The workforce is changing: How automation is reshaping the potato industry – and the people who keep it running

Labour shortages are accelerating investment in optical sorting, autonomous equipment and data-driven processing. The technology can reduce repetitive work and improve consistency, but its success will still depend on experienced people, new technical skills and a fair transition for the workers affected.

By Lukie Pieterse, Editor and Publisher, Potato News Today

Across the international potato industry, one production constraint increasingly follows the crop from the field to the factory: finding enough dependable people to do physically demanding, repetitive and highly seasonal work.

Growers struggle to recruit tractor and harvester operators. Packing facilities need people capable of inspecting potatoes accurately for long shifts. Storage managers must monitor increasingly complex buildings, while processors require employees who can maintain continuous production under strict food-safety and quality standards.

The causes differ between regions.

In North America and Western Europe, ageing rural populations, competition from other industries, immigration policy and rising wages are important factors. In parts of Asia, rapid industrialisation is drawing workers away from farms. In Africa and Latin America, mechanisation may be limited by capital, farm size, technical support and access to finance—even where manual work is scarce or arduous.

Automation is emerging as a common response. But describing it simply as replacing labour misses the more important transformation.

The potato industry is not becoming people-free. It is changing what it requires people to do.

Potatoes are already highly mechanised—but not fully automated

Commercial potato production has relied on machinery for decades.

Planters meter seed into rows. Sprayers, cultivators and irrigation systems manage the crop. Windrowers and harvesters lift potatoes and separate soil, vines and stones. Conveyors move tubers into trucks, stores, packing lines and processing plants.

Yet many operations still depend heavily on human judgement.

Harvester operators adjust digging depth and separation intensity as soil and crop conditions change. Workers remove clods, stones, damaged tubers and foreign material from inspection tables. Graders decide whether potatoes meet changing customer specifications. Storage managers interpret temperature, humidity, carbon dioxide and crop behaviour.

Automation is now moving into these judgement-intensive tasks.

Machine vision, artificial intelligence, connected sensors and automatic controls are enabling equipment not merely to move potatoes, but to assess them and respond.

Optical sorting is leading the transition

Sorting is among the clearest examples of automation driven by labour pressure.

Manual inspection is demanding. Workers must make rapid decisions while thousands of potatoes pass in front of them. Accuracy can decline with fatigue, changing light, difficult batches or inconsistent interpretation of specifications.

Optical sorters use cameras, lasers, near-infrared or other sensors to evaluate colour, shape, size, surface condition and foreign material. Artificial-intelligence models can be trained to recognise complicated defect patterns and classify individual potatoes according to predetermined standards.

The advantages can extend beyond reducing headcount:

More consistent grading across shifts.
Stable throughput when raw-material quality changes.
Better separation of saleable product from genuine defects.
Reduced exposure to repetitive work.
Detailed information about each batch.
Faster response to changing customer specifications.
Improved traceability and production planning.

Flikweert Vision reported in August 2026 that users of its AI-powered QualityGrader had achieved processing-capacity increases of 10–20% in some operations. The company emphasised that customers often invest initially to reduce manual inspection costs but later place greater value on consistent capacity, lower unnecessary rejection and improved batch information.

Those are manufacturer-reported results rather than universal guarantees. Actual performance depends on line design, crop condition, defect type, settings, maintenance and the quality of the training data.

Nevertheless, the pattern is significant. The business case for automation is shifting from “replace several workers” towards “make the entire operation more predictable.”

A machine can be consistently wrong

Automated inspection is not automatically accurate.

A sorter applies the criteria it has been given. If those criteria are poorly configured, it may reject good potatoes with impressive consistency.

Different markets tolerate different characteristics. A superficial skin blemish may matter greatly in a premium fresh pack but be irrelevant for a peeled product. A defect that appears minor externally may indicate an internal quality problem. Varieties also differ naturally in colour, shape and skin characteristics.

