AU2020277109A1 - Computational assessment system - Google Patents

Abstract

A computational assessment system (1) comprising: a server (3) including at least one processor (5), wherein the server (3) is in communication with a plurality of distributed nodes (9); and wherein the at least one processor (5) is configured to: receive (111) an assessment request (13) including: a purported image (15) of a product (17) associated with the assessment request (13); a description (18) associated with the assessment request (13). The processor is further configured to: send (113), to the plurality of distributed nodes (9), the purported image (15) of the product (17) and the description (18), wherein in response each of the distributed nodes (9) performs an individual node assessment (19) indicative of: (i) a result of product verification (21) that the purported image (15) of the product (17) matches a proof imagery (23); and (ii) a result of description verification (25) that the description (18) associated with the assessment request (13) matches the purported image (15) of the product (17). The processor is further configured to: receive (115) a combined node assessment (27) comprising the individual node assessment (19) from each the plurality of distributed nodes (9); and determine (117) a final assessment (29) based on statistical analysis of the combined node assessment (27). 2/7 119 5 101 Capture an image of there product 103 Receive a descriptions associated with an assessment request 105 Send an assessment request including the image and descrptionReceive an assessment description request including a purported image of the product and a description associated with the assessment request Send, to a plurality of distributed nodes, the image and the description Receive image and description 11310 3106 Perform individual node assessment 107 Send individual node assessment Receive a combined node assessment comprising the individual node assessment 10 from the plurality of distributed nodes 115 Determine a final assessment based on statistical analysis of the gross node assessement 117 Fig. 2

Description

2/7

119 5

101 Capture an image of there product

103 Receive a descriptions associated with an assessment request 105 Send an assessment request including the image and descrptionReceive an assessment description request including a purported image of the product and a description associated with the assessment request

Send, to a plurality of distributed nodes, the image and the description Receive image and description 11310 3106 Perform individual node assessment 107 Send individual node assessment Receive a combined node assessment comprising the individual node assessment 10 from the plurality of distributed nodes 115

Determine a final assessment based on statistical analysis of the gross node assessement 117 Fig. 2

"Computational assessment system"

Technical Field

[0001] The present disclosure relates to a computational assessment system and method. In some examples, the system and method can be used for making a computational assessment based on captured images.

Background

[0002] The insurance industry typically relies on human operators to review and assess insurance claims. This is a costly exercise that involves substantial human effort to review the information and make respective determinations. It involves having a dedicated person spending significant time and effort in relation to paperwork and associated communication with the claimant.

[0003] This issue is amplified in first world countries where human wages and salaries mean that the cost of processing an insurance claim can be approximately $150 AUD. This can exceed the cost of the insured product itself, where in an ecommerce environment it may be as low as several dollars AUD making insurance uneconomical as the processing costs exceed the claim amount.

[0004] There is an absence of an automated technical solution for affordable insurance claims management at scale in relation to high volume, low value goods that are purchased online.

[0005] Throughout this specification the word "comprise", or variations such as "comprises" or "comprising", will be understood to imply the inclusion of a stated element, integer or step, or group of elements, integers or steps, but not the exclusion of any other element, integer or step, or group of elements, integers or steps.

[0006] Any discussion of documents, acts, materials, devices, articles or the like which has been included in the present specification is not to be taken as an admission that any or all of these matters form part of the prior art base or were common general knowledge in the field relevant to the present disclosure as it existed before the priority date of each claim of this application.

Summary

[0007] There is disclosed a computational assessment system comprising: a server including at least one processor, wherein the server is in communication with a plurality of distributed nodes; and wherein the at least one processor is configured to: receive an assessment request including: a purported image of a product associated with the assessment request; a description associated with the assessment request; send, to the plurality of distributed nodes, the purported image of the product and the description, wherein in response each of the distributed nodes performs an individual node assessment indicative of: (i) a result of product verification that the purported image of the product matches a proof imagery; and (ii) a result of description verification that the description associated with the assessment request matches the purported image of the product; receive a combined node assessment comprising the individual node assessment from each the plurality of distributed nodes; determine a final assessment based on statistical analysis of the combined node assessment.

[0008] In some examples, the individual node assessment is further indicative of: (iii) a result of quality assessment that the purported image of the product is sufficient to verify the product and the description.

[0009] In some examples of the computational assessment system, the at least one processor is further configured to: store a result of the final assessment for the assessment request in a data store.

[0010] In some examples of the computational assessment system, the at least one processor is further configured to: send a result of the final assessment to a third-party node.

[0011] In further examples of the computational assessment system, the third-party node is associated with one or more of: an insurer; ecommerce platform or ecommerce marketplace; a retailer; payment gateway or providers; finance providers.

