RU2772835C2 - System for aggregation of metadata of events of consumer behavior in a store, data verification and their analysis using artificial intelligence to interpretate data and launch a related action - Google Patents

Abstract

FIELD: computer technology.

SUBSTANCE: invention relates to the field of computer technology for aggregating metadata of consumer behavior events in a store. The effect is achieved by a trained artificial intelligence-based analytical engine, which may have a data interpretation controller configured to intelligently interpret such aggregated metadata of consumer behavior events in the store and trigger actions accordingly, which are then sent electronically to the consumer's electronic devices. The data interpretation controller may have a data verification controller configured to optimize the data integrity of the consumer product metadata database representing the plurality of consumer products according to availability and other metadata.

EFFECT: invention improves the accuracy of in-store consumer behavior event metadata aggregation, verification, and analysis using artificial intelligence to interpret the data and trigger the associated action.

18 cl, 7 dwg

Description

Technical field

[1] The present invention generally relates to AI-based analytics engines, and more specifically, the present invention relates to a system for aggregating store behavior event metadata, verifying the data, and analyzing it using artificial intelligence to interpret the data and trigger the associated action. suitable for real-world store applications.

State of the art

[2] Analyzing the aggregation of consumer behavior data and triggering the event usually occurs online. For example, various providers track consumer behavior, including the consumer's interest in multiple websites, using tracking technologies, including the use of cookies, after which, once the consumer's interest has been observed, targeted advertising is served to consumers accordingly.

[3] However, such systems cannot easily be applied in the real world, for example, for use in analyzing consumer behavior in a store and modifying consumer behavior.

[4] Problems encountered in deploying such analytics engines in the real world include issues with the integrity of consumer product metadata in that products vary greatly in both availability and pricing among different stores.

[5] In addition, there are difficulties in obtaining metadata of consumer behavior from the real world.

[6] Further, online targeted advertising analytics engines of the prior art generally go as far as issuing targeted advertisements selected from an advertisement database according to a user's identified interest. However, it would be ideal to have a system that could do much more, including generating additional AI-assisted inferences useful to the consumer in the real world.

[7] The present invention is directed to providing a system and associated methodology that eliminates or substantially reduces at least some of the shortcomings of the prior art, or at least provides an alternative.

[8] It should be understood that if reference is made herein to any prior art information, such reference is not an admission that the information forms part of the general knowledge of the art in Australia or any other country.

Disclosure of the essence of the invention

[9] The following embodiments provide a system for aggregating metadata of consumer behavior events in a store, verifying the data, and analyzing it using artificial intelligence to interpret the data and trigger an associated action that overcomes or at least reduces the problems of prior art systems, or at least provides an alternative.

[10] As will be described in more detail below, the system collects in-store consumer behavior event metadata from a variety of consumer electronic devices and then uses a trained AI-based analytics engine to generate various AI-based inferences useful to those consumers, which can then modify consumer behavior.

[11] As will be described in more detail below, the trained AI analytics engine has a data interpretation controller capable of intelligently interpreting such aggregated store behavior event metadata of a consumer in a store and triggering actions accordingly, which are then sent electronically to the consumer's electronic devices. .

[12] In embodiments, the data interpretation controller comprises a data verification controller configured to optimize the data integrity of a consumer product metadata database representing a plurality of consumer products according to availability and other metadata (such as store locations, cost, and the like). In view of this, the data verification controller is configured to firstly build a relatively accurate consumer product metadata model in a manner that solves the problems of aggregating such data in the real world, after which, once such a relatively accurate data model is built, the system can make further findings of artificial intelligence.

[13] In this regard, the system may further have intelligent bidding capabilities, wherein the data interpretation controller comprises a bidding controller.

[14] The present system uses a shopping list graphical user interface (GUI) displayed by a display device on a plurality of consumer electronic devices, wherein the shopping list GUI is configured to display a plurality of consumer-specific consumer products and a list format, and each displayed the consumer product contains an associated purchase confirmation input (check box or the like) indicating its purchase.

[15] In this regard, through the use of such consumer-specific shopping lists and, in particular, consumer products listed therein, the product offering controller is able to intelligently suggest products for consideration by the consumer.

[16] In embodiments, the product suggestion controller is configured to suggest specific products among product categories in an intelligent manner such that, for example, may be inferred from consumer-specific parameters or obtained from machine learning metadata of the consumer's in-store behavior event received from other consumers. .

[17] Also, in embodiments, the product suggestion controller is configured to suggest probable unrelated products, but which may be relevant to the user, as identified by analysis performed by a machine learning algorithm.

[18] In addition, in embodiments, the trained AI-based analytics engine is configured to generate intelligent information notifications regarding the consumer product that may be useful to consumers. Such consumer product informational notices can also be obtained from a machine learning algorithm.

[19] In accordance with one aspect, a system is provided for aggregating consumer behavior event metadata in a store, verifying the data, and analyzing it using artificial intelligence to interpret the data and trigger an associated action, comprising: a data model having: consumer product metadata containing at least: consumer product ID metadata; and location metadata of the consumer product; wherein the system is configured to: receive metadata of consumer behavior events in the store from a plurality of consumer electronic devices, wherein the metadata of consumer behavior events in the store is at least partially obtained from the shopping list GUI displayed by the display device of each of the consumer's electronic devices, wherein the GUI the shopping list contains a list of consumer-specific consumer products and associated purchase confirmation inputs indicating their purchase, the input of metadata of consumer behavior events in the store into a trained AI-based analytic engine, while the AI-based analytic engine is trained using a machine learning algorithm , whose input is the training metadata data of consumer behavior events in the store, and wherein the machine learning algorithm is trained according to at least the purchase confirmation data obtained from the purchase confirmation inputs , wherein the analytic engine comprises: a data interpretation controller comprising: a data verification data interpretation controller and an action launch controller comprising: a data verification action controller, wherein the data verification data interpretation controller is configured to identify a data verification capability match according to behavior event metadata consumer in the store; and generating a data verification action with the data verification action controller, wherein the data verification action is configured according to the store behavior event metadata of the consumer in the store; sending an electronic message about the data verification action to at least one of the plurality of consumer electronic devices; receiving response data regarding item data metadata in response to the data verification action email message from at least one consumer electronic device; and updating the consumer product data with the response data on the product data metadata.

