DETAILED ACTION
1. The present application, filed on or after March 13, 2013, is being examined under the first inventor to file provisions of the AIA .
This is a regular utility application without a claim of priority.
No IDS has been filed in connection with this Application.
Response to Amendment
2. An RCE with accompanying Amendment was filed June 19, 2026 (hereinafter “Amendment”) and has been entered into the record and fully considered. The Amendment was filed in response to a Final Rejection dated February 20, 2025.
Despite the Amendment to the Claims and Applicant’s remarks, the Rejections under §101 and §103 as set forth in the Non-Final Rejection are hereby maintained.
While advancing prosecution in a helpful manner, the Amendment does not provide the concrete specificity required under §101, nor does it distinguish over the cited combination of references with respect to §103. That said, as noted below, an interview would help advance prosecution in a timely manner.
An explanation of the maintained Rejections and a response to Applicant’s arguments are set forth below. Please see the “Conclusion” section of this Action below for important information regarding responding to this Action.
NOTE: Interviews are welcome at any stage of prosecution. Please use the AIR form, the link for which can be found at the end of this action, to schedule the interview. Despite repeated suggestions, no interview has been scheduled by Applicant.
Status of the Claims:
Claims 1 – 22 are pending in this Application.
Dependent Claims 21 – 22 are NEW; however, they do not change the analysis set forth below. They are addressed separately below.
Independent Claims 1 and 11 are similar in structure and scope and amended in virtually an identical manner.
The dependent Claims were not amended.
Therefore, the following explanation of the maintained rejections with respect to Claim 1 is considered explanatory of the Rejection as a whole.
With regard to the Amendment:
Claim 1 was amended as follows:
PNG
media_image1.png
462
665
media_image1.png
Greyscale
PNG
media_image2.png
702
558
media_image2.png
Greyscale
PNG
media_image3.png
287
636
media_image3.png
Greyscale
New Claims 21 – 22 read as follows:
PNG
media_image4.png
432
681
media_image4.png
Greyscale
Summary of the Amendment and Broadest Reasonable Interpretation:
Claim terminology is to be given its plain and ordinary meaning to a person of ordinary skill in the art, consistent with the specification. This is true, unless the terms are given a special meaning. See MPEP §2111.01
Here, no special meaning is detected. That said, the term “graph-based link structures” may deserve some attention from a broadest reasonable interpretation standpoint.
The specification describes Fig. 13 as depicting “a graphical representation of an alerted transaction.” This figure appears as follows:
PNG
media_image5.png
580
592
media_image5.png
Greyscale
Thus, this term appears – subject to further consideration – to relate to a tree-like structure with nodes, branches, and edges or links which illustrate a relationship between various parties and events pertaining to an alerted transaction. This structure is therefore related to the “link analysis” illustrations shown in Figs. 14 – 15. Therefore, giving this term its broadest reasonable interpretation, it clearly relates to a graphical representation involving nodes and links/edges that connect the various attributes of an alerted transaction. See also 00123.
This would be the common and ordinary meaning of the quoted term. Thus, the claim terms are to be given their broadest reasonable interpretation based on their plain and ordinary meaning.
With regard to §101:
Respectfully, the claim only slightly advances prosecution. The Claim now recites these additional limitations:
A structured query which comprises machine readable fields.
A first, second, and third server
The graph-based link structure discussed above
A natural language response comprising a structured output (e.g. “report”) formatted in a certain way
Respectfully, each of these is a high level, generic concept. Virtually all structured queries and prompt are “machine readable.” That’s why they are structured. It’s the structure that makes them machine readable. This is a very common and well known and generic concept. It is abstract. The number of servers utilized in a system is arbitrary. This is a generic concept as well to a person of ordinary skill in the art. The graph structure is recited at such a high level as to be very generic. These are not specific limitations. These are generic concepts. An LLM – which is recited as used in this claimed system – outputs a natural language output. This is what LLM’s do. Essentially and virtually, all outputs or reports are “structured” at some high level.
These are abstract ideas and a generic concepts. This abstract idea is the same as set forth in previous Actions - organizing human activity; namely, fraud detection.
These recitations do not alter the 101 analysis set forth in the Final Rejection. They painfully emphasize the failings of the Claims to be directed to eligible subject matter.
Moreover, the Claim merely recites very common computerized functions. Computers generically analyze data according to the manner in which they are programmed. Here, no special programming is recited. These limitations are recited at a very high level of generality.
The Claim provides no specificity in terms of how this analysis is accomplished or what algorithms are used or whether they are special or used or applied in any special way. Only the mere outcome or result analyzing relationships is recited. These are common functions. No special functionality is recited. No new computerized components are recited.
These limitations recite results or “outcome” of computer processing without specifying “how” a technical problem is solved. That is, the solution of a technical problem is not reflected in the Claim.