[0012] In some examples, the assessment request is received from the third-party node.

[0013] In some examples of the computational assessment system, the assessment request is received from a claimant node associated with a customer.

[0014] In some examples of the computational assessment system, the at least one processor is further configured to: send an authorisation for a claim settlement, associated with the assessment request based on the result of the final assessment.

[0015] In some examples of the computational assessment system, the statistical assessment comprises: determining a statistical majority corresponding to the results of product verification and results of description verification in the combined node assessment.

[0016] In some examples of the computation assessment system, to determine a final assessment further comprises applying one or more threshold rule.

[0017] There is also provided a computational assessment system comprising: a plurality of distributed nodes; a server including at least one processor is configured to: receive, an assessment request including: a purported image of a product associated with the assessment request; a description associated with the assessment request; send, to the plurality of distributed nodes, the purported image of the product and the description, wherein each of the plurality of distributed nodes are configured to: verify that the purported image of the product matches a proof imagery; verify that the description associated with the assessment request matches the purported image of the product; send, to the server, an individual node assessment based on the results of verification of the purported image and the description wherein the at least one processor of the server is further configured to: receive a combined node assessment comprising the individual node assessment from each the plurality of distributed nodes; determine a final assessment based on statistical analysis of the combined node assessment.

[0018] In some examples, the computational assessment system further comprises: a claimant node configured to: capture the purported image of the product; receive the description associated with the assessment request; send the assessment request to the at least one processor of the server.

[0019] In some examples, the computational assessment system further comprises: a third party node to: receive a result of the final assessment from the server; and send an authorisation for a claim settlement associated with the assessment request based on the result of the final assessment.

[0020] The method may further comprise: receiving an assessment request including: a purported image of a product associated with the assessment request; a description associated with the assessment request; sending, to a plurality of distributed nodes, the purported image of the product and the description, wherein in response each of the distributed nodes performs an individual node assessment indicative of: (i) a result of product verification that the purported image of the product matches a proof imagery; and (ii) a result of description verification that the description associated with the assessment request matches the purported image of the product; receiving a combined node assessment comprising the individual node assessment from each the plurality of distributed nodes; determining a final assessment based on statistical analysis of the combined node assessment.

[0021] The method may further comprise: sending an authorisation for a claim settlement, associated with the assessment request based on the result of thefinal assessment.

[0022] The method may further comprise: receiving, an assessment request including: a purported image of a product associated with the assessment request; a description associated with the assessment request; sending, to a plurality of distributed nodes, the purported image of the product and the description, verifying, at each of the plurality of the distributed nodes, that the purported image of the product matches a proof imagery; verifying, at each of the plurality of distributed nodes, that the description associated with the assessment request matches the purported image of the product; sending, by each of the plurality of distributed nodes, an individual node assessment based on the results of verification of the purported image and the description; receiving a combined node assessment comprising the individual node assessment from each the plurality of distributed nodes; determining a final assessment based on statistical analysis of the combined node assessment.

[0023] The method may further comprise: capturing the purported image of the product; receiving, from a user interface, the description associated with the assessment request.

[0024] The method may further comprise: sending an authorisation for a claim assessment associated with the assessment request based on the result of the final assessment.

Brief Description of Drawings

[0025] Fig. 1 is a schematic of a computational assessment system;

[0026] Fig. 2 is a flow diagram illustrating a method of computational assessment;

[0027] Fig. 3 is a representation of an assessment request;

[0028] Fig. 4 is a representation of five separate individual node assessments;

[0029] Fig. 5 is a representation of a combined node assessment and resultant final assessment;

[0030] Fig. 6 is a graphical representation of the flow of information in the computational assessment system; and

[0031] Fig. 7 is a schematic example of a processing device.

Description of Embodiments

[0032] Brief overview of the system 1 and method 100

[0033] An example of the computational assessment system 1 will be described with reference to Fig. 1. The system 1 includes a server 3 having at least one processor 5 and associated data store 7. The server 3 is a node that, through a communication network 6 such as the internet, communicates with other nodes to perform the assessment method 100.

[0034] This include a first node 11 which is typically associated with a customer or claimant who in turn owns, possesses, or have other rights or claims in relation to a product 17. The first node 11 (in some examples herein referred to as a claimant node), may include a smart phone, tablet computer, laptop computer, desktop computer, or other communication device. An assessment request 13 may be initiated at the first node 11.

[0035] A plurality of distributed nodes 9 are in communication with the server 3, whereby the plurality of distributed nodes 9 each function to perform individual node assessments 19 in relation to the assessment requests.

