Lesson 1: get your item data right before you automate
The most common problem is not the pick software, but the data it depends on. Pick locations that are incorrect, article numbers that have been duplicated over years of growth, dimensions that were never filled in. A scanner or voice system needs that data to guide someone through the warehouse. If pick location A12-03 is in the system but the goods have been sitting at B07-11 for months, automation becomes an error-amplification machine. The sequence is therefore always: clean up location management and item master data, set up logical routes in the system, and only then switch on automation. That clean-up takes time, sometimes weeks. Anyone who skips this step and starts implementing straight away pays double: first in incorrectly picked orders, then in the loss of staff confidence in the new system. Automating a disorganized pick process simply makes that process disorganized at speed.
Lesson 2: scanning, voice, or WMS integration, what fits and when?
There is no universally best approach to picking. The three common options each have their own logic. Scanning on paper or with a standalone scan app has the lowest barrier to entry: low investment, quick to roll out, but limited feedback to the system. Suitable if you process relatively few order lines per day and want to take a step toward digital without a major project. Voice-picking adds hands-free working and helps with higher pick volumes or complex warehouse environments, but requires a solid data stream from the source system. Full WMS integration, where pick orders are created in real time, prioritized, and booked back, offers the most control but also demands the most from your master data and the systems around it. The question is not what is most advanced, but what fits your order volume, your product range, and the state of your current software. We see wholesale companies introduce voice-picking while their ERP does not systematically track pick locations. The result: a lot of confirmations for locations that are already empty.
Lesson 3: the link to inventory management determines whether time savings become visible
Suppose your pick process is well set up and staff are working with a scanner or voice system. The next question is: is inventory also updated in real time? And are orders coming in through the customer portal, by phone, or via EDI immediately visible in the pick queue? This is where time savings on the warehouse floor evaporate in the departments around it. If a pick order is created manually after a phone order, or if inventory mutations are only processed at the end of the day, pick time may have been halved but total lead time has barely changed. The link between order processing and inventory management is not a technical detail, it is the prerequisite for visible results. In practice, this means that when automating the pick process, you also need to address the upstream processes: how do orders enter the system, how quickly is inventory updated after receipt, and who manages the exceptions? Only when those chains are closed does faster picking translate into shorter delivery times.
When does pick process automation go wrong quickly?
The most dangerous situation is a project that starts with the technology too soon. A system is selected, a scanner solution purchased, or a WMS module activated, and only then does it become clear that the underlying data is incorrect, the connection to the ERP is half-baked, or staff have not been brought along in the new way of working. Warehouse automation requires buy-in on the floor. A warehouse worker who knew his own pick sequence and now has a scanner telling him to go to the wrong location will have restored his own route within a day and set the scanner aside. Technical rollout without operational preparation does not work. The sequence that does work: get the data right, design the process together with the people who carry it out, start small with a representative part of the product range, and then scale up.
What does a well-configured pick process deliver?
When data, method, and integration are in order, the gains are tangible: fewer incorrectly picked orders, shorter routes through the warehouse, better control over inventory levels, and less correction work afterwards. For wholesale companies with a broad product range and multiple order streams, that difference is noticeable in daily operations. But it does not start with the technology. It starts with the question: is our foundation strong enough to build on? If not, the first step is not automation but clean-up. Bonsai helps wholesale companies map out that foundation and build systems that fit the way the operation actually works, whether that is an AI Worker processing order streams or a custom-built WMS that connects to your existing ERP.