Authors:
Susan Brosnan, Daniel Goins, Patricia Hogan, Jessica Snead
Abstract:
While there are several solutions that aim to verify the groceries picked in the store for Click-and-Collect, there are not good solutions for Verifying the content of the bags that are delivered to a customer. This solution uses cameras and Artificial Intelligence to ensure the actual groceries moved from the storage lockers into the customer's vehicle contain the ordered items.
Background:
There are many solutions for assisting pickers in the store with purchasing the correct merchandise, as well as solutions to tie a particular shopper to the bag(s) for that shopper (labeling, numbering, and/or bar-coding bags). But there currently is not a good solution for ensuring that the contents of the bag match what the shopper ordered. Very often, picked groceries include several Shelf-Stable Items, as well as one or more Cold/Frozen and Hot bags. Mistakes are not infrequently made when combining those bags into the Shelf-stable, such that Shopper A ends up with Kielbasa and Sliced Roast Beef, meant to be combined from Cold Lockers into Shopper B's bag. With the benefit of Click-and-collect being that the shopper never has to exit the vehicle, it is typically not until Shopper A arrives home that the mistake is discovered.
Description:
The solution is to use mesh or see-through bags (perhaps reusable totes) to store the items in the lockers and then use cameras/technology in the Storage Locker area to identify types of items that seem unlikely to belong to the customer being served at that time.
Case where AI is not connected to the Store Current Transaction Engine:
In one embodiment for unusual purchase for shopper history:
-- Shopper A arrives to pick up items, which is known by the Pick-up Application (either through the App, Car Detection, manual Shopper assistant Entry, etc.)
-- Shopper assistant retrieves the bags for the self-stable items for shopper A
-- Shopper assistant 1 and Shopper assistant 2 pull items from the Cold and Hot Storage Lockers
-- Shopper assistant 2 is confused and pulls a cold bag for Shopper B from the shared Cold Locker
-- AI recognizes sliced deli meat in that cold bag, which it knows is unusual for Shopper A from Shopper A's past purchases; it flags the item for double-check
-- Shopper assistants are cued to double check the suspected bags against the order, prior to taking the bags to the car, so that any mistakes can be corrected
-- The sliced meat bag is returned to the Cold Storage Locker for the later arrival of Shopper B
In another embodiment comparing to Shopper Preferences/Allergies:
-- Shopper A arrives to pick up items, which is known by the Pick-up Application (either through the App, Car Detection, manual Shopper assistant Entry, etc.)
-- Shopper assistant 1 retrieves the bags for the self-stable items
-- Shopper assistant 1 pulls a bag with a Mrs. Smith’s Pecan Pie from the Freezer Storage Locker
-- AI recognizes the pecan wording/picture and verifies this common allergen against the Loyalty Setting for Shopper A and sees that Shopper A has a nut allergy; this is flagged for double-check
-- Alternatively, the AI could be loaded ahead of time with allergies of concern for the arrived customer.
-- The Pecan Pie was in the order, but the Shopper assistant double-checks with Shopper A when delivering the bags to the trunk, since this is an allergy item.
-- Shopper A made an ordering mistake, meaning to order Pumpkin Pie, and the Shopper assistant can make that adjustment for Shopper A on site.
Case where AI is connected to the Store Transaction Engine:
In yet another embodiment with more connectivity between Store Systems, where an item is compared to Shopper Preferences/Allergies:
-- Shopper A arrives to pick up items, which is known by the Pick-up Application (either through the App, Car Detection, manual Shopper assistant Entry, etc.)
-- Shopper assistant 1 retrieves the bags for the self-stable items
-- Shopper assistant 1 pulls a bag with a Mrs. Smith’s Pecan Pie from the Freezer Storage Locker
-- AI recognizes the pecan wording/picture and verifies this common allergen against the Loyalty Setting for Shopper A and sees that Shopper A has a nut allergy; this is flagged for double-check
-- The AI system requests Shopper A's purchase list for the Store Transaction Log and compares for any "pecan pie" entries; Note that the AI/cameras do not have to specifically identify the item - this is a cross-check on the items being pulled from the Locker, so finding something similar is sufficient (i.e., Mrs. Smith’s and Sara Lee do not have to be distinguished)
-- Shopper A does have a Pecan Pie item in the order (Shopper A is having company), so the AI does not flag that item
In a fourth embodiment using the additional connectivity between Store Systems, where Items are compared to the Current Transaction from Hot/Cold/Frozen Storage Lockers:
-- Shopper A arrives to pick up items, which is known by the Pick-up Application (either through the App, Car Detection, manual Shopper assistant Entry, etc.)
-- Shopper assistant 1 retrieves the bags for the self-stable items
-- Shopper assistant 1 pulls a bag with kielbasa from the Cold Storage Locker (system checks things removed from Cold, Frozen, Hot storage, since they are more likely to be confused when adding to the shelf-stable bags)
-- AI recognizes a sausage product and compares to the Shopper A's current purchase list from the Store Transaction Log; it would not check everything -- just the combining bags to verify it contains the right content
-- AI/TLOG comparison shows that no sausage was purchased by this customer in the current order
-- Shopper assistant is flagged to remove the kielbasa from Shopper A's bag
-- AI then determines the likely bag from the current orders that have been picked.
-- If the found Order is still on site, the Shopper assistant is notified to move the kielbasa to that order; If the order has already been distributed, the customer is either auto-notified by the System or Management is auto-notified to contact the real Kielbasa Shopper.
Supporting Art:
There are warehouse applications that check the correct order items are added to the correct customer order carton/box/bag. The system can see the order. This is a little different to our idea. Our idea is after the picking but at the part where the shelf-stable, hot, or cold items are being combined to be taken to the shopper’s car.
TGCS Reference 2945