[20] The data interpretation controller may further comprise: a product offer data interpretation controller, and wherein the action trigger controller may further comprise: a product offer action controller, and while operating: the product offer data interpretation controller can be configured to identify a match opportunity product offers according to the metadata of consumer behavior events in the store; and the product offer action controller may be configured to generate and send the product offer action to display the shopping list on the GUI.

[21] The product offer action may contain location data in the product offer store.

[22] The store location data may indicate at least one of a row number and a shelf number.

[23] The electronic device may further comprise a sensor for detecting store location data, wherein the store consumer behavior event metadata may further comprise store location data.

[24] The sensor may include at least one of a near field communication sensor and a Bluetooth Beacon (BLE) sensor.

[25] The electronic device may further comprise an image sensor for capturing an image of the product, and where the product data can be derived from the image.

[26] The item data may comprise at least one of item ID and item cost data.

[27] The data interpretation controller may further comprise: an information notification data interpretation controller; and wherein the action launch controller may further comprise: an information notification action controller, and while operating: the information notification data interpretation controller may be configured to identify and match the information notification capability in accordance with the consumer behavior event metadata in the store, and the information notification action controller may be configured to send a notification to at least one electronic device.

[28] The data interpretation controller can be optimized with a machine learning algorithm.

[29] The trigger action controller can be optimized using a machine learning algorithm.

[30] The machine learning algorithm may be configured to optimize the product offer controller according to the purchase confirmation input received from the shopping list GUI.

[31] The machine learning algorithm may be configured to optimize the data verification data interpretation controller according to the probability of receiving the response data with respect to the product data metadata.

[32] The machine learning algorithm may be configured to optimize the data verification data interpretation controller according to the integrity of the response data with respect to product data metadata.

[33] The machine learning algorithm may be configured to optimize the information controller according to consumer interactions with information notification prompts.

[34] Consumer product metadata may contain at least one product category.

[35] The response data regarding item data metadata may contain item price metadata.

[36] The in-store consumer behavior event metadata may comprise image data, and wherein a machine learning algorithm may be configured to optimize the product offer action controller by object recognition of the image data.

[37] Other aspects of the invention are also disclosed.

Brief description of the drawings

[38] Despite any other forms that may fall within the scope of the present invention, the following will be described, by way of example only, preferred embodiments of the disclosure with reference to the accompanying drawings, in which:

[39] FIG. 1 shows a computer network for aggregating store behavior event metadata, verifying the data, and analyzing it with artificial intelligence to interpret the data and trigger an associated action, in accordance with an embodiment;

[40] FIG. 2 is a computing architecture illustrating in more detail the server and electronics of the network shown in FIG. 1, in accordance with an embodiment;

[41] FIG. 3 shows an example of a data model, controller, and interface module view for each of a server and an electronic device, in accordance with an embodiment;

[42] FIG. 4 shows an example of a preliminary data flow for aggregating store behavior event metadata, verifying the data, and analyzing it with artificial intelligence to interpret the data and trigger the associated action, according to an embodiment;

[43] FIG. 5 shows an example of a shopping list GUI displayed by an electronic device of the system shown in FIG. 2, according to an embodiment;

[44] FIG. 6 shows an example of an information notification prompt displayed by an electronic device of the system shown in FIG. 2, according to an embodiment; and

[45] FIG. 7 shows an example of a personal behavioral interaction with the system.

Description of implementation options

[46] Referring to FIG. 1, it shows an example of a network 100 of consumer electronics 213. As can be seen, the network 100 contains a plurality of consumer electronic devices 213, each of which is in operative communication with the analytic engine server 222 via the data network.

[47] The consumer electronic devices 213 belong to different consumers and may be located in different locations, including those that may be classified by the network 100 into virtual geofences 101, which in an embodiment may represent different stores.

[48] As will be described in more detail below, consumers use such in-store consumer electronic devices 213 such that the network 100 is capable of receiving in-store consumer behavior event metadata from such consumer electronic devices 213 to be able to appropriately perform machine learning and analysis with using artificial intelligence described in this document.

[49] Turning now to FIG. 2, it shows the computer system 200 in more detail, showing the analytic engine server 222 and consumer electronics 213 in the network 100, in accordance with an embodiment.

[50] As can be seen, each of the analytic engine server 222 and consumer electronics 213 may take the form of a computing device with a digital data processor 209.

[51] In embodiments, as compared to having an analytics engine server 222, the system 200 may use distributed processing of a decentralized smart contract blockchain platform, such as the Ethereum™ blockchain platform.

[52] The processor 209 may be in operative communication with the memory 223 via the system bus 208. The memory 223 is configured to store digital data containing computer program code instructions. As a result, during operation, the processor 209 calls these computer code instructions from the memory 223 for execution, and the resulting data can be stored again in the memory 223.

[53] For ease of illustration, these computer code instructions have been divided into modules of data model 201, controller 202, and interface 203.

[54] In general, the data model 201 contains a structure for storing necessary data (such as relational database tables) and data stored therein.

[55] In addition, the interface module 203 manages various aspects of various user interface GUIs.

[56] In addition, the controller 202 performs various computing tasks, including the interaction of the interface modules 203 and the data model 201.

[57] The storage device 223 may further comprise an operating system 207, such as a Linux kernel, or a mobile operating system that is retrieved by the processor 209 during the boot phase. In an implementation utilizing a decentralized smart contract blockchain platform, storage device 223 may store all or part of a corresponding blockchain ledger including data from various transactions described herein.

[58] Each computing device may further comprise an input/output interface 210 for interfacing with various computer peripherals, including storage, sensor, and user interface peripherals.

