these data to identify patterns that conventional monitoring may miss.

Predictive maintenance is one established application. Instead of waiting for machinery to fail or relying solely on fixed maintenance schedules, AI systems can identify patterns associated with equipment problems and help operators intervene earlier.

Computer vision provides another important application. AI-enabled inspection systems can analyse packaging and printed materials for defects during production. McKinsey has highlighted applications including image and video processing for issue identification and AI-enhanced visual inspection of waste in paper and cardboard production.

The commercial case is straightforward.

Earlier defect detection can reduce material waste, rework and the risk of defective products reaching customers. Better production planning can improve asset utilisation, while predictive maintenance can reduce unplanned downtime.

For packaging manufacturers operating at high volumes and on tight margins, relatively small improvements in yield, uptime, energy use or scrap rates can translate into significant financial gains.

Crucially, these improvements can be measured against established operational metrics.

AI in packaging supply chains

Packaging decisions do not stop at the production line.

Material availability and prices, customer demand, manufacturing capacity, inventory and transport requirements can change quickly. AI can analyse historical orders alongside other internal and external data to support demand forecasting and production planning.

It can also support warehouse design, shipment optimisation and route planning. These are among the applications identified by packaging executives in McKinsey’s research.

The greater opportunity is to connect these decisions with packaging design itself.

A seemingly small change to a package’s dimensions can determine how many units fit into a case, how cases are arranged on a pallet, how efficiently pallets fill a vehicle and how much warehouse space a product requires.

Optimising the package without considering those consequences can therefore create costs elsewhere in the system.

AI can help companies model these relationships, assessing how changes in packaging design affect manufacturing, inventory, warehousing and transport.

Rather than treating the package as an isolated product, companies can increasingly evaluate it in the context of the entire supply chain.

AI could change packaging businesses beyond the factory

AI’s impact on packaging companies will not be limited to physical production.

McKinsey’s 2025 survey found that 56% of respondents reported machine learning or generative AI use in commercial excellence, compared with 43% in procurement and 37% in supply chain and logistics.

Potential applications extend across sales, pricing, customer analysis, procurement and market intelligence.

A converter or materials supplier, for example, can use AI to analyse customer and market information, identify sales opportunities, prepare proposals or accelerate routine commercial work.

Procurement teams can process supplier information and support purchasing decisions, while sales teams can analyse large volumes of customer and market data more quickly.

These applications may receive less attention than AI-generated packaging designs, but they can have a direct effect on revenue, margins and operating costs.

They also illustrate a broader point: AI in packaging is not simply a technology for designing or manufacturing packages. It is increasingly becoming a tool for running packaging businesses.

Smart packaging and AI

AI could also increase the value of information generated by smart and connected packaging.

Sensors, digital identifiers and other connected-packaging technologies can provide information about products and the environments through which they move. AI can analyse that information and identify patterns that may otherwise be difficult to interpret.

Food packaging is an important area of research, with studies examining AI applications for freshness monitoring, quality prediction, food safety, traceability and supply-chain optimisation.

Combining smart packaging with AI could help businesses make better decisions about storage, distribution and product quality. In food supply chains, better information about product condition could also contribute to efforts to reduce waste.

Many applications, however, remain at different stages of technological and commercial development. Wider adoption will depend on factors including sensor costs, infrastructure, data integration and validation under real operating conditions.

The immediate opportunity may therefore be less about making every package intelligent and more about making the information generated by packaging useful.

AI in recycling and the circular economy

The packaging industry’s transition towards a more circular economy creates another potential role for AI.

Computer vision and machine learning can help identify materials and contaminants in waste streams, supporting automated sorting. AI can also help analyse packaging formats and materials against recycling requirements.

These capabilities could become increasingly useful as packaging companies face more complex requirements covering recyclability, recycled content and producer responsibility.

The Sustainable Packaging Coalition’s 2026 trends report highlights continuing challenges around data, consistent definitions and recyclability assessments. It also points to greater data sharing across the value chain as a way to improve understanding of packaging recyclability.

AI cannot resolve those structural problems by itself. Recycling performance depends on packaging design, collection systems, sorting infrastructure, markets for recycled material and consumer behaviour.

Better analysis of design, material and waste-stream data can nevertheless help packaging companies make more informed decisions about how their products are likely to perform within those systems.

Why optimisation could be AI’s biggest opportunity

Taken together, these applications suggest that AI’s most consequential contribution to packaging may not come from any single technology.

Generative AI can accelerate design development. Machine learning can improve forecasting and production. Computer vision can identify defects and support material sorting. Predictive analytics can support maintenance and quality control.

The larger opportunity emerges when these capabilities are connected.

A future packaging development process could start with a product’s requirements and assess multiple package formats, materials and structures. Those options could then be evaluated against manufacturing constraints, production cost, transport efficiency, product protection and environmental criteria before the strongest alternatives are presented to packaging engineers and designers.

The same principle could apply after a package enters production. Manufacturing data could inform design decisions. Sales and demand data could influence production planning.

Logistics data could identify opportunities to change package dimensions. Recycling data could feed back into future material and structural choices.

This is where AI could create a step change: connecting decisions traditionally made by different teams using different data.

The opportunity is not simply to make individual decisions faster. It is to understand how one packaging decision affects the rest of the system.

Human expertise would remain essential. AI can identify patterns, analyse data and generate options, but packaging decisions require technical judgement and knowledge of materials, manufacturing processes, regulations, product safety, markets and consumers.

The technology is therefore more likely to augment packaging professionals than replace them.

Data and governance will determine the value of AI

AI’s potential does not mean every packaging company will benefit equally.

The quality of AI outputs depends heavily on the information available to the system. Packaging businesses need reliable data about products, customers, materials, equipment and processes, as well as the digital infrastructure required to connect and analyse it.

The nature of the challenge is also changing. McKinsey’s 2025 survey found that intellectual property and privacy concerns, along with limited understanding of which use cases can create value, had become the leading barriers to gen-AI adoption among respondents.

Only 7% identified limited access to data and modern data stacks as their primary barrier, down from 21% in 2024.

Human oversight is equally important. AI-generated designs, forecasts and recommendations can contain errors and should not automatically be treated as correct.

That is particularly important where decisions involve product safety, food-contact materials, regulatory compliance or environmental claims.

The strongest strategy for packaging companies is therefore unlikely to be deploying AI wherever technically possible. The better approach is to identify specific business problems where the technology can deliver measurable improvements in cost, productivity, quality, resource efficiency or customer value.

From faster design to smarter packaging decisions

AI is extending across packaging design, manufacturing, quality control, logistics, procurement and recycling. Its most important long-term effect, however, could be the connections it creates between them.

Packaging companies must simultaneously balance cost, material performance, manufacturability, product protection, logistics and environmental considerations. AI offers the possibility of analysing more of those relationships at once and testing many more potential solutions.

The technology is already producing practical applications in packaging businesses, but most companies are still working towards meaningful, scaled impact. McKinsey’s latest research shows that adoption has moved beyond experimentation, while much of the industry’s AI activity remains at an early stage.

For an industry under pressure to control costs, improve sustainability, increase efficiency and respond faster to customers, the ability to connect data and decisions could prove more consequential than any individual AI application.

AI’s biggest impact on packaging may ultimately be its ability to optimise the package as part of a wider system – rather than simply making the package itself faster or cheaper to design.