Ocado Swarm Robotics & The Next Frontier in Warehouse Automation

This is a snippet of the transcript, sign up to read more.

On the human labor side, is it the pick and pack process that's one of the final obstacles? When we look at more line manufacturing and autonomy, do you think there might be more "intelligence" involved? On the retail side, do you see further enhancements in efficiency being done by process on high? Because every single step relies on the previous one. Is that final pick and pack the last sticking point in terms of achieving a fully automated system?

Yes, pick and pack is definitely the last main source of labor. There are a few other sources. In a facility like Ocado, you've got the decanters, who take items from pallets and put them into the storage system. That's a manual process and accounts for about 20% of labor on site. It's not as significant as pick and pack, which accounts for about 70% of labor in those facilities. The pick and pack step is the biggest bulk at 70%.

This is a snippet of the transcript, sign up to read more.

During your time at Ocado, I'm not sure how much you can share, but I imagine a major part of your work was to reduce costs. You likely collaborated with the financial team to review the P&L and track progress. Are there any data points you can provide on how much cost was reduced through swarm or other projects? How did you approach this with your finance team, and what were some of the major goals or milestones you aimed to achieve?

Regarding cost reductions, I can't provide specific numbers, but I can discuss throughput rates. In a conveyor-based system, you might see 150 to 200 units per hour per person. This measure reflects site productivity and labor efficiency by dividing total units produced by the total staff, including managers. For conveyor-based sites, you'd get around 150 to 200 units, while for our sites, it's 550 to 600, indicating higher labor efficiency.

This is a snippet of the transcript, sign up to read more.

When considering the modern warehouse, which human jobs do you think are the low-hanging fruit for automation with robots? Conversely, are there any jobs within a warehouse that you believe will always require human operation?

Picking is also starting to be replaced by robotics, but it's still in its infancy. For example, the project at Ocado is live and picking production items, but it's only handling a small percentage of the total picks in the facility. This is partly due to capital expenditure and the challenges of retrofitting, as well as the current capabilities of robots. They need to handle a wide variety of items, from a floppy bag of salad to a postage stamp to a crate of beer, and even a sack of potatoes where the weight shifts as you move it. The challenges include dealing with spillages and damages.

This is a snippet of the transcript, sign up to read more.