Operators must therefore understand the product as well as the machine.

They need to validate grading decisions, examine rejected material, adjust recipes when varieties or crop conditions change and recognise when cameras, lighting or sensors require cleaning or calibration.

The worker has not disappeared. The role has moved from watching every potato to supervising the decision system.

Harvest automation faces a more difficult environment

A packing plant offers relatively controlled lighting, speed and product presentation. A potato field does not.

Soil texture, moisture, stones, slopes, vines, tuber depth and weather can change within one field. Harvest machinery must remove soil and debris while protecting tubers from cuts, bruising and drops.

Modern harvesters already use automatic depth control, load monitoring, camera systems and adjustable separation equipment. GPS and guidance technologies improve row alignment and reduce operator fatigue. Electronic controls can coordinate harvesters, tractors and transport vehicles more efficiently.

Full autonomy is more difficult.

A driverless machine must safely detect people, vehicles, field boundaries and unexpected obstacles. It must continue operating in dust, mud, darkness and vibration. It also needs to respond when the crop suddenly behaves differently on the machine.

Research continues into autonomous platforms and fleets of smaller harvesters, but large-scale potato harvesting remains a demanding test.

In the near term, the most practical development may be supervised autonomy: equipment controls routine driving or adjustment while a skilled person monitors performance and intervenes when conditions change.

That can reduce fatigue and allow one operator to oversee more than one function without pretending the field no longer needs human attention.

Automation is spreading through storage

Potato storage once depended heavily on periodic manual readings and an experienced manager’s senses.

Modern stores may contain temperature cables, humidity and carbon-dioxide sensors, variable-speed fans, refrigeration controls, weather stations and remote alarms. Newer analytical systems can compare conditions across zones and identify unusual trends.

Automation can regulate airflow and temperature more precisely than continuous manual switching. Predictive tools may help operators prioritise inspections and recognise deterioration earlier.

But storage control remains biologically complicated.

A system that responds aggressively to one reading may increase dehydration, energy use or condensation elsewhere. The correct response depends on variety, market, crop condition, outdoor air, dormancy and the stage of storage.

The storage manager of the future will need fewer hours beside switches and more knowledge of plant physiology, airflow, data interpretation and control logic.

Processing plants are moving towards fewer—but more technical—operators

Automation in processing extends from receiving and peeling to cutting, frying, freezing, seasoning, packaging and pallet handling.

Highly integrated lines can maintain enormous throughput with relatively few operators. Predictive maintenance systems monitor vibration, temperature and equipment performance to identify developing mechanical problems. Digital dashboards bring information from several production stages into one control environment.

This reduces dependence on manual handling and repetitive observation. It also raises the consequences of technical failure.

When a highly automated line stops, production can be lost rapidly. A plant with few employees cannot afford to discover that none understands the interaction between software, sensors, mechanical systems and food-processing conditions.

Technical support therefore becomes part of production capacity.

The industry increasingly needs:

Electrical and mechanical technicians.
Automation and control specialists.
Machine-vision operators.
Data analysts.
Food-safety professionals.
Process engineers.
Refrigeration specialists.
Employees capable of working across mechanical and digital systems.

Interpack 2026 identified both smart manufacturing and future skills as central themes for the processing and packaging industries. That combination is appropriate. Equipment investment without workforce development creates expensive vulnerability.

The experienced worker carries knowledge that was never written down

Automation programmes often begin by analysing formal tasks: remove defective potatoes, operate a grader, adjust a harvester or monitor a control panel.

What they can overlook is tacit knowledge.

An experienced worker may recognise a developing bearing problem by sound. A harvester operator feels when separation has become too aggressive. A grader notices that a defect pattern is associated with a particular field or storage lot. A storage manager understands that a small temperature change matters because of what happened during harvest.

This knowledge may not appear in an operating manual, yet it protects product quality and prevents downtime.

Before automating a task, businesses should document what experienced employees actually do—not merely what their job description says they do.