[59] As shown with respect to consumer electronics 213, I/O interface 210 can interact with digital display 218 to display digital data. The digital display 218 may be covered with a tactile sensor to enable gestures to be set on the user interface.

[60] The input/output interface 210 may also interact with the GPS sensor 216 to enable location of the consumer electronic device 213, including for the purposes of detecting the intersection of the virtual geofence 101 described above.

[61] In yet another embodiment, consumer electronics devices 213 may include other sensors 217, which may include sensors for reading product data from consumer products (e.g., nutritional information and product appearance), row labels (e.g., product location, and row groups of product categories within a row), shelf labels (e.g. specific product code/barcode from a retailer in a store, product name, shelf price, special price status, product range groups), store location, e.g. a number of near field positioning technologies (such as near field communication (NFC) systems) and the like.

[62] Each computing device may further comprise a network interface 211 for transmitting and receiving data via the data network 212.

[63] In embodiments, the analytics engine server 222 resides "in the cloud" and may take the form of a physical rack-mounted server or, alternatively, a virtualized server instance, such as may be implemented through, for example, Amazon Web Services™ (AWS). Alternatively, with reference to the above, the analytics engine server 222 may be in the form of a smart contract blockchain platform, where smart contracts are used to describe the terms of peer-to-peer transactions between blockchain users without the need for a centralized server.

[64] In addition, consumer electronics 213 may take the form of a small form factor electronic device containing suitable electronic circuitry for the purposes of collecting data described herein and sending it to analytic engine server 222 over data network 212.

[65] In an embodiment, consumer electronic device 213 may take the form of a mobile communication device such as a smartphone such as Apple's iPhone and/or Google's Android device or the like. In this embodiment, in order to configure the mobile communications device for the specific computing processes described herein, the user may download the modules 201, 202, and 203 to the storage device 223 via a downloadable software application "app", which may be, for example, downloaded for installation and execution by the consumer's electronic device 213 from a software application store such as the Apple App Store™, Google Chrome App Store™ or Firefox App Store™, or the like.

66] Referring now to Figure 3, it shows example views of the data model 201, controller 202, and interface 203 for each of the analytic engine server 222 and consumer electronics 213.

[67] As shown, the AI-trained analytics engine of the analytics engine controller 202 of the analytics engine server 222 may comprise a data interpretation controller comprising a consumer product offering controller 308, a data verification controller 313, and an information controller 317.

[68] The data interpretation controller modules can identify various matching actions 323 from the data model 201 according to the received in-store consumer behavior event metadata 324 that are received from a plurality of consumer electronic devices 213.

[69] These actions 323 can then be communicated to consumers via the interface module 203 of the analytic engine server 222 and associated consumer electronics 213.

[70] In particular, the interface module 203 of the consumer electronic device 213 may display a shopping list user graphical interface 304 (which is essentially shown in Figure 5).

[71] However, the information may be displayed on the consumer's electronic devices 213 in other ways, for example, through the warning interface 310.

[72] The data model 201 in the analytics engine server 222 may contain consumer product and purchase metadata 301 . In an embodiment, the consumer product and purchase metadata 301 is in the form of a product tree, which, as described in more detail below, is continually updated to reflect changes and the availability of product information, and the like. among various stores.

[73] In implementations, the product tree can be organized into a hierarchy of product categories (such as "fast goods", services such as milk, bread, sugar, haircuts, phone screen repairs, car adjustments, etc.), and in these categories consumer products and services - specific consumer products such as Dairy Farmers milk, Coles milk, Devondale The Creamy One Full Cream 1L milk, Lube Mobile 50000 km service for RENAULT KOLEOS 2013, 2.5L 4 cyl., front wheel drive, petrol, MPFI, 2TRA, DOHC 16V (08-16), etc.

[74] Products in the product tree 301 can have product data 307 associated with various metadata such as product price, product popularity, product group metadata (such bonus packs, Partner product price changes, Nutrition information, Recipes, etc.), Product purchase incentives, Product metadata accuracy, Product metadata update incentives, product location, including store location and store location data ( such as row number, shelf number, etc.) and other applicable product metadata.

[75] In addition, the data model 201 may contain user data 312 representing various consumers using the system 200 (eg, demographics, personal information, personal preferences, friend relationships, family, user-provided metadata, and the like).

[76] Shopping lists 316, the aforementioned consumer-specific shopping lists 316, events 324, actions 323, product data 307, product tree elements 301, locations 319, and user data 312 may be stored in association with users.

[77] Shopping list data 316 may represent various consumer products in each consumer shopping list (including lists of recipe ingredients, Lists of incentives to update product metadata (such as Required price confirmations, Required barcodes), list of informational updates (for example, with similar to ratings, which may encourage switching to other products)), as well as other data, such as an indication whether the consumer bought this item or not, the status of confirmation from other consumers to send the price of the product from the local store, valuable analytical data (in including cost savings for a consumer changing list items to another product or store, forgotten and probably needed items that are not on the shopping list, the location of these items in the consumer's chosen store, items that are likely to better meet the needs of the consumer, than items that are already on the shopping list, etc.), st Imules issued for consumer action events to update metadata, and the like.

[78] As can be seen, the data model 201 of the consumer electronic device 213 may further contain the corresponding list data 306 in the data model 201.

[79] Thus, using the shopping list user GUI 304, the consumer electronic device 213 can update the list 306 using the update controller 305.

[80] The data model 201 of the analytic engine server 222 may further comprise training data 321 used to train the machine learning algorithm 322. In embodiments, the training data 321 may use training data of a consumer's in-store behavior obtained from a plurality of consumer electronic devices 213, and may use data from third-party sources, including social networking sites such as Facebook, statistical data services such as the Australian Bureau of Statistics, and the like.

[81] As shown, each of the product offering, data verification, and data interpretation controller information notification controllers can update the list interface 303 or alternatively generate a notification 309 displayed by an alert that is displayed on the shopping list user GUI 304 or on the the warning interface 310 of a plurality of consumer electronic devices 213 .