[0036] A third-party node 31 may be in communication with the server 3 to perform additional actions. This can include authorising additional actions based on the result of the computational assessment, such as an authorisation of claim settlement. In some examples, the third-party node 3 is associated with an insurer. In other examples the third-party node 3 is associated with a merchant, ecommerce marketplace, ecommerce platform, payment service, payment providers, financial institution, etc.

[0037] Fig 2 shows and example of a method 100 of computational assessment. This includes making an assessment request 13, where the request includes an image of the product 17 and a description 18 associated with the request (as illustrated in Fig. 3). This information can be captured at the claimant node 11 and sent 105 to, and received 111 by, the processor 5 of the server. The server 5 then sends 113 the image 15 and description to the plurality of distributed nodes 9.

[0038] The distributed nodes 9 function to perform individual node assessments (19). This includes each node individually, and independently, verifying the product in the purported image 15 matches a proof imagery 23 (e.g. a reference image). This also includes each node individually, and independently, verifying that the description 17 associated with the assessment request matches what is shown in the purported image of the product 17. Examples of the result of each individual node assessment is shown in Fig. 4 that shows five separate results of individual node assessments performed by five different nodes of the distributed nodes 9.

[0039] The multiple individual node assessments are then sent to, and received 115 by the server 3 to form a combined node assessment 27 as illustrated in Fig 5. The server 3 performs analysis of the combined node assessment 27, and based on this analysis determines a final assessment 29 for the assessment request 13. This final assessment 29 can include an approval or rejection of the assessment request 13.

[0040] An advantage of this system and method is to obtain multiple assessments (from the plurality of nodes) that can diversify risks compared to assessment made by a single assessor. Each of the plurality of nodes 9 may include a different assessment methodology, but having a number of different nodes 9 making the assessment can lower the risk of error in the final assessment.

[0041] A further advantage is that using a plurality of distributed nodes 9 can reduce costs. This includes sending the information to distributed nodes 9 that have spare capacity, abundant resources to perform the tasks, or can perform the tasks at the lowest cost. Thus in some examples, the server 3 may send 113 the image and description 18 to distributed nodes 9 that are available at the lowest cost.

[0042] Detailed examples of components will now be described in further detail. The non limiting examples will be described in the context of using the computational assessment system aiding assessment of insurance claims (as illustrated in the schematic diagram in Fig. 6). In particular an insurance claim by the insured person, claimant 10, in relation to the product 17, whereby the server 3 in conjunction with the distributed nodes 9 provide a final assessment. The final assessment may be sent to a third-party node 31 that is associated with an insurer that provided insurance coverage to the claimant 10 and product 17.

[0043] The first node 11

[0044] The first node 11, or claimant node 11 in some specific examples, is a computer or other communication device. Typically, this includes a smart phone, tablet, or other computer that is connected to the intemet and has access to a camera. The camera is used to capture an image of a product to be assessed.

[0045] The first node 11 may also include a webpage browser or application to communicate with the server 3 and/or third-party node 31. This communication can include accessing a portal or application to make an assessment request 13, that enables the claimant to upload the image 15 and to include a description 18 associated with the assessment request 13. The description may be entered as text via a user interface of the first node 11, or a selection of pre-defined text options, or a combination of both.

[0046] In alternative examples, the assessment request 13 may be send as a text message, chat message, or electronic mail (email) from the first node 11 to the server 3 and/or third party node 31.

[0047] The server 3

[0048] The server 3 functions to receive the assessment requests 13 and forward information to the plurality of distributed nodes 9, and to determine the final assessment 29.

[0049] In some examples, the server 3 is a dedicated server 3 for these functions and operated by a service provider to perform and coordinate the computational assessment on behalf of another party (such as an insurer). In some examples, the server 3 is a centralised server. In some other examples, the server 3 is a distributed server 3 comprising a network of nodes (and corresponding processors) in communication over a network. In yet further examples, the server 3 may be performed by a cloud-based service or service provider.

[0050] In some other examples, the server 3 is part of (or otherwise associated with) the third-party node (e.g. the insurer). That is, the insurer may operate the computational assessment system internally.

[0051] The server 3 is in communication with the first node, plurality of distributed nodes, and (if separate) the third-party node 31 over a network 6. The network 6 may include a wide area network, such as the internet, a local area network, and/or other combinations. For example, first node 11 and server 3 may communicate via the internet, whilst the server 3 and the plurality of distributed nodes 9 communicate via a separate network for security reasons.

