DETAILED ACTION
Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
A request for continued examination under 37 CFR 1.114, including the fee set forth in 37 CFR 1.17(e), was filed in this application after final rejection. Since this application is eligible for continued examination under 37 CFR 1.114, and the fee set forth in 37 CFR 1.17(e) has been timely paid, the finality of the previous Office action has been withdrawn pursuant to 37 CFR 1.114. Applicant's submission filed on December 1, 2025 has been entered.
Status of Claims
Claims 1, 5 - 6, 9, 14 - 15, 19 - 20, 23 and 28 have been amended and are hereby entered.
Claims 8, 12 – 13, 22 and 26 – 27 were cancelled.
Claims 1 - 7, 9 - 11, 14 - 21, 23 - 25 and 28 - 30 are pending and have been examined.
This action is made NON-FINAL.
Response to Arguments
Applicant's arguments filed December 1, 2025 have been fully considered but they are not persuasive.
Regarding to Applicant's arguments against the 101 rejection of pending claims on pages 10 -19: Applicant’s arguments directed to 101 analysis were considered. However, these arguments are not persuasive and the Examiner respectfully disagrees for the following reasons:
For Step 2A-Prong 1 starting in p. 11: Applicant argues that the amended limitations recited in the claims 1 and 15 are not directed to the abstract ideas identified and the Applicant further disagrees by comparing the claimed invention to the Enfish court decision. However, the Examiner find these arguments unpersuasive and respectfully disagrees. Because for this claimed invention, the Examiner identified the abstract idea of a method of organizing human activity recited (e.g. “set forth or described in the claim”; see MPEP 2106.04 (II)(1)) in the specific limitation(s) of establishing “a first chat-based communication session…” and identifying a “predicted user intent” and its corresponding “prediction confidence value” that when the “prediction confidence value” is above a threshold “an agent” is selected to “join the first chat-based communication session based upon the predicted user intent and a complexity value…”, or if the “prediction confidence value” is below the threshold, select the agent based “upon the topic value”. Next, “establish a second chat-based communication session…” and “join the second chat-based communication session to the first chat-based communication session…” for simultaneous connection. Because establishing a first and second chat-based communication sessions to simultaneously connect the user with an agent and an application while identifying the user’s intent and the complex and topic values to provide an agent service encompasses commercial interactions related to business relations by providing customer support for a product or service.
As for the Applicant’s arguments regarding the claims reciting mental processes in p. 14 from Remarks, are found to be unpersuasive and the Examiner disagrees. Because “identifying” user intent based on user’s message, even when “invoking the trained intent classification model with the user message as input” to “select” an agent to “join the chat first chat-based communication session” based on the “user intent and a complexity value” or based on “a topic value” requires evaluation, judgement and opinion. But also, and in light of the USPTO memorandum from August 4, 2025, the particular “way” that these claim limitations are recited and are generally applying (i.e. “invoking”) the AI model of the “intent classification model” can still be read as mental process and does not negate the mental nature of these specific limitation(s). Also, these steps do not further limit or differentiate how the AI model is performing these functions differently from being practically performed in the human mind or with pen and paper. Thus, the claim limitations still recite the abstract idea of a mental process even if they require at least one of: (B) physical aid (e.g. pen and paper) and/or (C) a computer (see MPEP 2106.04(a)(2)(III)(B & C)) to identify predicted user intents associated to user messages and a prediction confidence value to further select an agent to join a communication session.
For Step 2A-Prong 2 and Step 2B starting in p. 15: Applicant argues that “the claimed invention clearly includes features that integrate the claimed invention into a practical application” because the combination of claim 1 limitations “reflect the substantive improvements to the computing system” and the Applicant disclosure in ¶0042 sets forth “the specific training of the intent classification model [which] realizes a more accurate intent prediction”. However, the Examiner find these arguments unpersuasive and respectfully disagrees. Because the claim limitations recited are invoking the use of a computer that “trains” and “intent classification model” with multiple types of data and “invokes” (i.e. applies) the “trained” model as a tool to perform an abstract idea (see MPEP 2106.04(d)(I) and MPEP 2106.05(f)) for “selecting” an agent to the “first chat-based communication session”, “establish second chat-based communication session”, “join the second and first chat-based sessions” to simultaneously connect the “client computing device, the agent computing device, and the virtual assistant software application” which are the other generic computers used. Thus, not providing an inventive concept at Step 2B. These claims, when compared to the Enfish case (see MPEP 2106.05(a)(I)) as asserted by the Applicant in p.12 from Remarks, does not reflect an improvement to the way the computer is working (i.e. functionality) to specifically achieve the concurrent “multiple independent chat-based communication sessions” for simultaneous and seamless connection. Rather, the claim limitations are recited in a high level of generality by disclosing the end result without providing details on how this alleged “improvement” to the computer functioning and/or to the existing technology of “chat-based communications processing between multiple computing devices” is achieved. Moreover, the specification in ¶0042 – 43 lacks discussion of or generally discuss the prior art and how the invention improved the way the computer identifies a predicted user intent and selects an agent to join the communications based on the confidence value and its threshold in combination with the additional element of the trained AI model that is broadly recited. Thus, the claim limitations are recited in a high level of generality by disclosing the end result without providing details on how this alleged “improvement” to the computer functioning and/or to the existing technology of “chat-based communications processing between multiple computing devices” is achieved.