[82] Consumer electronic device 213 may include an invitation controller 311 configured to generate invitation emails according to a relevant action, and wherein, in embodiments, especially for data verification actions, information notification actions, and product offer actions, the user can respond with the response interface 314 managed by the response controller 318 with the product metadata response data, which is further received by the receive controller 320 of the analytic engine server 222 to update the product tree 301 using the product tree update controller 302 . In addition, implementations may also provide update controllers configured to update other purchase information such as user data 312, list data 316, store data, and the like.

[83] Referring to figure 4, it shows an example of the movement of data 400 in the system 200.

[84] As shown, store behavior event metadata 425 is received from a plurality of consumer electronic devices 213 .

[85] As mentioned above, the store behavior event metadata 425 of the consumer can be obtained at least in part from the shopping list GUI 304 (which is essentially shown in Figure 5) displayed by the display device of each of the consumer's electronic devices 213, which may further comprise purchase confirmation data indicative of a consumer's purchase of each of the displayed consumer products.

[86] As will be described in more detail below, the shopping list user GUI 304 further comprises a list of consumer-specific items and an associated purchase confirmation input (checkbox or other type of input) indicative of their purchase. Thus, using such a GUI, the system can identify relevant consumer products for each consumer and, in addition, verify whether each listed item has been purchased by the consumer or not, metadata related to relevant products (including availability, likely current price, fast products and the like) at the consumer's chosen store, user metadata (including consumer segmentation and incentive threshold for event metadata collection actions based on relevant products and purchase event duration), store metadata (including likely efforts to complete purchases, and the like) and events with associated metadata (including Metadata Update Incentives for Products, Stores, Listings, Users, Reviews, Prices, Availability, and the like).

[87] Additional information may also be received from each electronic device, such as location data received by the GPS sensor 216.

[88] In addition, other information may be received from each consumer electronic device 213 using other types of sensors 217.

[89] For example, the consumer electronic device 213 may include an image sensor that enables the consumer to, for example, capture product image data so that the system 200 can recognize the product, for example, according to various external features of the product, including reading a barcode indicated on it (including reading the product image into a trained object recognition AI engine).

[90] Alternatively, as opposed to capturing item image data, price data can be captured, for example, by capturing an image of a price tag next to items using sensor 217.

[91] Alternatively, in contrast to capturing product image data, the level of activity, aesthetic appeal, desirability, and the like can be captured, for example, by capturing an image of store fronts, adjacent stores, or a car park using sensor 217.

[92] Alternatively, in contrast to capturing product image data, proof of purchase, store visit, and the like can be captured, for example, by capturing an image of a QR code using the sensor 217.

[93] Alternatively, in contrast to capturing product image data, cost information (price, unit price, special price, discount amount, product, proof of purchase), list information (including purchased products, metadata store, total purchase price, purchase method, registration number, registry operator details, store manager details, volume purchased, quantity purchased, item receipt alias, tax status, etc.), for example, by capturing an image of a purchase receipt from using sensor 217.

[94] In still other embodiments, the sensor 217 is configured to locate a consumer in a store, specifically with row resolution (including row number, row group, and nearby items, and the like), and in still other embodiments, realizations - with the permission on regiments.

[95] For example, a plurality of near field communication tags may be provided in various rows or shelves in a store, which may then be scanned by sensor 217 to determine the location of consumer electronic device 213 in the store.

[96] Alternatively, Bluetooth beacon technology can be used, in which the sensor 217 uses received signal strength measurements to determine the location of the consumer electronics device 213 relative to one or more Bluetooth beacons in the store.

[97] In embodiments, and also referring to Figure 5, such store location data may be obtained from user input, including for consumer electronic devices 213 that do not have such sensors 217.

[98] As shown, event metadata 402 is obtained from consumer behavior event metadata 425 in the store.

[99] Such event metadata 402 may represent various consumer events such as purchasing a product, entering a store (determined by crossing the virtual geofence 101 inward), returning home (determined by crossing the virtual geofence 101 outward), selecting a Fast Item, Product Optimization, Coming Home (crossing the intended location of the "Home" geofence inward), Adoption Incentives (such as Metadata Update, Information Incentives, etc.), Redemption Incentives, and other consumer events.

[100] As shown, event metadata 402 may include event type metadata 403 indicating the type of consumer behavior event and location metadata 404 indicating the location of the event.

[101] Location metadata 404 and consumer behavior events may be obtained from GPS sensor 216 and, in embodiments, may be used by system 200 to determine which store the consumer is currently in, as well as important locations for events outside of the store (including Home, Work, Start and End of the trip, etc.). Locations of important events can be determined by AI models trained to recognize the most optimal locations (for notification, suggestion, inducement, etc.). Models take into account not only the current locations of users, but can also take into account other locations and attributes of other users on the network at a point in time.

[102] The event metadata 402 may further comprise higher resolution store location data 405 representing the user's location in the store, such as a particular row, shelf, and the like.

[103] The event metadata 402 may further comprise list metadata 407, which may include a consumer-specific shopping list, including associated metadata, including metadata indicating the consumer's purchases.

[104] In addition, the event metadata 402 may further comprise product metadata 408, which is metadata applicable to various consumer products. Such item metadata 408 may represent the price of the item, the location of the item, the availability of the item, or other applicable item metadata.

[105] As shown, the event metadata 402 is fed into the trained AI analytics engine 413.

[106] With such data as input, the trained AI analytics engine 413 is configured to generate AI smart inferences 423, which can then be fed back to relevant electronic devices 213 on the network. Such findings are any intelligent data and/or notifications that may be useful to the consumer network during the shopping process.

[107] In particular, the trained AI-based analytic engine 413 may comprise a data interpretation controller 414 . The data interpretation controller 414 interprets the event metadata 402 to be able to match different possibilities to generate related actions.

[108] In particular, as shown, the trained AI analytic engine 413 may comprise an action launch controller 418 .

