Systems and methods for obtaining incident information to reduce fraud

1. A computer-implemented method of analyzing device-submitted documentation, the method comprising:

receiving, by one or more processors, an initial set of documentation associated with a claim filing related to a damaged property, the initial set of documentation including one or more initial images depicting the damaged property and textual information describing the damaged property;

analyzing, by the one or more processors using a machine learning model, the one or more initial images and determining a depicted amount of damage to the damaged property based upon the analysis;

comparing, by the one or more processors using the machine learning model, the depicted amount of damage to the damaged property to the textual information describing the damaged property and calculating an initial likelihood of fraud based upon the comparison;

determining, by the one or more processors using the machine learning model, that additional documentation is needed when the initial likelihood of fraud exceeds an initial threshold level;

in response to determining that the additional documentation is needed, automatically opening, by the one or more processors, a communication channel with a user device;

transmitting in response to the opening of the communication channel, by the one or more processors to the user device via the communication channel at a first time, a set of instructions for capturing an additional set of one or more images further depicting the damaged property via an image sensor of the user device, the set of instructions identifying one or more locations on the damaged property and a capture time limit;

causing a capture time remaining indicator based upon the first time and the capture time limit to be displayed on the user device;

receiving, by the one or more processors from the user device via the communication channel at a second time, the additional set of one or more images captured according to the set of instructions and an image capture time indicating when the additional set of one or more images was captured;

analyzing, by the one or more processors using the machine learning model, the additional set of one or more images and image capture time, and calculating a likelihood of fraud in association with the claim filing based upon the analysis, wherein the analysis includes determining whether the one or more images of the additional set correspond to the one or more locations on the damaged property identified by the instructions, and whether the image capture time is within the capture time limit following the first time;

determining, by the one or more processors using the machine learning model, whether the likelihood of fraud in association with the claim filing exceeds a threshold level;

in response to determining that the likelihood of fraud in association with the claim filing exceeds the threshold level, notifying, by the one or more processors to the user device, that the claim filing is denied; and

in response to determining that the likelihood of fraud in association with the claim filing does not exceed the threshold level, notifying, by the one or more processors to the user device, that the claim filing is approved.

2. The computer-implemented method of claim 1 , further comprising:

outputting, by the one or more processors using the machine learning model, an indication of completeness of the initial set of documentation and

determining, by the one or more processors using the machine learning model, a confidence level based at least in part upon the indication of completeness.

3. The computer-implemented method of claim 1 , further comprising:

using, by the one or more processors, a set of training data to train the machine learning model using a set of training data.

4. The computer-implemented method of claim 1 , wherein the image sensor is configured to capture the additional set of one or more images in accordance with the set of instructions.

5. The computer-implemented method of claim 1 , further comprising:

analyzing, by the one or more processors using the machine learning model, metadata associated with the additional set of one or more images to generate a metadata analysis result; and

determining, by the one or more processors using the machine learning model, the likelihood of fraud based at least in part on the metadata analysis result.

6. A system for analyzing device-submitted documentation, the system comprising:

a memory storing a set of instructions including a machine learning model; and

a processor coupled to the memory, and configured to:

receive an initial set of documentation associated with a claim filing related to a damaged property, the initial set of documentation including one or more initial images depicting the damaged property and textual information describing the damaged property;

analyze, using the machine learning model, the one or more initial images and determining a depicted amount of damage to the damaged property based upon the analysis;

compare, using the machine learning model, the depicted amount of damage to the damaged property to the textual information describing the damaged property and calculating an initial likelihood of fraud based upon the comparison;

determine, using the machine learning model, that additional documentation is needed when the initial likelihood of fraud exceeds an initial threshold level;

in response to determining that the additional documentation is needed, automatically opening a communication channel with a user device;

transmit, in response to the opening of the communication channel, to the user device via the communication channel at a first time, a set of instructions for capturing an additional set of one or more images further depicting the damaged property via an image sensor of the user device, the set of instructions identifying one or more locations on the damaged property and a capture time limit;

causing a capture time remaining indicator based upon the first time and the capture time limit to be displayed on the user device;