[0052] The server 3 is associated with a data store 7 that can be part of the server 3, or accessible to the server. The data store 7 can store data such as proof imagery 23. In this context, the proof imagery 23 is a control/reference image of the product(s). For example, if the insured product was a shoe, the data store 7 may have a proof imagery 23 of that product to provide to the distributed nodes 9. The data store 7 may also store other information, such as details of the claimant 10, the product 17, purchasing history, claim history, assessment request(s) 13, results of individual node assessments 19, combined node assessments 27, and final assessments 29. The data store 7 may also store details of the distributed nodes 9.

[0053] The plurality of distributed nodes 9

[0054] The plurality of distributed nodes 9 each include respective processing devices. The distributed nodes 9 are configured to receive data associated with the assessment request 13, in particular the purported image 15 and description 18, and to send a result of verification of that data.

[0055] The plurality of nodes 9 may include operators to perform or approve, at least part, of the verification process. This may include entering or confirming the results of the product verification, description verification, or quality assessment. In some examples the operators of the nodes are microworkers to provide such inputs for the verification process.

[0056] The plurality of nodes 9 may be part of a wide network that includes nodes distributed around the world. This can advantageously provide availability and short response times to ensure the overall computational assessment is finalised in a timely manner.

[0057] The server 3 may select specific nodes to make up the plurality of distributed nodes 9 for each assessment request based on availability, service record, and costs. This may include a reverse auction for the lowest cost node 9 to perform the individual node assessment 19.

[0058] The third-party node 31

[0059] The third-party node 31 includes a processor and a data store. In the example where the third-party node 31 is associated with an insurer, the data store may store details of the claimant(s) 10, respective insurance policies, details of the product 17, proof imagery of the product 15, details of the product transactions, payment details, etc.

[0060] In some examples, the third-party node 31 may share information in the data store with the server 3 (and conversely the data store 7 of the server 3 shares information with the third-party node 31). In some other examples, the server 3 and third-party node 31 share a common data store. This can include accessing common data in cloud data storage.

[0061] In some examples, where the insurer operates the computation assessment themselves, the third party node 31, may be common with the server 3.

[0062] Example of a computational assessment of an insurance claim

[0063] i. Selecting insuranceand initiatinga claim

[0064] Insurance coverage to the claimant 10 can be offered before, during or after purchase of the product 17. For example, insurance coverage can be an option at checkout when purchasing the product 17. That is, the end consumer, purchaser, or their agent may elect to select insurance coverage. This may be offered by the merchant website, or as an option at the payment website, application, platform, marketplace or portal. In another example, insurance coverage can be applied across all of the merchant's orders. In some examples, this may include the merchant (e.g. online retailer) buying the insurance for one, a subset, or all goods sold and/or dispatched. Therefore the insured person may be the end consumer, purchaser, their agent, or the merchant.

[0065] In this example, the product 17 is delivered to the claimant 10 but in a damaged state. The claimant 10 then proceeds to initiate an insurance claim which, in part, initiates an assessment request 13. This can include the claimant 10 visiting a website (or application) of the insurer or payment website, application, platform, marketplace or portal, such as a website associated with the third-party node 3. In some examples, the claimant may be the insured person(s) noted above. However, in other examples another entity or agent may initiate the insurance claim. This may be a third party platform designated to process and such claims on behalf of the insured persons. In yet another example, the merchant, on receiving information such as a complaint from the insured person such as the customer, may initiate the insurance claim process on behalf of the customer. In alternative examples, the claimant 10 may make the claim through the merchant website, application, which in turn passes the request to the third-party node 31 or server 3. The website or application then prompts the claimant 10 to provide claim information.

[0066] Referring to Fig. 2, this can include asking the claimant 10 to capture 101 an image of the damaged product 101. This can include using a camera of the claimant node 11. The claimant 10 is prompted to enter a description 18 associated with the claim. For example, the claimant 10 may describe the claim as associated with "broken parts" to the product 17. This received 103 description and image 15 forms at least part of the assessment request 13 (visually represented in Fig. 3) that is then sent for computational assessment.

[0067] In one example the assessment request 13 is sent 105 directly to the server 3. In other examples, the assessment request 13 is sent indirectly to the server 3, such as first sending the claim information to the third-party node 31 that in turn sends an assessment request 13 to the server 3.

[0068] ii. Sending tasks to plurality of distributednodes 9 andprocessing the tasks

[0069] Once the assessment request is received 111 at the server 3, the server then sends assessment tasks to individual nodes, in the plurality of nodes 9. In preferred examples, the server 3 automatically sends tasks to these nodes. In some examples, the nodes a selected of factors including availability of the node to perform the tasks and costs. In other examples, the nodes may be selected based on regulatory reasons, such as location requirements. Other factors may include ratings, historical accuracy, etc. of results from the node.