[109] Thus, for any potential match with the capabilities identified by the data interpretation controller 414 using the aggregated event metadata 402, the trained AI analytics engine 413 can trigger the associated and applicable action using the action trigger controller 418 .

[110] As shown and also referring to the above, the data interpretation controller 414 may comprise a product suggestion controller 308 configured to intelligently suggest various consumer products. Such consumer products may be offered according to consumer-specific parameters (including demographics, past purchasing behavior) as well as other consumer-specific parameters such as the behavioral habits of other consumers.

[111] The data interpretation controller 414 may further comprise a data verification controller 313 configured to identify opportunities to improve the integrity of consumer product metadata 301, list metadata, store metadata, and the like.

[112] For example, if the price of a particular consumer product is "soft" (i.e., identified as potentially unreliable), the data verification controller 313 may identify this to generate an associated data verification action from the data verification action controller 420, which may be passed to the relevant electronic devices 213 of the network for verification by the consumer, for example, by checking the price by the consumer, adding a barcode, checking the row, checking how busy the store is, checking the quality of the goods, and the like.

[113] The data interpretation controller 414 may further comprise an information controller 317 configured to generate various information notification actions from the information notification action controller 421 that can be communicated to relevant electronic devices 213 of the network as a whole by way of push notifications, other GUI alerts. (for example, posts in the news feed, advertised items on the comparison screen), emails, and the like.

[114] As also shown in data flow 400, such aggregated event metadata 402 taken over a number of people can be fed into a machine learning algorithm 322.

[115] In embodiments, the machine learning algorithm 322 may comprise an optimizer configured to optimize each of the data interpretation controller 414 modules. In particular, as shown, the machine learning algorithm 322 may include a product offer model optimizer 410 to generate trained data 422 to optimize the product offer controller 308 . Similarly, the machine learning algorithm 322 may comprise a data verification model optimizer 411 and an information notification model optimizer 412 to optimize the respective data verification controller 313 and information controller 317.

[116] The product offer model optimizer 410 may be configured to optimize the purchase of the offered product. The data verification model optimizer 411 may be configured to optimize the likelihood or probability of accuracy of the wide network metadata and receive feedback from the consumer or, alternatively, the integrity of the received feedback.

[117] The information notification model optimizer 412 may be configured to optimize notifications according to the receipt of such notifications, such as receiving notification 601, as generally shown in Figure 6.

[118] As an example, the machine learning algorithm 322 may use a recurrent neural network (RNN) model trained on aggregated event metadata 402 and may identify that a particular product offering is being purchased by consumers more frequently than another product offering (as determined from shopping list user GUI 304) and thus can adjust the product offer accordingly.

[119] It should be noted that the machine learning algorithm 322 can also be configured to optimize the trigger action controller 418 in order to optimize the triggering of related actions.

[120] As an example, the machine learning algorithm 322 can identify that a consumer is more likely to purchase an offer if such an offer is made before the consumer's electronic device 213 enters the corresponding store's virtual geofence 101, and thus can optimize the trigger action controller 418 accordingly.

[121] As another example, machine learning algorithm 322 can identify that women are more likely to purchase an item if it is offered before crossing virtual geofence 101 inwards, while men are more likely to purchase an item only when it is offered in a store (i.e. e. when the consumer electronic device 213 has already crossed the virtual geofence 101 of the respective store), and can thus optimize the action trigger controller 418 accordingly.

[122] In embodiments, the trained AI analytic engine 413 may take the form of an artificial neural network, and thus the trained data 422 may represent optimized weights for each neural network node.

[123] Referring to figure 5, it shows an example of a graphical user interface 500 displayed on the electronic device 213 consumer.

[124] As shown, the shopping list contains many quick items, which are milk, spaghetti, and bread, and in addition, a specific item, which is Colgate Sensitive toothpaste.

[125] The user could create such a list by entering these positions. In one embodiment, as the user types, the interface 500 uses text prediction to suggest products from the product tree 301. In yet another implementation, the list may be populated with consumer products previously purchased, purchased frequently, or predicted to be needed by the consumer using AI inferences (such as offer actions from previous events by that user, or even by other users on the network). In yet another implementation, the text of the web page data (of the entire web page, or text selected by the user, and the like) can be parsed against a list of product links, or alternatively, mined to overlay element(s) control on the interface 500 (including for offers 501 of the product (for example, adding a product, comparing a product, searching for a product and a new product), as well as invitations 503 for data verification (submission of an overview of the product (including over the network by sending friends in a social networks)).

[126] Now, as shown, interface 500 contains product offers 501 generated by product offer controller 308 and product offer controller 419, as well as data verification prompts 503 generated by data verification controller 313 and data verification action controller 420.

[127] In particular, for the quick product milk, the shopping list interface 501 generates product suggestions, where, for example, this category of fast product offers specific and intelligently suggested products such as Dairy Farmer milk and Paul's milk.

[128] Additional artificial intelligence inferences may be associated with it, including inferences regarding location and prices, and, for example, as can be seen, for the Daily Farmers milk offered, a specific location is shown in the store (third row and second shelf from the top), and for Paul's milk, that Paul's milk is currently the cheapest in the store.

[129] However, as also shown, a product offer can offer a product in a different location given a low price. As shown, item offer 501 additionally contains an offer for Coles milk, which costs only $1 but requires walking 250 m to another location.

[130] For the fast spaghetti item as shown, the item offer 501 can further intelligently recommend a particular type of spaghetti. As mentioned above, the suggested type of spaghetti can be intelligently generated by the trained AI analysis engine 413 to be optimized for consumer purchase. For example, a particular type of consumer product can be derived from consumer-specific metadata such that a particular consumer is offered gluten-free spaghetti.

[131] Alternatively, a specific suggested type of spaghetti may be suggested according to an analysis of event metadata 402 received from other consumers, including similar consumers, and other parameters.

[132] As shown, product suggestion 501 can also intelligently suggest related items such as pasta sauce, with machine learning algorithm 322 taking into account that consumers often buy these types of products together, and they may fall into different product categories.

