Compliance-to-policy detection method and system

• US 6,029,144 A
• Filed: 08/29/1997
• Issued: 02/22/2000
• Est. Priority Date: 08/29/1997
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Ho, Keith Ky Trieu, Barrett, Merwin John, Woisard, Eric, D'Andria, Karen Marie, Gearing, Michael Wayne, Miller, Hope Eleanor, Paradis, Rosemary DeSantis, Cason, Stanley Philip
• Primary Examiner(s): Stamber, Eric W.
• Assistant Examiner(s): Kazimi, Hani M.
• Application Number:
  US08/920,553
• Time in Patent Office: 907 Days
• Field of Search: 705/1, 705/10, 705/11, 705/22, 705/28, 705/30, 235/375, 235/437, 707/101
• US Class Current: 705/30
• CPC Class Codes: G06Q 40/12 Accounting
• CPC Class Codes: G06Q 99/00 Subject matter not provided for in other groups of this subclass

Abstract

According to the present invention, a system and method for checking expense entries for compliance with policy rules and detecting the possibility of fraud is provided. The preferred embodiment includes a policy checker comprising a knowledge based system designed to determine expense entries'"'"' compliance with policy rules and detect a possibility of fraud. The preferred embodiment also includes an auditor workflow system that works with the policy checker to guide manual audits of those expense entries that are not in compliance with the policy rules. The preferred embodiment also includes a data pattern analyzer that detects patterns of behavior that can be indicative of fraud. The preferred embodiment also includes a prioritizer for ranking detected policy violations. These systems work together to provide an automated system for checking expense entries for compliance with company rules, detecting instances where fraud is likely, and generating the appropriate reports.

Claims

1. An apparatus comprising:

at least one processor;

a memory coupled to the at least one processor;

a computer program residing in the memory, said computer program including;

a policy checker, said policy checker comparing an expense entry against at least one of a plurality of policy rules, said policy checker generating a recommendation for changing the expense entry to comply with the at least one of the plurality of policy rules, if the expense entry does not comply with the at least one of the plurality of policy rules; and

an audit workflow program, wherein said policy checker sends the expense entry and the recommendation for changing the expense entry to said audit workflow program if the expense entry fails the at least one of the plurality of policy rules, and wherein said audit workflow program guides a manual audit of the expense entry.

2. The apparatus of claim 1 wherein said policy checker comprises a knowledge based expert system.

3. The apparatus of claim 1 wherein said plurality of policy rules comprise air travel rules, car expense rules and hotel rules.

4. The apparats of claim 1 wherein said audit workflow program guides a manual audit of the expense entry by guiding a human auditor to choose to accept, reject or modify the recommendation for changing the expense entry to comply with the at least one of the plurality of policy rules.

5. The apparatus of claim 1 wherein said plurality of policy rules are organized into a plurality of categories, each category comprising a plurality of high level rules, and wherein said policy checker checks said expense entry against at least one of said plurality of high level rules, sending said expense entry to an audit workflow program when said expense entry fails said at least one of said plurality of high level rules before checking said expense entry against other of said plurality of policy rules in the same category.

6. The apparatus of claim 5 wherein said expense entry is sent back to said policy checker after said expense entry is corrected by a manual audit.

7. The apparatus of claim 1 wherein s policy checker flags said expense entry when said policy checker detects spouse expenses for reimbursements that are taxable.

8. The apparatus of claim 1 further comprising a receipt database, wherein said plurality of policy rules include a check to determine if expense entry data is consistent with receipt data on the receipt database.

9. The apparatus of claim 1 wherein said plurality of policy rules include a check to determine if expense entry data is consistent with credit card data.

10. The apparatus of claim 1 wherein said plurality of policy rules include a check to determine if expense entry data is consistent with air ticket data.

11. The apparatus of claim 1 wherein said plurality of policy rules include a check to determine if expense entry data is consistent with car rental data.

12. The apparatus of claim 1 wherein said plurality of policy rules include a check to determine if a receipt has been submitted for said expense entry above a predetermined dollar amount.

13. A computer-implemented method for checking an expense entry for compliance with company policy, the method comprising the steps of:

a) selecting from a plurality of rules at least one rule that said expense entry should be checked against;

b) checking said expense entry against said at least one selected rule;

c) recommending changes to put said expense entry in compliance with company policy if a violation of said at least one selected rule is found in said expense entry;

d) passing said expense entry and said recommended changes to an auditor system for guiding a manual verification of said expense entry when a violation of said at least one selected rule is found.

14. The method of claim 13 wherein the step of checking said expense entry against said at least one selected rule comprises running said expense entry through a knowledge based expert system.

15. The method of claim 13 wherein the step of checking said expense entry against said at least one selected rule comprises checking said expense entry against at least one car rule when said expense entry includes car expenses.