Taking the claim elements separately, the function performed by the computer elements at each step of the process is purely typical of processing identifiers for authentication purposes. Without greater specificity as to “how” certain functions solve a technical problem, the currently recited limitations can be achieved by any general purpose computer without special programming. In short, each step does no more than require a generic computer to perform generic computer functions. Considered as an ordered combination, the computer components of the Claim add nothing that is not already present when the steps are considered separately.
Claim 1 does not, for example, purport to improve the functioning of the computer elements nor does the claim reflect how an improvement in any other technology or technical field is achieved. Thus, Claim 1 amounts to nothing significantly more than instructions to “apply” the abstract idea of detecting fraud using an LLM and machine learning. Such is not sufficient to integrate a practical application in the abstract idea.
Moreover, the new Claims 21 – 22 do not alter this analysis. They relate to “suggested actions” or “analytical suggestions.” These are high level concepts and do not recite a technical solution to a technical problem.
Accordingly, the Rejection is maintained.
With regard to §103:
It is respectfully submitted that the currently applied combination of Cernat in view of Wali renders the Claims obvious, even in view of the Amendment.
In this case, the teachings of Wali appear to be directly on point with the changes recited in the Amendment. Wali is directed to the generation of “structured” reports – namely a SAR – suspicious activity report. In fact, such reports are mandated by federal statute and are highly regulated and therefore structured. See 0002. The structured nature of this output of the system of Wali is illustrated clearly in Fig. 3C:
PNG
media_image6.png
513
707
media_image6.png
Greyscale
This figure is described as follows in the specification:
“[0094] As shown by reference number 376, the server 340 generates a machine-readable SAR. For example, the server 340 may generate the machine-readable SAR based on the server providing the LLM prompt as input to the model. In some implementations, the output of the model is a machine-readable SAR. In other implementations, the output of the model includes data associated with one or more strings of text. For example, the output of the model may include data associated with one or more strings of text, where each string of text corresponds to a field of a SAR report. In this example, the server 340 may generate the machine-readable SAR by including the one or more strings of text in corresponding fields of a SAR (e.g., of a blank or unfilled SAR).” (Emphasis Added)
Thus, it is clear that Wali teaches the newly added limitations relating to a structured fraud assessment output comprising a fraud classification (that which is specified in one or more of the fields of the SAR report) and formatted in accordance with regulatory requirements (e.g. required “fields”). This is clearly a downstream report and relies on external data, i.e. an API. See at least 0113 relating to API calls. No manual intervention is required in Wali. Thus, Wali teaches the use of various “personas” which are well understood to comprise a non-human, structured representation of a user or fictional character or as a digital role.
Wali also teaches the other newly added limitations:
Multiple servers – See at least 0051
A structured query which comprises machine readable fields – See Fig. 3C above. The transaction data is clearly shown in a structured, matrix, or tabular format. See also Fig. 3B which illustrates the feature vectors in a structured, machine readable format:
PNG
media_image7.png
415
752
media_image7.png
Greyscale
Wali teaches that the various features are equivalent to those recited in the Amendment: see at least 0062 – 0065
As to the graph structure, Wali teaches a similar structure to that shown in Applicant’s specification: see at least 0082
Wali clearly teaches the analysis of these relationships. It teaches the use of clustering to allow and facilitate the analysis of relationships between entities and transactions among entities. For example, Fig. 3A illustrates the analysis of relationships between and among various transactions by various users or “clients:”
PNG
media_image8.png
306
735
media_image8.png
Greyscale
In detecting money laundering, Wali clearly teaches analyzing the relationship between the transaction in question (i.e. a “potentially suspicious transaction”) and what appears to be “non-affiliated” transactions:
“[0068] In some implementations, the server may cluster one or more transactions not associated with the user (referred to as non-affiliated transactions). In this example, the server may cluster the one or more non-affiliated transactions using nearest neighbor clustering techniques, K-means clustering techniques, and/or the like. The server may then associate each cluster with a tag corresponding to the non-affiliated transactions in the cluster. For example, the server may associate each cluster with a suspicious transaction tag or a not suspicious transaction tag. In this way, the server may then compare the potentially suspicious transaction to the clustered non-affiliated transactions. The server may then determine whether to proceed with one or more operations described herein based on the server determining that the potentially suspicious transaction is associated with a cluster of non-affiliated transactions having a suspicious or not suspicious transaction tag. In this way, the server can choose to forgo continued processing and generation of a SAR report if the potentially suspicious transaction would unnecessarily cause the generation of the SAR report (e.g., where the potentially suspicious transaction is a false-positive).” (Emphasis Added)
Accordingly, the Rejection must be maintained.