[0070] In one example, sending the task includes sending 111 the received image 15 of the product 17, which can be described as a purported image of the product since this has not been verified. Sending the task may also include sending 113 the description 18, or a representation of or part of that description, to the nodes 9. In some examples, this also includes sending proof imagery 23 to the node 9.

[0071] Upon receiving 106 the purported image 15 and description 18, each node 9 individual performs an assessment. This can include:

(i) verifying that the purported image 15 of the product 15 matches proof imagery 23 of the product;

(ii) verifying that the description 18 associated with the assessment request 13 matches the purported image 15;

(iii) a quality assessment 30 that the purported image 15 of the product 17 is sufficient to verify the product 17 and the description 18.

[0072] The above examples are not exclusive and other variations and combinations of assessments may be applied.

[0073] To (i) verify the purported image 15 matches the proof imagery 23 of the product 17 can include a comparison of the image. For example, say the insured product 17 is a shoe. The proof imagery 23 may be a photograph of that product 17 obtained from the merchant, payment website, application, platform, marketplace, portal or insurer. It is to be appreciated that in some examples it does not need to be the exact same product item that was shipped to the claimant, but instead be a photograph of the same model of shoe. The proof imagery may also include other related items, such as the box or tags associated with the product.

[0074] In some examples, the verification that the purported product image 15 matches the proof imagery 23 involves input from an operator at the node 9 to "approve" (that includes "pass") or "fail" that verification.

[0075] To (ii) verify that the description 18 matches the purported image 15 of the product 17 involves assessments of that the problem with the product 17 provided in the description 18, such as "broken parts", matches what is shown in the purported image 15. In this example, this includes determining if the purported image 15 includes evidence of broken parts with the product 15. This may, for example, include identifying broken parts to the product. In some examples, verification of the description 18 also involves input from an operator at the node 9 to "approve" or "fail" that verification.

[0076] To (iii) perform a quality assessment 30 that the purported image 15 is sufficient to verify the product 17 and description 18 involves a determination of whether the purported image 15 is of a good enough quality. For example, if the image is blurry or partially obscured. This may be important, for example, if the insurance policy requires an image showing the serial number/barcode of the product 17 and that portion in the purported image is obscured.

[0077] In some examples, this quality assessment may also be aided by an operator at the node to "approve" or "fail" that assessment.

[0078] Fig. 4 illustrates the results of five separate individual node assessments 19', 19", 19'", 19'", 19"' and associated details. The first individual node assessment 19' in this example includes:

• Session token 41, which in the first example is "B0A7579F". The session token 41 is used as an identifier to track each individual response from the node or operator at the node. This can include, information to enable identification of the node, operator, and time of the session. In other examples, this token can be encrypted, or includes a randomised component.

• Processing time 43, which in the first example is "51" seconds. The processing time is the time taken for the node 9 to respond with the individual node assessment 19. IN some examples, this is based on the time the image and description is sent 113 or received 106 and the time the individual node assessment is sent 108 or received 115.

• Work state 45 is the status of the work at the node 9. In thisfirst example, the work state 45' is "approved". Examples of other statuses can include "pending", "processed", or "rejected". The work state 45 may have additional rules that can be customised. For example, the individual node assessment 9 may be rejected if the response was provided in an abnormally short time period compared to other nodes 9. For example, if the processing time was two seconds when all other nodes returned results over thirty seconds or more. This may be indicative of an operator not properly considering the information for verification and assessment. The work state 45 would be pending once the node 9 has receive the task. This can be indicative of the operator (e.g. microworker) that has started work. The work state 45 can be "processed" once the node has performed the individual node assessment 19 and sent the response to server 3. The work state 45 "approved" is used once the system 1 has checked that the work (i.e. the result of the individual node assessment 19) should not be "rejected" based on the processing time (or other prescribed reason).

• Quality assessment 30, which in the first example is "pass". This is the result of the quality assessment that the purported image 15 is of a sufficient quality by the node.

• Product verification 23, which in the first example is "pass". This is the result, by the node 9, of the verification that the purported image 15 of the product matches the proof imagery.

Description verification 25, which in the first example is "pass". This is the result, by the node 9, of the verification that the description 18 for that assessment request 13 matches what is shown in the purported image of the product 17. For example, this may be verifying that a description 18 that there is "broken parts" matches the purported image 15 that shows the broken parts. For the second and fifth individual node assessments 19", 19''', the result of description verification 25", 25'"" 'is "fail". That is, those two nodes returned a result that the description 18 does not match the image 15.

[0079] iii. Combined node assessment and statisticalanalysis

[0080] The separate individual node assessments 19, such as those shown in Fig. 4, are then sent 108 to the server 3 for further assessment. It is to be appreciated that the individual node assessments 19 in some examples many include a subset of data that is shown in Fig. 4. For example, the processing time 43 and work state 45 may be added at the server 3.