16. The method of claim 13 wherein the step of checking said expense entry against said at least one selected rule comprises checking said expense entry against at least one hotel rule when said expense entry includes hotel expenses.

17. The method of claim 13 wherein the step of checking said expense entry against said at least one selected rule comprises checking said expense entry against at least one air travel rule when said expense entry includes air travel expenses.

18. The method of claim 13 wherein the step of checking said expense entry against said at least one selected rule comprises checking said expense entry against data from receipts submitted for said expense entry.

19. The method of claim 13 wherein the step of checking said expense entry against said at least one selected rule comprises checking said expense entry against credit card data from a supplying credit card company.

20. The method of claim 13 wherein the step of guiding a manual verification of said expense entry comprises the steps of:

reviewing said recommended changes, and

determining whether to accept, reject or modify said recommended changes.

21. The method of claim 20 wherein said at least one selected rule is a high level rule, and further comprising step of checking said expense entry against at least one lower level rule selected from the plurality of rules after determining whether to accept, reject or modify said recommended changes.

22. The method of claim 13 further comprising the step of comparing said expense entry with a group of related expense entries to detect any patterns of noncompliance that are indicative of fraud.

23. The method of claim 22 wherein the step of comparing said expense entry wtih a group of related expense entries to detect any patterns of noncompliance comprises analyzing said expense entry and said group of related expense entries against a plurality of pattern detecting rules, said pattern detecting rules generated from statistical distributions of past groups of related expense entries.

24. A program product comprising:

(A) a computer program including;

a policy checker, said policy checker comparing an expense entry against at least one of a plurality of policy rules, said policy checker generating a recommendation for changing the expense entry to comply with the at least one of the plurality of policy rules, if the expense entry does not comply with the at least one of the plurality of policy rules; and

an audit workflow program, wherein said policy checker sends the expense entry and the recommendation for changing the expense entry to said audit workflow program if the expense entry fails the at least one of the plurality of policy rules, and wherein said audit workflow system guides a manual audit of said expense entry; and

(B) computer-readable signal bearing media bearing said policy checker.

25. The program product of claim 24 wherein the computer-readable signal bearing media comprises recordable media.

26. The program product of claim 24 wherein the computer-readable signal bearing media comprises transmission media.

27. The program product of claim 24 wherein said policy checker comprises a knowledge based expert system.

28. The program product of claim 24 wherein said plurality of policy rules comprise air travel rules, car expense rules and hotel rules.

29. The program product of claim 24 wherein said plurality of policy rules are organized into a plurality of categories, each category comprising a plurality of high level rules and low level rules, and wherein said policy checker checks said expense entry against at least one of said plurality of high level rules, sending the expense entry to said audit workflow program when said expense entry fails one of said plurality of high level rules before checking said expense entry against at least one of said plurality of low level rules in the same category.

30. The program product of claim 24 wherein said policy checker flags said expense entry when said policy checker detects spouse expenses for reimbursements that are taxable.

31. The program product of claim 24 wherein said plurality of policy rules include a check to determine if expense entry data is consistent with receipt data on a receipt database.

32. The program product of claim 24 wherein said plurality of policy rules include a check to determine if expense entry data is consistent with credit card data.

33. The program product of claim 24 wherein said plurality of policy rules include a check to determine if expense entry data is consistent with air ticket data.

34. The program product of claim 24 wherein said plurality of policy rules include a check to determine if expense entry data is consistent with car rental data.

35. The program product of claim 24 wherein said plurality of policy rules include a check to determine if a receipt has been submitted for said expense entry above a predetermined dollar amount.

36. An apparatus comprising:

at least one processor;

a memory coupled to the at least one processor;

a policy checker residing in said memory, said policy checker including a knowledge based expert system to compare a plurality of expense entries against a plurality of policy rules, said policy checker providing a recommendation for correcting any of said plurality of expense entries that fail at least one of said plurality of policy rules, said recommendation comprising at least one correction that will put said any of said plurality of expense entries in compliance with company policy; and

an audit workflow program in memory, said audit workflow program receiving said any failing expense entries and said recommendation for correcting said any failing expense entries, wherein said audit workflow program directs a manual verification of said recommendation.

37. The apparatus of claim 36 wherein said plurality of policy rules comprise air travel rules, car expense rules and hotel rules.

38. The apparatus of claim 36 wherein said audit workflow program directs an auditor to accept, reject or modify the recommendation for correcting said any failing expense entries during a manual audit.