For the same reasons stated above, the Examiner finds Applicant’s arguments related to Step 2B for the claimed invention that further compares it to federal court decisions, in pp. 18 – 19 from Remarks, are unpersuasive and the Examiner respectfully disagrees. Because the claims are not reciting a “specific technical improvement to computing technology”, but rather invoking computers being applied with machine learning model that is used as a tool while being recited in a high level of generality (see MPEP 2106.04(d)(I) and MPEP 2106.05(f)). Thus, not providing an inventive concept at Step 2B. Therefore, the Examiner respectfully disagrees, and maintains 35 USC § 101 rejection for these pending claims.
Regarding to Applicant's arguments against 35 USC § 103 rejection for the amended claims on pages 19 – 22: Applicant’s arguments regarding these amended limitation steps in claims 1 - 7, 9 - 11, 14 - 21, 23 - 25 and 28 - 30 are not persuasive. Firstly, because Applicant's arguments fail to comply with 37 CFR 1.111(b) since they amount to a general allegation that the claims define a patentable invention without specifically pointing out how the language of the claims patentably distinguishes them from the Dunn and Hill references. Also, the Applicant is focusing on each prior art teaching, rather than focusing on the actual language claimed in each claim limitation and how their corresponding limitation steps are different from the prior art teachings while considering the broadest reasonable interpretation (BRI) of the claim.
As for Applicant assertion in pp. 21 and 22 from Remarks, regarding to “one of ordinary skill in the art at the time of invention would not look to combine Dunn, Hill and Booher to arrive at each and every element” of claims 1 and 15. The Examiner find this unpersuasive because the following Graham factual inquiries were resolve and articulated the motivation stated in each 103 rejection: (1) “teaching, suggestion, or motivation” were found in the “references themselves” as a reason to combine reference teachings, (2) a “reasonable expectation of success” was found and/or (3) if additional findings based on the Graham factual inquiries were necessary, in view of the facts of the case under consideration, these were provided to explain a conclusion of obviousness. Moreover, the courts have made clear that the teaching, suggestion, or motivation test is flexible and an explicit suggestion to combine the prior art is not necessary. The motivation to combine may be implicit and may be found in the knowledge of one of ordinary skill in the art, or, in some cases, from the nature of the problem to be solved. An implicit motivation to combine exists not only when a suggestion may be gleaned from the prior art as a whole, but when the ‘improvement’ is technology-independent and the combination of references results in a product or process that is more desirable, for example because it is stronger, cheaper, cleaner, faster, lighter, smaller, more durable, or more efficient. Because the desire to enhance commercial opportunities by improving a product or process is universal—and even common-sensical—we have held that there exists in these situations a motivation to combine prior art references even absent any hint of suggestion in the references themselves (see MPEP 2143 (G)). Therefore, for these reasons stated above, the Examiner respectfully disagrees, and maintains 35 USC § 103 rejection for these pending claims.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1 - 7, 9 - 11, 14 - 21, 23 - 25 and 28 - 30 are rejected under 35 U.S.C. 101 because the claimed invention is directed to an abstract idea without significantly more. The analysis of this claimed invention recited in the claims begins in view of independent claim 1, the most representative claim of the independent claims set 1 and 15, as follows:
At Step 1: Claims 1 – 7, 9 – 11 and 14 falls under statutory category of a system, while claims 15 – 21, 23 – 25 and 28 – 30 are directed to a process.