New Claims 21 – 22:
Each of these dependent Claims relate to the generation of a “suggestion” or “suggested action” by the model. In Claim 21, the suggestion is used to updating a ranking of the suggested actions. In Claim 22, each analytical suggestion relates to an attribute of fraud. As taught in Wali, it is well known that a SAR report identifies a type of fraud. That is the purpose of the “narrative” section. (See at least Wali: 0035) It also contains a “mitigation” section which is considered to constitute the recited term “suggestion” or “suggested actions.” Moreover, 0076 – 0079 of Wali teaches the implementation of “feedback” for finalizing the SAR which is considered to constitute the recited term “update” the score which is used in Wali to determine the reporting threshold. (See at least 0072)
Accordingly, these Claims are rejected under §103 based on the same combination of references.
Response to Arguments
3. Applicant's arguments set forth in the Remarks section of the Amendment have been fully considered but they are not persuasive.
With regard to section 101 rejection, Applicant argues as follows:
PNG
media_image9.png
195
682
media_image9.png
Greyscale
The specificity illustrated in these Figures, including Fig. 8, may indeed contain the subject matter which, if recited with specificity, would be likely to constitute eligible subject matter. However, the Claim as it presently stands lacks such specificity. The Examiner is willing, via an interview, to suggest proposed amendments to Applicant which would likely render the Claim eligible. However, at present, the Claim is ineligible and the Rejection is maintained.
With regard to 103, Applicant argues as follows:
PNG
media_image10.png
172
693
media_image10.png
Greyscale
However, it is unclear in the Claim where these limitations are recited. For example, the specific “architecture” and the server-side relationship to the “repositories” are not recited, or at least are unclear from the Claim as presently drafted. Moreover, the alleged “iterative process” is not apparent in the Claim nor is it clear how this process is “tied to each relationship.”
If these features were recited with greater clarity, progress toward compact prosecution could be achieved. As presently broadly recited, the Claim reads on the combination of Cernat in view of Wali, and the teachings of the latter are significant, as explained in details above. That is, Wali clearly teaches customer data (0047) and attribute data (0034 – 0037 and 0062 – 0065).
Thus, the Rejections must be maintained.
Conclusion
4. Applicant should carefully consider the following in connection with this Office Action:
A. Search and Prior Art
The search conducted in connection with this Office Action, as well as any previous Actions, encompassed the inventive concepts as defined in the Applicant’s specification. That is, the search(es) included concepts and features which are defined by the pending claims but also pertinent to significant although unclaimed subject matter. Accordingly, such search(es) were directed to the defined invention as well as the general state of the art, including references which are in the same field of endeavor as the present application as well as related fields (e.g. using LLM prompts to detect money laundering). Indeed, there is a plethora of prior art in these fields.
Therefore, in addition to prior art references cited and applied in connection with this and any previous Office Actions, the following prior art is also made of record but not relied upon in the current rejection:
U.S. Patent Publication No. 2025/0252445 to Gaddam. This reference relates to the concept of an LLM to generate a chatbot for detecting fraud.
U.S. Patent Publication No. 2024/0202687 to Kolchin. This reference relates to the concept of an LLM-based operating system for detecting fraud.
B. Responding to this Office Action
In view of the foregoing explanation of the scope of searches conducted in connection with the examination of this application, in preparing any response to this Action, Applicant is encouraged to carefully review the entire disclosures of the above-cited, unapplied references, as well as any previously cited references. It is likely that one or more such references disclose or suggest features which Applicant may seek to claim. Moreover, for the same reasons, Applicant is encouraged to review the entire disclosures of the references applied in the foregoing rejections and not just the sections mentioned.
C. Interviews and Compact Prosecution
The Office strongly encourages interviews as an important aspect of compact prosecution. Statistics and studies have shown that prosecution can be greatly advanced by way of interviews. Indeed, in many instances, during the course of one or more interviews, the Examiner and Applicant may reach an agreement on eligible and allowable subject matter that is supported by the specification.
Interviews are especially welcomed by this examiner at any stage of the prosecution process. Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool (e.g. TEAMS).
To facilitate the scheduling of an interview, the Examiner requests the use of the AIR form as follows:
USPTO Automated Interview Request http://www.uspto.gov/interviewpractice.
Other forms of interview requests filed in this application may result in a delay in scheduling the interview because of the time required to appear on the Examiner's docket. Thus, the use of the AIR form is strongly encouraged.
E. Communicating with the Office
Any inquiry concerning this communication or earlier communications from the examiner should be directed to WILLIAM BUNKER whose telephone number is (571)272-0017. The examiner can normally be reached on M - F 8:30AM - 5:30PM, Pacific.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Abhishek Vyas, can be reached at 571-270-1836. Information regarding the status of an application, whether published or unpublished, may be obtained from the “Patent Center” system. For more information about the Patent Center system, see https://patentcenter.uspto.gov/
/William (Bill) Bunker/
U.S. Patent Examiner
AU 3691
william.bunker@uspto.gov
(571) 272-0017
July 10, 2026
/ABHISHEK VYAS/Supervisory Patent Examiner, Art Unit 3691
Prosecution Insights