39. The apparatus of claim 36 wherein said plurality of policy rules comprise a plurality of high level rules and a plurality of low level rules, and wherein said plurality of expense entries are compared against at least one of said plurality of high level rules first, and wherein any of said plurality of expense entries are sent to said audit workflow system when a high level rule failure is determined, and wherein the plurality of expense entries are compared against at least one of the plurality of low level rules only after compliance with said at least one of said plurality of high level rules has been established.

40. The apparatus of claim 36 wherein said policy checker compares said plurality of expense entries that contain spouse expenses for reimbursements that are taxable.

41. The apparatus of claim 36 further including a receipt database, and wherein said plurality of policy rules include a check to determine if expense entry data is consistent with receipt data on said receipt database.

42. The apparatus of claim 36 wherein said plurality of policy rules include a check to determine if expense entry data is consistent with credit card data.

43. The apparatus of claim 36 wherein said plurality of policy rules include a check to determine if expense entry data is consistent with air ticket data.

44. The apparatus of claim 36 wherein said plurality of policy rules include a check to determine if expense entry data is consistent with car rental data.

45. The apparatus of claim 36 wherein said plurality of policy rules include a check to determine if a receipt has been submitted for expense entries above a predetermined dollar amount.

46. The apparatus of claim 36 further comprising a data pattern analyzer having a self organizing map neural network.

47. The apparatus of claim 36 further comprising a data pattern analyzer residing in said memory, said data pattern analyzer analyzing a selected group of the plurality of expense entries to detect any patterns of noncompliance by comparing said selected group of expense entries against a plurality of pattern detecting rules, said pattern detecting rules generated from statistical distributions of past selected groups of expense entries.

48. The apparatus of claim 36 further comprising a prioritizer program residing in said memory, said prioritizer program generating a prioritized list of any of said plurality of expense entries with violations ranked in order of their likelihood of containing fraud.

49. The apparatus of claim 48 wherein said prioritizer program comprises a plurality of backpropagation neural networks.

50. A computer-implemented method for checking a plurality of expense entries for compliance with company policy, the method comprising the steps of:

a) determining which rules selected from a plurality of rules said each of said plurality of expense entries should be checked against, such that any of said plurality of expense entries with car expenses are checked against car rules, any of said plurality of expense entries with hotel expenses are checked against hotel rules, and any of said plurality of expense entries with air travel expenses are checked against air travel rules;

b) checking each of said plurality of expense entries against said selected rules by running said each of said plurality of expense entries through a knowledge based expert system;

c) passing to an auditor system any of said plurality of expense entries that violate any of said selected rules, said passing including a recommendation for action to correct said any of said plurality of expense entries to put said any of said plurality of expense entries in compliance with company policy.

51. The method of claim 50 wherein the step of checking said plurality of expense entries against said selected rules further comprises checking each of said plurality of expense entries against data from receipts submitted for each of said plurality of expense entries.

52. The method of claim 50 wherein the step of checking said plurality of expense entries against said selected rules further comprises checking each of said expense entries against credit card data from a supplying credit card company.

53. The method of claim 50 further comprising the steps of reviewing said recommendation for action, and determining whether to accept, reject or modify said recommendation for action.

54. The method of claim 53 wherein further comprising step of checking said plurality of expense entries against selected lower level rules from the plurality of rules after determining whether to accept, reject or modify said recommendation for action.

55. The method of claim 50 further comprising the step of analyzing a group of expense entries in the plurality of expense entries to detect any patterns of noncompliance that are indicative of fraud.

56. The method of claim 55 wherein the step of analyzing a group of expense entries to detect any patterns of noncompliance comprises comparing said group of expense entries against a plurality of pattern detecting rules, said pattern detecting rules generated from statistical distributions of past groups of expense entries.

57. The method of claim 50 further comprising the step of generating a report, said report prioritizing violations of said selected rules according to the severity of the violations.

58. An apparatus comprising:

at least one processor;

a memory coupled to the at least one processor;

a policy checker residing in said memory, said policy checker including a knowledge based expert system to compare a plurality of expense entries against a plurality of policy rules, said policy checker providing a recommendation for action in regards to any of said plurality of expense entries that fail one or more of said plurality of policy rules;

an audit workflow program in said memory, said audit workflow program receiving any of said plurality of expense entries that have failed one or more of said plurality of policy rules and said recommendation for action, wherein said audit workflow program directs a manual verification of said recommendation for action;

a data pattern analyzer program in said memory, said data pattern analyzer program analyzing a group of expense entries within the plurality of expense entries to detect any patterns of noncompliance by comparing said group of expense entries against a plurality of pattern detecting rules, said pattern detecting rules generated from statistical distributions of past groups of expense entries, said data pattern analyzer including a self organizing map neural network; and

a prioritizer program residing in said memory, said prioritizer program generating a prioritized list of expense entries with violations ranked in order of their likelihood of containing fraud, said prioritizer including at least one back propagation neural network.