At Step 2A Prong 1: Claim 1 (representative of claim 15) recites an abstract idea in the following limitations:
establish a first chat-based communication session…;
receive a request from a user… to join an agent to the first chat-based communication session;
capture a topic value and a user message from the first chat-based communication session;
train …on (i) historical user messages exchanged between…, (ii) known intents corresponding to the historical user messages, and (iii) user context data, to predict a user intent for an input user message;
identify a predicted user intent associated with the user message and a prediction confidence value for the predicted user intent by invoking…to the user message;
when the prediction confidence value for the predicted user intent is above a threshold:
select an agent to join the first chat-based communication session based upon the predicted user intent and a complexity value associated with the predicted user intent,
establish a second chat-based communication session…associated with the selected agent, and
join the second chat-based communication session to the first chat-based communication session so that each…are simultaneously connected to each other via the first chat-based communication session; and
when a user intent is not identified:
select an agent to join the first chat-based communication session based upon the topic value,
establish a second chat-based communication session…associated with the selected agent, and
join the second chat-based communication session to the first chat-based communication session so that each…are simultaneously connected to each other via the first chat-based communication session.
when the prediction confidence value for the predicted user intent is below the threshold:
select an agent to join the first chat-based communication session based upon the topic value,
establish a second chat-based communication session…associated with the selected agent, and
join the second chat-based communication session to the first chat-based communication session so that each…are simultaneously connected to each other via the first chat-based communication session.
Generally, these limitations, describe an intent-aware virtual assistant chat routing that identifies and evaluates user’s inquiry data and intent to select and connect the user with an agent in a chat communication session. As disclosed in the specification in ¶1, p.2, this invention allows to “select a qualified agent for participation in a live chat session” and “prioritize agent selection for live chat servicing based upon both agent availability and competence for a given intent in order to make effective chat routing decisions that are predicted to resolve the customer interaction favorably with minimal difficulty.” However, the abstract idea(s) of a certain method of organizing human activity (See MPEP 2106.04(a)(2), subsection II) is recited in claim 1 in the form of “commercial or legal interactions”. Specifically, the abstract idea is recited in the steps of establishing “a first chat-based communication session…” and identifying a “predicted user intent” and its corresponding “prediction confidence value” that when the “prediction confidence value” is above a threshold “an agent” is selected to “join the first chat-based communication session based upon the predicted user intent and a complexity value…”, or if the “prediction confidence value” is below the threshold, select the agent based “upon the topic value”. Next, “establish a second chat-based communication session…” and “join the second chat-based communication session to the first chat-based communication session…” for simultaneous connection. Thus, the abstract idea is recited in these steps because establishing a first and second chat-based communication sessions to simultaneously connect the user with an agent and an application while identifying the user’s intent and the complex and topic values to provide an agent service encompasses commercial interactions related to business relations by providing customer support for a product or service.
Moreover, the steps of “identify a user intent associated with the user message…”, “select an agent to join the first chat-based communication session based upon the user intent and a complexity value…” and “select an agent to join the first chat-based communication session based upon the topic value” fall under the abstract idea of mental processes that can be practically be performed in the human mind or in pen and paper (See MPEP 2106.04(a)(2), subsection III). Because “identifying” user intent based on user’s message as well as “selecting” an agent to join the chat first chat-based communication session based on the user intent and a complexity value or based on a topic value requires evaluation, judgement and opinion.
Step 2A Prong 2: For independent claims 1 and 15, The judicial exception(s) or abstract idea previously identified is not integrated into a practical application (see MPEP 2106.04 (d)). The claims recite the additional element(s) of a server computing device, memory, a processor, (from claim 1); a virtual assistant software application, a client computing device(s), train an intent classification model and an agent computing device (from claims 1 and 15). These additional elements, individually and in combination, and while considering the claims as a whole, are merely used as a tool to perform the abstract idea (See MPEP 2106.05(f)). These recited element features of a computer and Machine Learning (ML) model being trained and then applied (e.g. “invoking the trained intent classification model” as claimed; see ¶0030 from Applicant disclosure) are recited at a high level of generality that these are being used as a tool to perform the generic computer functions for identifying the predicted user intent based on the user’s request and the message received as input as well as user’s historical messages, its known intents and context data to then further select and connect an agent to the first chat-based communication session, establish second chat-based communication session, join the second and first chat-based sessions to simultaneously connect the “client computing device , the agent computing device, and the virtual assistant software application” which are the other generic computers used. Thus, these steps mentioned above are further describing and applying the abstract idea without placing any limits on how the technological components are being improved, while distinguishing in the claim language, the performing limitations from functions that generic computer components can perform.