[0081] The received 115 individual node assessments 19 form a combined node assessment 27 (represented in Fig. 5). This is a combination of the responses for processing by statistical analysis.

[0082] Statistical analysis is automated based on specified rules. In one example, the specified rules are based on a statistical a majority of the validity response from the nodes 9 that are aggregated in the combined node assessment 27. In this example, there are three "pass" and two "fail" in the description verification 25, which will give a 60% "pass" score.

Since this is above 50%, this will satisfy the specified rules that are based on a statistical majority. Furthermore, the product verification 23 has all five results as "pass" to give 100% (and similarly for quality assessment 30). Thus the statistical assessment (based on majority) will "pass" each category of description verification 25, product verification 23, and quality assessment 30. As a result, the processor 5 will determine 117 a final assessment 27 as "approved".

[0083] It is to be appreciated that the statistical analysis can be based on other rules. In one example, the statistical analysis is based on passing a threshold. For example, this can be %, 80%, or even 100%. The statistical confidence level can be adjusted based on specified requirements. In yet other examples, the threshold can be different for different categories. For example, the quality assessment may be provided with a lower value, say 20%, whilst the product verification is, at a 75% threshold.

[0084] In yet further examples, statistical analysis includes specified rules that are dependent on the results of other verifications. For example, the specified rules may include:

• Determining results that pass the quality assessment 30. This may have a low threshold, such as 60% that must pass the quality assessment or else thefinal assessment is "fail".

• Select the individual node assessments that pass the quality assessment 30, and then subject those selected results for statistical analysis for product verification 23. This product verification may require a higher threshold, such as 80%. If the analysis of those results is less than 80% the final assessment is "fail".

• Then select the results that pass the product verification and perform an analysis of the number that pass the description verification 25. This may include, for example a threshold that is greater than 85% to pass the specified rule. Then based on the result of this last score. Assuming the result is over 85%, the system will give an "approved" results as a final assessment. Otherwise failing this statistical analysis will

provide a "fail" as a final assessment.

[0085] iv. Result offinal assessment andfurther action

[0086] On determining 117 the final assessment 29 to "approve" or "fail" the assessment request 13, a number of action can be initiated by the processor 5 at the server 3.

[0087] In one example, the server 3 can automatically notify the stakeholders of the result that can include sending a notification to the claimant 10 (such as to the first node 11), the insurer (such as sending 121 to the third-party node 31), the merchant, payment, etc. This notification can include an email, text message, or other notification through an API associated with the nodes.

[0088] In some examples, the server 3 authorises automatic payment (or other resolution for claim settlement such as sending a replacement product) to the claimant based on an "approved" final assessment 29. For automatic payment, this can include sending payment directly to a banking details provided by the claimant 10 during the claim submission and notification process. In other alternatives, this may include a refund to a credit card associated with initial purchase of the product.

[0089] In yet another example, the notification may include presenting options to the claimant 10 to select a preferred means to resolve the insurance claim, such as payment or a replacement product.

[0090] In a scenario where the final assessment 29 is "fail", the claimant may be provided with an automatic notification of that fact and an invitation to amend the insurance claim for a second assessment. This may include, for example, requesting a better quality image 15 if the cause of failing the final assessment 29 included failing the quality assessment 30. Similarly, this may include requesting a description with more information, or amended information, if the statistical analysis failed in the description verification 25. Furthermore, this may include asking the claimant to check that they took the image of the correct product (or they selected the right product for making the insurance claim) if the statistical analysis failed due to the result of the product verification 21.

[0091] The result of the final assessment 29 can also be stored 119, along with other information related to the assessment request 13, in the data store 7 associated with the server 3 and/or the third-party node 31.

[0092] Advantages

[0093] As noted above, an advantage of this system and method is to obtain multiple assessments (from the plurality of nodes) that can diversify risks compared to assessment made by a single assessor.

[0094] A further advantage is that using a plurality of distributed nodes 9 can reduce costs. This includes sending the information to distributed nodes 9 that have capacity. The distributed nodes, in some examples, are paid on a per assessment basis. Thus the server may reduce overheads of maintaining the distributed nodes 9 but instead only pay for the service of the distributed nodes 9 as needed. This is in contrast with other systems that may require maintenance of those resources by an insurer - that could lead to waste of resource during idle times. The present system allows easy scaling up or down of microworkers to perform at least part of the functions at the distributed nodes.

[0095] The automated system may also allow insurance companies and other business to scale down the number of employees handling insurance and warranty claims. In some example, the method and system may be used as automated quality control to cross-check manual insurance and warranty claims performed by employees and agents processing such claims.