receive, from the user device via the communication channel at a second time, the additional set of one or more images captured according to the set of instructions and an image capture time indicating when the additional set of one or more images was captured;

analyze, using the machine learning model, the additional set of one or more images and image capture time, and calculate a likelihood of fraud in association with the claim filing based upon the analysis, wherein the analysis includes determining whether the one or more images of the additional set correspond to the one or more locations on the damaged property identified by the instructions, and whether the image capture time is within the capture time limit following the first time;

determine, using the machine learning model, whether the likelihood of fraud in association with the claim filing exceeds a threshold level;

in response to determining that the likelihood of fraud in association with the claim filing exceeds the threshold level, notifying the user device that the claim filing is denied; and

in response to determining that the likelihood of fraud in association with the claim filing does not exceed the threshold level, notifying the user device that the claim filing is approved.

7. The system of claim 6 , wherein the processor is further configured to:

output, using the machine learning model, an indication of completeness of the initial set of image data documentation; and

determine, using the machine learning model, a confidence level based at least in part upon the indication of completeness.

8. The system of claim 6 , wherein the processor is further configured to:

using a set of training data to train the machine learning model using a set of training data.

9. The system of claim 6 wherein the image capture sensor is configured to capture the additional set of one or more images in accordance with the set of instructions.

10. The system of claim 6 , wherein the processor is further configured to:

analyze, using the machine learning model, metadata associated with the additional set of one or more images to generate a metadata analysis result; and

determine, using the machine learning model, the likelihood of fraud based at least in part on the metadata analysis result.

11. A computer-implemented method of analyzing device-submitted documentation, the method comprising:

receiving, by one or more processors, an initial set of documentation associated with a claim filing related to a damaged property, the initial set of documentation including one or more initial images depicting the damaged property and textual information describing the damaged property;

analyzing, by the one or more processors using a machine learning model, the initial set of documentation and calculating an initial likelihood of fraud based upon the analysis;

determining, by the one or more processors using the machine learning model, that additional documentation is needed when the initial likelihood of fraud exceeds an initial threshold level;

in response to determining that the additional documentation is needed, automatically opening, by the one or more processors, a communication channel with a user device;

generating and transmitting in response to the opening of the communication channel, by the one or more processors to the user device via the communication channel at a first time, a set of instructions for capturing an additional set of one or more images further depicting the damaged property via an image sensor of the user device, the set of instructions identifying one or more locations on the damaged property and a capture time limit;

causing a capture time remaining indicator based upon the first time and the capture time limit to be displayed on the user device;

receiving, by the one or more processors from the user device via the communication channel at a second time, the additional set of one or more images captured according to the set of instructions and an image capture time indicating when the additional set of one or more images was captured;

analyzing, by the one or more processors using the machine learning model, the additional set of one or more images and image capture time, and calculating a likelihood of fraud in association with the claim filing based upon the analysis, wherein the analysis includes determining whether the one or more images of the additional set correspond to the one or more locations on the damaged property identified by the instructions and whether the image capture time is within the capture time limit following the first time;

determining, by the one or more processors using the machine learning model, whether the likelihood of fraud in association with the claim filing exceeds a threshold level;

in response to determining that the likelihood of fraud in association with the claim filing exceeds the threshold level, notifying, by the one or more processors to the user device, that the claim filing is denied; and

in response to determining that the likelihood of fraud in association with the claim filing does not exceed the threshold level, notifying, by the one or more processors to the user device, that the claim filing is approved.

12. The computer-implemented method of claim 11 , further comprising:

outputting, by the one or more processors using the machine learning model, an indication of completeness of the initial set of documentation; and

determining, by the one or more processors using the machine learning model, a confidence level based at least in part upon the indication of completeness.

13. The computer-implemented method of claim 11 , wherein the image sensor is configured to capture the additional set of image data in accordance with the set of instructions.

14. The computer-implemented method of claim 11 , further comprising:

analyzing, by the one or more processors using the machine learning model, metadata associated with the additional set of image data to generate a metadata analysis result; and

determining, by the one or more processors using the machine learning model, the likelihood of fraud based at least in part on the metadata analysis result.