Late in my career I worked with technology delivery and process redesign within the corporation I worked for, I really enjoyed that work. Everyone knows that technology is a business disruptor. Here are my thoughts.
The current systems are designed around the retail shopper. The question is, "will consumers accept technology advances and automated curbside or home delivery?" Redesigning the current process is a choice of evolution or revolution. Existing corporation are taking the "toe dip" redesign approach. They are mixing automated delivery with the current retail markets to see how that works. It's inefficient, confusing and alienating. But the automated delivery segment is small right now. And these inefficiencies can be tolerated.
If I were starting a new grocery system, I would separate the online process from the conventional store design. Build separate central "warehouses" designed specifically for online delivery which use autopicking technology and automated deliveries. These systems would be designed to allow for easy update as technology improves. I would convert the current large grocery store "department" model to smaller local store fronts with larger produce and perishable selections. The central warehouse will deliver to the home for consumers who elect that convenience or to the smaller storefronts for customers who want to select their own perishables.
The consumer gets the best of all options, more selection for non-perishables since the "large" warehouse will hold all non-perishables, more selection for perishables, since the "local storefront" will have larger perishable selection, cost efficiency since the automated processes will not interfere with local shopping and the whole inventory can be converted to more "just in time" management, and the consumer can take advantage of the convenience of home delivery at less or no cost. Of course there are issues to worked out, like how will consumers "shop" for items they aren't ready to select or how to manage the perishable inventory. Both of which are issues within the current model.
Obviously redesigning the current model for large corporations will be expensive. But the potential in cost savings, customer approval and the lost opportunity cost will eventually see new designs happen organically. Might as well get ahead of the curve. Of course there are risks, but there are always risks in business.