[0096] The present system and method may be particularly advantageous for insuring low value goods. Processing of claims with traditional employees managing claims have high overheads (such as ~$150 AUD per claim). This can be incredibly problematic where the value of the premium gathered from the insured product is substantially less that the claim processing fee. The system and method can reduce the processing costs by using distributed nodes 9 that typically have a much smaller service cost - that can bring the processing costs below $1 AUD (depending on market rate). This can allow profitability and viability of insurance for low value products, especially in an ecommerce environment.

[0097] Another advantage includes automatic tasks and determinations on the status of an insurance claim. The system provides automated administration and correspondence work required to request information from and send determinations to the insured or payment website, application, platform, marketplace or portal regarding their claims. This is significant as it would otherwise have to be done manually by some claims representative where as in this case all correspondence and claim determination is made automatically based on the response given from the distributed nodes.

[0098] In some examples the system may allow entities other than the consumer to initiate the claims process easily or automatically. For example, if a customer provides feedback, questions, or other queries to the merchant (or payment website, marketplace, portal, or other application and platform, etc.) that indicates an event that can trigger an insurance claim, that entity may initiate the insurance claim on behalf of the customer to improve the process and customer experience

[0099] The end-to-end speed of starting a claim to approving (and paying the claim) using examples of this system may be done within seconds or minutes. This is in stark contrast to the usual turnaround time of insurance claim processes of 1-4 weeks. Furthermore, the ease of the process and the fact that the claimant does not have to interact with humans makes the claim experience so much more enjoyable for the claimant and may help build brand loyalty.

[0100] Variations

[0101] Although the above example has been described with reference to insurance for products, it is to be appreciated that the assessment can be used in applications such as processing warranty claims. Furthermore, this system and method could be applied to any number of other sectors within the insurance industry such as travel, car insurance, home and content insurance, or any other insurances where by data can be shared to tasks to distributed nodes (associated with microworkers) with questions in order to either partially, or completely automate the entire insurance claim process.

[0102] Processing device

[0103] Fig. 7 illustrates an example of a processing device than maybe associated with a first node 11, server 3, distributed node 9, third-party node 31. The server/nodes includes a processor 1510, a memory 1520 and an interface device 1540 that communicate with each other via a bus 1530. The memory 1520 stores instructions and data for implementing one or more of the methods 100 described above, and the processor 1510 performs the instructions (such as a computer program) from the memory 1520 to implement the methods 100. The interface device 1540 may include a communications module that facilitates communication with the communications network 6 and, in some examples, with the user interface and peripherals such as data store 7. It should be noted that although the processing device may be independent network elements, the processing device may also be part of another network element. Further, some functions performed by the processing device may be distributed between multiple network elements. For example, the server 3 may be associated with multiple processing devices and steps of the method 100 may be performed, and distributed, across more than one of these devices.

[0104] It will be appreciated by persons skilled in the art that numerous variations and/or modifications may be made to the above-described embodiments, without departing from the broad general scope of the present disclosure. The present embodiments are, therefore, to be considered in all respects as illustrative and not restrictive.

Claims (18)

CLAIMS:

1. A computational assessment system (1) comprising:

- a server (3) including at least one processor (5), wherein the server (3) is in communication with a plurality of distributed nodes (9); and

wherein the at least one processor (5) is configured to:

- receive (111) an assessment request (13) including:

o a purported image (15) of a product (17) associated with the assessment request (13);

o a description (18) associated with the assessment request (13);

- send (113), to the plurality of distributed nodes (9), the purported image (15) of the product (17) and the description (18), wherein in response each of the distributed nodes (9) performs an individual node assessment (19) indicative of:

o (i) a result of product verification (21) that the purported image (15) of the product (17) matches a proof imagery (23); and

o (ii) a result of description verification (25) that the description (18) associated with the assessment request (13) matches the purported image (15) of the product (17);

- receive (115) a combined node assessment (27) comprising the individual node

assessment (19) from each the plurality of distributed nodes (9);

- determine (117) a final assessment (29) based on statistical analysis of the combined node assessment (27).

2. A computational assessment system (1) according to claim 1, wherein the individual node assessment (19) is further indicative of:

(iii) a result of quality assessment (30) that the purported image (15) of the product (17) is sufficient to verify the product (17) and the description (18).

3. A computational assessment system (1) according to any one of the preceding claims, wherein the at least one processor (5) is further configured to:

- store (119) a result of the final assessment (29) for the assessment request in a data store (7).

4. A computational assessment system (1) according to any one of the preceding claims, wherein the at least one processor (5) is further configured to:

- send (121) a result of the final assessment (29) to a third-party node (31).