[133] Using the trained AI analysis engine 413 can further generate non-intuitive AI outputs when, for example, for a quick item of bread, the interface 500 can suggest shoe polish. While it may seem counter-intuitive, shoe polish can actually be quite relevant to the user.

[134] As also shown, the interface 500 displays the data verification prompt 503 when, as mentioned above, the data verification controller 313 has identified an opportunity to improve the integrity of the product data 307.

[135] As shown, for Colgate Sensitive toothpaste, a 503 data verification prompt asks the user if the price is indeed $2.99 and furthermore, if it is indeed in row 3. As shown, the user can check the box if it exists, or alternatively enter the correct number.

[136] In yet another implementation, the system 200 may be configured to perform an OCR analysis of a consumer's purchase receipts to be able to collect pertinent product data 307 as appropriate, including price data.

[137] As shown, the interface 500 includes a check OCR button 504 that a user can use when capturing an image of a purchase receipt using the image sensor 217 in the consumer electronics device 213.

[138] For each of the consumer products on the shopping list interface 500, the user can use the associated purchase confirmation input (checkbox or the like) to indicate whether the product has been purchased or not.

[139] Thus, by using such purchase behavior feedback obtained from purchase confirmation inputs, the machine learning algorithm 322 is further configured to train the AI analytics engine 413 appropriately.

[140] Using the trained AI-based analytics engine 413 can further generate AI anti-fraud inferences where, for example, a user can maliciously make incorrect price updates or deceptive photographic evidence of their purchase, possibly to claim a reward from a manufacturer.

[141] The trained AI analytic engine 413 can be configured to identify outliers, for example, by using a single class support vector machine (SVM) or a single class neural network model. The data may be associated not only with purchase behavior, but also with pre-purchase behavior, such as when and where items were added to the list, as well as the user's past history of such updates and queries; the data verification controller 313 can identify deviations and unusual behavior. This can generate an associated data verification action from the data verification action controller 420, which can be communicated to the user's electronic device to ask the user to further prove the purchase of the item (eg, scan a barcode at home).

[142] As another example, a manufacturer may wish to know the current status of their products in a store on a shelf, in a row, in a particular local supermarket, to make sure that their product is positioned correctly, or that the shelf is properly and correctly stocked. When a user makes purchases at that local supermarket, the data verification controller 313 may be able to predict from the user's purchase confirmation input of other items in the store (such as input via a check box on a user interface or the like) and other current application behavior (such as, for example, crossing the perimeter of the geofence inward) that the user can be in the same row as the product for which the manufacturer requires information about the current state of the store. Thus, the trigger action controller 418 can generate an associated data verification action with the data verification action controller 420, which can be communicated to the user's electronic device to prompt the user to take a photo of the shelf for the manufacturer, possibly for a reward.

[143] As another example, a user may provide a description of a new product. The trained AI analytics engine 413 can be configured to generate AI smart categories for fast goods, for example by using a multiclass decision forest or a multiclass neural network model. This will generate associated data verification actions from the data verification action controller 420, which can then be communicated to the relevant network electronic devices 213 so that these users can verify and validate the proposed categorizations.

[144] Referring to Figure 6, it shows an information notification 601 that was generated by the information controller 317.

[145] As can be seen, the notification 601, after locating the consumer's electronic device 213, alerts the user to proximity to a smart bid and furthermore displays information about the set price of such a bid.

[146] Using button inputs, the user can initiate a navigation interface to navigate to a product or, alternatively, dismiss the warning.

[147] Various exemplary implementations will now be presented to further illustrate the functionality of the system 200. It should be noted that these implementations are provided by way of example only and that not all implementations need to be applied to the technical limitation as appropriate.

[148] In the first example, the user selects a fast item using the shopping list GUI 500.

[149] In response, the product suggestion controller 308 intelligently suggests specific products associated with the fast product. In addition, if any item metadata associated with the fast item has been identified by the data verification controller 313 as unreliable ("soft"), the data verification controller 313 may trigger a data verification action from the data verification action controller 420 to collect additional item metadata with the appropriate manner.

[150] In addition, the information controller 317 can identify relevant information regarding the fast item and the notification action accordingly.

[151] As mentioned above, the machine learning algorithm 322 could train the data interpretation controller 414 and/or the action trigger controller 418 to optimize the likelihood of a consumer accepting or otherwise interacting with the product offering, data verification request, and/or informational notification.

[152] In another example, the system 200, by locating the consumer electronic device 213, determines that the consumer electronic device 213 has crossed the perimeter of the virtual geofence 101 surrounding the store inward. Thus, the system 200 is able to logically conclude that the user is likely to go shopping.

[153] Upon identification of such a possible event, the data interpretation controller 414 may implement any potential appropriate actions from the action trigger controller 418 .

[154] In addition, various relevant data can be updated in response to such an event, in which, for example, the system 200 can retrieve product data, product special prices, and the like from the corresponding store. In addition, system 200 can parse metadata 402 of recent or substantially real-time events associated with other consumers within the same area. Thus, the data interpretation controller 414 can function using such updated information.

[155] In yet another implementation, the user confirms the purchase or selection (eg, by taking an item from a store and placing it in a cart or cart) of an item using a purchase confirmation input, such as a check box input, as essentially shown in figure 5.

[156] Thus, based on the specific product indicated, system 200 can infer other information, such as the location of the consumer in the store now, for example, with row resolution. Thus, the product suggestion controller 308 then has the ability to suggest other products at the same location in the store, such as other products in the same or adjacent rows.

[157] Also, after receiving confirmation of taking such an item, the item suggestion controller 308 thus may not recommend the same or similar items.

[158] In addition, the shopping list interface 500 may suggest the next item to take according to proximity.

[159] Other information may be obtained from the consumer's interaction with the shopping list GUI 500, such as the time to complete purchases, for example, between the text timestamps of the first and last items on the shopping list interface 500.