Pairing senior operators with younger technicians can be particularly valuable. One brings crop, equipment and situational knowledge; the other may bring confidence with software, sensors and data.

The strongest operation does not choose between them. It deliberately connects them.

Small and medium operations face a different calculation

Large processors can spread automation costs across high volumes. A smaller packer or grower may struggle to justify the same investment.

The relevant question is not whether the technology is impressive, but whether it solves a clearly defined bottleneck.

A business should calculate:

Current labour and overtime costs.
Recruitment and staff-retention difficulties.
Throughput lost during difficult batches.
Product rejected unnecessarily.
Downtime and maintenance expenditure.
Training and technical-support costs.
Financing and depreciation.
Expected equipment utilisation.
Availability of replacement parts and local service.

A sorter that operates only a few weeks annually may have a weak payback unless it also provides custom services to neighbouring farms.

Cooperative ownership, machinery-sharing, mobile equipment and contract services may allow smaller producers to benefit without carrying the entire capital cost.

This is especially important in emerging regions. Mechanisation designed for enormous European or North American farms may be financially and physically unsuitable for smaller fields.

FAO’s approach to sustainable agricultural mechanisation emphasises appropriate scale, local service capacity and the reduction of drudgery. Simply importing sophisticated machinery does not create a durable mechanisation system.

Automation can widen inequality

The move towards automation carries a risk of creating two potato industries.

Well-capitalised businesses can invest in advanced equipment, lower unit costs and satisfy strict customer specifications. Smaller operators may fall further behind, particularly when buyers begin treating automated traceability and grading as minimum requirements.

Access is also uneven within rural communities. FAO has reported persistent gender gaps in machinery ownership and access, with women-managed farms using less equipment than male-managed or jointly managed farms in most countries examined.

Training programmes must consequently be accessible to existing workers, young people, women and smaller businesses—not confined to already advantaged operations.

Otherwise, technology intended to solve labour shortages may intensify economic exclusion.

A responsible transition must include affected workers

It would be dishonest to claim that automation never removes jobs.

Repetitive grading, inspection and handling positions can be reduced when machines assume those tasks. The impact is real for the individuals whose income depends on them.

Businesses should not discuss that change only in terms of efficiency.

A responsible transition includes early communication, retraining, opportunities to move into machine operation or quality assurance and realistic recognition that not every displaced worker can become a programmer.

The International Labour Organization warns that digitalisation can produce job displacement, skills mismatches and wider inequality unless adjustment policies and training accompany technological change.

For the potato industry, this is both an ethical and practical issue.

A business that treats experienced workers as disposable may lose the knowledge needed to make its new system succeed.

The future operator will manage exceptions

Automation performs best when conditions are repetitive and clearly defined. People remain strongest when something unexpected happens.

The future potato professional will increasingly manage exceptions:

A crop arrives with an unfamiliar defect.
A sorter’s rejection rate suddenly rises.
A storage zone stops responding normally.
A harvester encounters different soil conditions.
Two sensors provide conflicting information.
Equipment recommends an action that does not make biological sense.

This requires judgement, curiosity and confidence to challenge the system.

Training should therefore go beyond button-pushing. Workers need to understand why the equipment makes decisions, which information it uses, where its limitations lie and how to test whether the outcome is credible.

The industry still runs on human responsibility

Automation can make potato production safer, more consistent and less dependent on physically exhausting labour. It can improve recovery, reduce waste and allow operations to continue where workers are increasingly difficult to find.

But the promise should be stated honestly.

Machines do not accept responsibility for food safety, crop damage, rejected product or a failed storage season. People do.

The next generation of potato operations will probably employ fewer people in some repetitive roles, but it will depend more heavily on the competence of those who remain.

That makes workforce development as important as machinery investment.

The successful potato business of the future will not be the one with the fewest workers. It will be the one that combines appropriate automation with experienced, respected and properly trained people.