5. A computational assessment system (1) according to claim 4, wherein the third-party node (31) is associated with one or more of:

- an insurer;

- ecommerce platform or ecommerce marketplace;

- a retailer;

- payment gateway or providers;

- finance providers.

6. A computational assessment system according to either claims 4 or 5, wherein the assessment request (13) is received from the third-party node (31).

7. A computational assessment system according to any one of claims I to 5, wherein the assessment request (13) is received from a claimant node (11) associated with a customer.

8. A computational assessment system according to any one of the preceding claims, wherein the at least one processor (5) is further configured to:

- send (123) an authorisation for a claim settlement, associated with the assessment request (13) based on the result of the final assessment (29).

9. A computational assessment system according to any one of the preceding claims wherein the statistical assessment comprises:

- determining a statistical majority corresponding to the results of product verification (21) and results of description verification (25) in the combined node assessment (27).

10. A computation assessment system according to claim 9 wherein to determine a final assessment further comprises applying one or more threshold rule.

11. A computational assessment system (1) comprising:

- a plurality of distributed nodes (9);

- a server (3) including at least one processor (5) is configured to:

- receive (111), an assessment request (13) including:

o a purported image (15) of a product (17) associated with the assessment request (13);

o a description (18) associated with the assessment request (13);

- send (113), to the plurality of distributed nodes (9), the purported image (15) of the product (17) and the description (18),

wherein each of the plurality of distributed nodes are configured to:

- verify that the purported image (15) of the product matches a proof imagery (23);

- verify that the description (18) associated with the assessment request (13) matches the purported image (15) of the product;

- send, to the server (3), an individual node assessment (19) based on the results of verification of the purported image (15) and the description (18)

wherein the at least one processor (5) of the server (3) is further configured to:

- receive (115) a combined node assessment (27) comprising the individual node assessment (19) from each the plurality of distributed nodes (9);

- determine (117) a final assessment (29) based on statistical analysis of the combined node assessment (27).

12. A computational assessment system according to claim 11, further comprising:

- a claimant node (11) configured to:

- capture the purported image (15) of the product (17);

- receive the description associated with the assessment request (13);

- send the assessment request (13) to the at least one processor (5) of the server

(3).

13. A computational assessment system according to either claim 11 or 12 further comprising:

- a third party node (31) to:

- receive a result of the final assessment (29) from the server (3); and

- send (125) an authorisation for a claim settlement associated with the assessment request (13) based on the result of the final assessment (29).

14. A method of computational assessment, the method comprising:

- receiving (111) an assessment request (13) including: o a purported image (15) of a product (17) associated with the assessment request (13); o a description (18) associated with the assessment request (13);

- sending (113), to a plurality of distributed nodes (9), the purported image (15) of the product (17) and the description (18), wherein in response each of the distributed nodes (9) performs an individual node assessment (19) indicative of:

o (i) a result of product verification (21) that the purported image (15) of the product (17) matches a proof imagery (23); and

o (ii) a result of description verification (25) that the description (18) associated with the assessment request (13) matches the purported image (15) of the product (17);

- receiving (115) a combined node assessment (27) comprising the individual node assessment (19) from each the plurality of distributed nodes (9);

- determining (117) a final assessment (29) based on statistical analysis of the combined node assessment (27).

15. A method according to claim 14 further comprising:

- sending (123) an authorisation for a claim settlement, associated with the assessment request (13) based on the result of the final assessment (29).

16. A method of computational assessment, the method comprising:

- receiving (111), an assessment request (13) including:

o a purported image (15) of a product (17) associated with the assessment request (13);

" a description (18) associated with the assessment request (13);

- sending (113), to a plurality of distributed nodes (9), the purported image (15) of the product (17) and the description (18),

- verifying, at each of the plurality of the distributed nodes, that the purported image (15) of the product matches a proof imagery (23);

- verifying, at each of the plurality of distributed nodes, that the description (18) associated with the assessment request (13) matches the purported image (15) of the product;

- sending, by each of the plurality of distributed nodes, an individual node assessment (19) based on the results of verification of the purported image (15) and the description (18);

- receiving (115) a combined node assessment (27) comprising the individual node assessment (19) from each the plurality of distributed nodes (9); and

- determining (117) a final assessment (29) based on statistical analysis of the combined node assessment (27).

17. A method according to claim 16 further comprising:

- capturing the purported image (15) of the product (17);

- receiving, from a user interface, the description associated with the assessment request (13).

18. A method according to claim 16 or 17 further comprising:

- sending (125) an authorisation for a claim assessment associated with the assessment request (13) based on the result of the final assessment (29).