A fully automated warehouse should be planned as an end-to-end operating system, not a collection of machines. The right design starts with the work to be done: what arrives, how inventory is stored, how orders are built, and where goods leave the site. Only then can a business decide which processes to automate and how equipment should connect. Mapping demand, space, product characteristics, exceptions, staffing and operating costs first helps expose bottlenecks before they become expensive parts of the layout. It also clarifies whether full automation is appropriate or whether a targeted system would meet the same operational need with less complexity.
There is no single equipment package that makes every warehouse fully automated. The term usually describes a facility where connected technologies handle much of the movement, storage, retrieval, sorting or picking of goods. The exact scope varies. A system might automate pallet storage and transport while people still pick individual items, or it might automate order picking but rely on people for receiving, replenishment and packing.
Even a highly automated site needs people to manage exceptions, maintain equipment, investigate inventory discrepancies and respond when a process cannot continue as planned. The useful planning question is therefore not “How do we remove every person?” but “Which tasks should be automated, and how will the operation work when an item, order or machine falls outside the normal process?”
Set boundaries before comparing suppliers. Identify the processes in scope, the service outcomes the system must support, and the work that will remain manual. This prevents a proposal from appearing complete because it includes sophisticated equipment while leaving a critical hand-off or exception process undefined.
Start with the physical and information flow from receipt to dispatch. Document how products arrive, are checked and identified, move into storage, become available for orders, and reach packing or shipping. Include returns, damaged stock, replenishment and other work that can compete for the same space or equipment. A process map should show both the normal path and the points where staff must make a decision.
This work helps establish whether automation will address a real constraint. If the process is inconsistent, stock data is unreliable or work arrives in large, uneven waves, automation may make those problems happen faster rather than remove them. Resolve what can be resolved first, then design the system around the remaining operational needs.
Different technologies solve different handling problems. A storage system does not automatically solve picking, and a mobile robot does not by itself organise order release or packing. The useful comparison is how each option fits the required flow, what it depends on, and what work remains around it.
| System type | Useful role | Key dependencies | Planning limitation |
|---|---|---|---|
| Automated storage and retrieval system (AS/RS) | Stores and retrieves loads or items within a designed storage structure. | Compatible load and product characteristics, defined access patterns, and suitable interfaces with inventory and order systems. | Storage capacity alone does not ensure goods can be retrieved, picked, packed and dispatched at the required pace. |
| Automated guided vehicles or autonomous mobile robots | Moves loads or containers between designated processes. | Clear travel routes, workable hand-off points, traffic rules and a process for handling blocked routes or unavailable destinations. | Congestion, shared aisles and poorly placed pickup or drop-off points can restrict system flow. |
| Conveyors and sortation equipment | Moves and directs goods along a repeatable path between work areas. | Suitable product or container handling, balanced upstream and downstream processes, and space for accumulation where needed. | A fixed route can be difficult to adapt when the operation changes; downstream queues can stop the line from clearing. |
| Goods-to-person workstations | Brings stored items or containers to a worker for order fulfilment. | Accurate inventory, replenishment, workstation capacity and a clear process for completed orders. | It can shift the constraint to replenishment, order consolidation, packing or dispatch if those steps are not sized to match. |
| Automated pallet handling | Transfers palletised loads between storage, staging or production-related areas. | Consistent load condition, compatible transfer points and sufficient space to stage incoming and outgoing pallets. | Damaged, unstable or non-standard loads may need an exception route and manual handling plan. |
These options can be combined, but each added interface creates a planning and integration requirement. Compare proposals against the same operating scenarios and ask suppliers to show how goods move through the complete process, including replenishment, order consolidation and exception handling. A machine’s stated rate is not the same as the rate the whole warehouse can sustain.
Capacity is a property of the connected operation. Receiving, putaway, storage, retrieval, picking, packing and dispatch need to work together. If one stage can process work faster than the next stage can accept it, queues build up and the nominal capacity of upstream equipment becomes irrelevant.
Review both the quantity of work and its pattern. A system must deal with the mix of products and orders, the timing of inbound and outbound work, replenishment activity and the space available for temporary buffers. Ask how work will be released to equipment, how priorities will be managed, and where work can wait safely without blocking travel paths or workstations.
Test more than a steady, average scenario. Consider shifts in order mix, inbound peaks, replenishment competing with picking, and a temporary reduction in the availability of a key process. The purpose is not to predict every possible event; it is to find out whether a reasonable disruption creates a manageable queue or stops the flow altogether. Where the design depends on a buffer, specify what it holds, how it is controlled and how work is cleared from it.
Normal movement is only part of the design. An item may arrive without usable identification, a load may not meet handling requirements, or an order may need review. Equipment can also require maintenance or become unavailable. If these cases have no defined route, they tend to occupy operating space, interrupt standard work or create informal workarounds.