[160] In addition, the frequency or number of interactions of consumers using the interface 500 can be used to infer the occupancy of a particular store to enable the trained AI-based analytics engine 413 to possibly suggest shopping at a different time or location. place. Such information may also be obtained from the user through an information notification prompt, wherein the prompt prompts the user for the current store occupancy status or other information in the store.

[161] In yet another implementation, using the data verification prompt 503, the user provides the price of a consumer product in a store.

[162] Thus, the data verification controller 313 may determine that the consumer is willing to provide such feedback and may therefore increase the rate of subsequent requests.

[163] In addition, in this case, for a particular product, the data verification controller 313 may request information regarding related consumer products, such as other consumer products in the same row.

[164] In yet another example, the system 200 detects the consumer's electronic device 213 crossing the virtual geofence 101 outwards, indicating that the consumer has left the store location. Thus, the system 200 is able to logically conclude that the user has completed the purchases.

[165] Thus, in embodiments, the product offering controller 308 may offer a consumer product that the consumer is likely to need at other locations near the store's location, such as by recommending a newspaper from a nearby news outlet.

[166] In addition, the event metadata 402 can be analyzed to identify correlations between consumers and store locations, for example, by evaluating the convenience of parking, by analyzing the volume of consumers within a parameter, and the time it took to complete a shopping trip.

[167] In yet another example, system 200 may detect an inward crossing of home virtual geofence 101 indicating that the user has returned home. Thus, the system 200 may logically conclude that the user is likely to unpack purchased consumer goods.

[168] Thus, during the unpacking process, the data verification controller 313 may issue an invitation to the user through a data verification action from the data verification action controller 420 to provide metadata related to the various purchased items.

[169] In addition, the data verification controller 313 may receive OCR data on the purchase receipt to be able to establish product metadata based on them, for example, by matching inline elements with a keyword and identifying related cost information (price, price per unit, special price, discount amount, product, proof of purchase), listing information (including purchased items, store metadata, total purchase price, purchase method, registration number, registry operator details, store manager details, volume purchased , purchased quantity, pseudo-name of the goods receipt, taxation status, etc.).

[170] In addition, the data verification controller 313 may receive OCR data on purchase-related documents (including product package labels, row labels, shelf labels/labels, and the like) to be able to establish purchase metadata (including product, row, shelf, etc.) based on them, for example, by matching characters to a keyword and identifying related shopping information (including nutritional information, rating, product location, grouping into a number of product categories within line, code/barcode of a specific product from a retailer in a store, product name, shelf price, special price status, product range groups, etc.).

[171] In addition, for any consumer products having associated special store requirements for purchasing such product, the information controller 317 may issue an invitation to the user through an information notification action from the information notification action controller 421, where, for example, the invitation may inform the user that if the user captures an image of a particular product, then the user will be eligible to claim a reward for a future shopping trip.

[172] Referring to Figure 7, an example of a user scenario 700 is shown that depicts the elements and functionality of system 200. It should be noted that scenario 700 is provided as an example only and that technical limitations should not necessarily be applied to all implementations of system 200. accordingly. In addition, in the scenario example 700, the store behavior event metadata 425 is shown as circles, the data verification actions from the data verification action controller 420 are shown as parallelograms, the information notification actions from the information notification actions controller 421 are shown as rectangles, the product offer actions from the action controller 419, product offers are shown as ellipses, and data updates for the data model 201 are shown as hexagons.

[173] Next, script 700 begins at step 712, where user A adds 120g HP Sauce without price or barcode to a shopping list for local Store 1 while at home. Thereafter, the event 712 may cause the system 200 to trigger a series of data verification actions from the data verification controller 313, comprising a first action 713 in which the system 200 sends data verification requests 713 to other consumer electronic devices 213 to request a price from users who have recently purchased that same goods. Alternatively or in addition, action 714 may be sent to other consumer electronic devices 213 to request barcodes from other users who have recently purchased the same item.

[174] After receiving data verification actions 713, 714, user B's electronic device may generate an event 716 in which user B updates the price of the item with a photo of the price entry on his purchase receipt while at home. Thus, the event 716 may also cause the system 200 to perform a data update 717 to update the price of the position with the cost information provided, but set a flag representing that the price is "not confirmed". In addition, the system 200 may then initiate a data verification action 718 to generate a data verification action to request confirmations of value from other users' other electronic devices, for example, to verify that a photo of a price entry on their purchase receipts matches the item and price from the action. 713 data verification.

[175] In addition, when User A performs an add product event 712 (by printing search terms that match the product, by capturing an image of the product packaging, by scanning a barcode, or the like), the system 200 may generate an AI-generated product suggestion action 711, which, for example, is recommended to User A for adding to the additional item category Oven-Baked Potatoes in the shopping list, from which the system 200 has learned using the machine learning algorithm 322 that it is usually associated with 120g HP Sauce.

[176] The event 712 may further cause the system 200 to initiate an action 715 to verify the request for price data for the consumer electronic device 213 at User C, given that the system 200 has determined that User C is likely to be in Store 1 or in the same aisle soon. Such a prediction may be generated by a machine learning algorithm 322 that is trained on various event metadata 402, including location metadata 404 and consumer behavior events.

[177] Upon receiving a price request data verification action 715, User C's electronic device may generate an event 701 in which User C updates the price in Store 1, for example, from $1 to $1.30, and captures an image of the price tag or label on the shelf. In response, the system 200 performs a data update 702 to update the store price and, if applicable, updates the corresponding price of the respective stores of the same retailer (price harmonization).

[178] In addition, the data verification controller 313 may generate a data verification action from the data verification action controller 420 and send a validation request 703 to the consumer electronic device 213 of the other User D to confirm the price. After User D confirms that the price on the shelf matches the photograph taken, the system 200 performs a price confirmation update 705.

[179] Additionally, in response to events 701, the system 200 may trigger an offer 706 to notify all other users of the system who have "Brown Sauce" (i.e., item category) on their respective shopping lists, which, for example, is in effect special price for 120g HP Sauce, $1.30. In this regard, an information notification action 707 can be sent to other electronic devices.