Plan for a degraded operating mode: what can continue, what must pause, and which people are authorised and trained to respond? A manual fallback does not need to reproduce the automated process at full capacity. It does need to protect people, preserve stock accuracy and give operators a controlled way to recover the operation.
Automation depends on accurate decisions as well as physical movement. Warehouse and order systems need to communicate with equipment controls so that work is released, assigned, tracked and confirmed correctly. System responsibilities differ by design: warehouse management software may manage inventory and orders, while execution or equipment-control layers coordinate tasks and machinery. Suppliers may combine or divide these functions, so confirm the actual responsibilities rather than relying on product labels.
During design, map the key events and data exchanged at each hand-off. For example, determine when an order becomes available to the automation, how a retrieval request is confirmed, and how a completed movement updates inventory. Define how errors are reported, who owns each interface and what happens if a message is delayed or rejected.
Test with realistic product, order and exception data before live operation. Include the people who use the systems day to day, not only technical teams. A technically connected system can still create operational delays if screens, alerts or recovery instructions do not help staff make the next decision.
Automation changes where people work and how they interact with moving equipment. Review pedestrian routes, vehicle movement, loading and unloading, access for maintenance, emergency arrangements and the separation of people from automated travel paths. The detailed controls must suit the selected equipment and the site; obtain appropriate safety and engineering input during design rather than treating safety as a final inspection.
Consider the building as well as the equipment. Check that the proposed layout can accommodate storage structures, travel paths, workstations, staging, utilities and maintenance access. Existing buildings may impose constraints that affect equipment choice or usable space. For a New Zealand facility, confirm relevant local building, fire, workplace safety and electrical requirements with qualified advisers before committing to a design.
Plan workforce changes alongside the technical project. Staff may need training for monitoring, exception handling, stock control, maintenance coordination and safe recovery. Define new responsibilities and escalation paths before commissioning so operators know who acts when the system behaves outside the normal flow.
Compare the cost of automation with the operational problem it is intended to solve. Include equipment and installation, software and integration, facility changes, training, maintenance, support, replacement parts and the cost of keeping the operation running during implementation. Also account for ongoing staffing and manual work that will remain. A lower purchase price does not necessarily mean a lower operating cost if a design is difficult to maintain or depends on costly changes to surrounding processes.
Use a consistent comparison between the proposed system and realistic alternatives, such as improving the current layout, adding targeted equipment or changing work allocation. Make assumptions visible: expected operating pattern, volume, product mix, service needs and the period over which the investment is assessed. Test how the case changes if demand grows more slowly than expected, the order mix changes, or operating costs differ from the plan.
Ask what is included in the supplier’s scope and what is not. Clarify responsibilities for integration, commissioning, staff training, service support and future modifications. The business case should show the operational benefit, the conditions needed to achieve it and the main cost or performance risks—not just a projected saving.
A structured implementation gives the business a chance to resolve design gaps before they affect the whole operation. The exact sequence depends on the project, but these steps provide a useful basis for planning:
Set acceptance criteria before commissioning. They should cover the operation as a whole, not merely whether each machine can run. Include successful order and inventory transactions, exception handling, recovery procedures and the hand-offs between equipment and people. Agree how problems will be recorded, resolved and retested.
No. Automation can reduce or change routine handling work, but people are still needed for supervision, maintenance, exceptions, inventory control and operational decisions. Plan roles and training around the tasks the system will not perform reliably or safely on its own.
Not necessarily. A phased or targeted project may suit an operation where demand, processes or data still need validation. The right approach depends on how closely the processes connect, what can be introduced without disrupting service, and whether the business can support parallel operations during a transition.
Prepare process maps, inventory and order data, product characteristics, operating constraints, service requirements and the known sources of delay or rework. Explain the exceptions and manual activities as well as the standard flow. Suppliers can then respond to a defined operation rather than make assumptions about the work.
Model the full route from inbound handling to dispatch, including replenishment, packing, staging and exceptions. Test different order mixes and periods of reduced equipment availability, and ensure each process has a clear downstream hand-off. Do not judge the design by one machine’s capacity in isolation.
Include installation and integration, facility changes, software, training, maintenance, support and the manual work that will remain. Compare the proposal with practical alternatives and make assumptions about demand, product mix and operating conditions explicit. Check that the projected benefits depend on changes the business can actually deliver.
A fully automated warehouse is most effective when storage, movement, software, people and facility design support one defined flow. Map the operation, identify its constraints, compare equipment by the job it must do, and test exceptions before committing to a layout. The next step is to build a clear operating brief and use it to evaluate complete system designs—not machines in isolation.