[180] The system 200 may optionally generate an additional product offer 708 to notify all other users who have 120g HP Sauce on their shopping lists that there is now a special price for this item of $1.30 in Store 1. Note that by User C updating the price during event 701, the system 200 can, by checking the relevant metadata, notify the first user of the new price of the item that was originally added by the first user, even if the first user was not specifically asked to update the price.

[181] The price update event 701 may further include sending an information notification action 709 that notifies users who may be shopping in Store 1" that there is currently at least one other active other user (which may be of interest to others) in the store. users for things like price check requests, availability requests, and the like). After User E receives the information notification action 709, the electronic device 230 may generate an event 719 in which User E wishes to know the price of Huggies diapers in Store 1 and suggests User C 50 outlets where they can be found. In response, system 200 may generate a data verification action 720 to send a notification to User C's consumer electronic device 213.

[182] Event 701 may additionally trigger another item offer 710 that notifies users who have 120g HP Sauce on their shopping lists in Store 2 that the item likely also has a special price of $1.30, given that that Store 2 is operated by the same retailer.

[183] In the above description, for purposes of explanation, specific nomenclature has been used to provide a thorough understanding of the invention. However, the person skilled in the art will be clear that the implementation of the invention in practice does not require specific details. Thus, the above description of specific embodiments of the invention has been presented for purposes of illustration and description. They are not exhaustive and are not intended to limit the invention to the exact forms disclosed; obviously, many modifications and variations are possible in view of the solutions presented above. Embodiments have been selected and described to best describe the principles of the invention and its applications in practice, thus, they provide other specialists in the art with the possibility of the best use of the invention and various embodiments with various modifications suitable for a particular intended application. The following claims and their equivalents are intended to define the scope of the invention.

Claims (34)

1. A system for aggregating metadata of consumer behavior events in a store and verifying data, comprising: an analytical engine server configured to communicate with a network of consumer electronic devices via a data network, comprising: a data model having: consumer product metadata, comprising at least: consumer product ID metadata; and location metadata of the consumer product; and trained analytical engine based on artificial intelligence; wherein the analytic engine server is configured to: receive metadata of consumer behavior events in the store from a plurality of consumer electronic devices, wherein the metadata of consumer behavior events in the store is at least partially obtained from the GUI of the shopping list displayed by the display device of each of the consumer electronic devices, when This shopping list GUI contains a list of consumer-specific consumer products and associated purchase confirmation entries indicating their purchase, input metadata of consumer behavior events in the store into a trained AI-based analytic engine, wherein the AI-based analytic engine is trained with a machine learning algorithm, the input of which is the training data of the metadata of consumer behavior events in the store, and the machine learning algorithm is trained at least according to purchase confirmation data obtained from purchase confirmation inputs, wherein the AI-based analytics engine comprises: data interpretation controller containing: a data interpretation data verification controller and an action trigger controller, comprising: a data verification action controller, wherein the data verification data interpretation controller is configured to identify a data verification capability match according to the store behavior event metadata of the consumer in the store; and generating a data verification action with the data verification action controller, wherein the data verification action is configured according to the metadata of consumer behavior events in the store; sending an electronic message about the data verification action to at least one of the plurality of consumer electronic devices; receiving response data with respect to product data metadata in response to the data verification action email message from at least one consumer electronic device; and updating consumer product metadata with response data on product data metadata. 2. The system according to claim 1, characterized in that the data interpretation controller further comprises: a product offer data interpretation controller, wherein the action trigger controller further comprises: a product offer action controller, and while functioning: the product offer data interpretation controller is configured to identify a product offer opportunity match according to the metadata of consumer behavior events in the store; and the product offer action controller is configured to generate and send the product offer action to display the shopping list on the GUI. 3. The system of claim. 2, characterized in that the product offer action contains location data in the product offer store. 4. The system of claim 1, wherein the store location data indicates at least one of a row number and a shelf number. 5. The system of claim 3, wherein the electronic device further comprises a sensor for detecting store location data, wherein the store behavior event metadata further comprises store location data. 6. The system of claim 5, wherein the sensor comprises at least one of a near field communication sensor and a Bluetooth Beacon (BLE) sensor. 7. The system according to claim. 1, characterized in that the electronic device further comprises an image sensor for capturing an image of the product, and the product data is obtained from the image. 8. The system according to claim 7, characterized in that the product data contains at least one of the product ID and product value data. 9. The system according to claim 1, characterized in that the data interpretation controller further comprises: an information notification data interpretation controller; and wherein the action trigger controller further comprises: an information notification action controller, and while in operation: the information notification data interpretation controller is configured to identify and match the information notification capability according to the store behavior event metadata of the consumer, and the information notification action controller is configured to send the notification to at least one electronic device. 10. The system according to claim 1, characterized in that the data interpretation controller is optimized using a machine learning algorithm. 11. The system according to claim 1, characterized in that the action launch controller is optimized using a machine learning algorithm. 12. The system of claim 1, wherein the machine learning algorithm is configured to optimize the product offer controller according to purchase confirmation input received from the shopping list GUI. 13. The system according to claim. 1, characterized in that the machine learning algorithm is configured to optimize the data verification data interpretation controller according to the probability of receiving response data in relation to the product data metadata. 14. The system according to claim. 1, characterized in that the machine learning algorithm is configured to optimize the data verification data interpretation controller according to the integrity of the response data in relation to the product data metadata. 15. The system of claim. 1, characterized in that the machine learning algorithm is configured to optimize the information controller according to consumer interactions with information notification prompts. 16. The system of claim. 1, characterized in that the consumer product metadata contains at least one product category. 17. The system of claim. 16, characterized in that the response data in relation to the product data metadata contains product price metadata. 18. The system of claim. 7, characterized in that the metadata of consumer behavior events in the store contains image data, and the machine learning algorithm is configured to optimize the product offer action controller by recognizing the image data object.

Images