As for the “establish a first chat-based communication session between a virtual assistant software application of the server computing device and a client computing device”, “receive a request…”, “capture a topic value…”, “select an agent…”, “establish a second chat-based communication session between the virtual assistant software application and an agent computing device…” and “join the second chat-based communication session to the first chat-based communication session…” steps in the claims are really nothing more than links to computer for implementing the use of ordinary capacity for economic or other tasks (e.g., to receive, store, or transmit data) or simply adding a general-purpose computer or computer components (refer to MPEP 2106.05 f (2)).
Step 2B: For independent claims 1 and 15, these claims do not provide an inventive concept. The recited additional elements of the claim(s) are the following: a server computing device, memory, a processor, (from claim 1); a virtual assistant software application, a client computing device(s), train an intent classification model and an agent computing device (from claims 1 and 15). These additional elements are not sufficient to amount significantly more than the judicial exception or abstract idea (see MPEP 2106.05). Because, as indicated in Step 2A Prong 2, these additional element(s) claimed are merely, instructions to “apply” the abstract ideas, which cannot provide an inventive concept. Thus, even when considered in combination, these additional elements represent mere instructions to implement an abstract idea or other exception on a computer, which do not provide an inventive concept at Step 2B.
For dependent claims 2- 7, 9 - 11, 14, 16 – 21, 23 - 25 and 28 – 30, these claims cover or fall under the same abstract idea of a method of organizing human activity and mental processes. They describe additional limitations steps of:
Claims 2- 7, 9 - 11, 14, 16 – 21, 23 - 25 and 28 – 30: further describes the abstract idea of the method for an intent-aware virtual assistant chat routing and further details of the type and processing of user data input and agent candidates’ data while determining confidence levels based on complexity values for the agent selection and user intent thresholds to have the agent concurrently joined to “a plurality of other chat-based communication sessions”. Further, after a period of time the “second chat-based communication session from the first chat-based communication session” is disconnected so that the agent is no longer connected, but the “virtual assistant software application and the client computing device” can still be connected via the “first chat-based communication session”. Thus, being directed to the abstract idea group of “engaging in commercial or legal interactions” to promote business relations by providing customer support for a product or service.
Step 2A Prong 2 and Step 2B: For dependent claims 5, 11, 19 and 25, these claims recite the additional elements of: an agent service window (claims 11 and 25), converting the user message into a machine-readable format suitable for input and executing the trained intent classification model using the converted user message to generate an intent classification (claims 5 and 19). These additional elements recited are invoking computers merely used as a tool to perform or “apply” the abstract idea(s) to the existing process of comparing a current time and determine that no agents are available for the captured topic when the current time is outside the agent service window and to generate a readable format and intent classification, respectively. Thus, amounting to no more than mere instructions to “apply” the exception using a generic computer component (MPEP 2106.05(f) and (f)(2)). Accordingly, for the same reasons stated above, these additional element(s) claimed cannot provide an inventive concept at Step 2B.
Finally, the additional elements previously mentioned above, are nothing more than descriptive language about the elements that define the abstract idea, and these claims remain rejected under 101 as well.
Claim Rejections - 35 USC § 103
In the event the determination of the status of the application as subject to AIA 35 U.S.C. 102 and 103 (or as subject to pre-AIA 35 U.S.C. 102 and 103) is incorrect, any correction of the statutory basis (i.e., changing from AIA to pre-AIA ) for the rejection will not be considered a new ground of rejection if the prior art relied upon, and the rationale supporting the rejection, would be the same under either status.
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claims 1 - 7, 9 - 11, 15 - 21, 23 - 25 and 29 - 30 are rejected under 35 U.S.C. 103 as being unpatentable over Dunn (U.S. Pub No. 20220060580 A1) in view of Hill (U.S. Pub No. 20040162724 A1).
Regarding claims 1 and 15:
Dunn teaches:
a server computing device having a memory that stores computer-executable instructions and a processor that executes the computer-executable instructions to: (In ¶0061; Fig. 1 (150); Fig. 6 (615); Fig. 7 (615 and 750): teaches that “a network device 105, terminal device 115, and/or client device 130 can include, for example, a portable electronic device (e.g., a smart phone, tablet, laptop computer, or smart wearable device) or a non-portable electronic device (e.g., one or more desktop computers, smart appliances, servers, and/or processors)” in which a “Connection management system 150 can be separately housed from network, terminal, IOT and client devices or may be part of one or more such devices (e.g., via installation of an application on a device)”. Refer ¶0157 – 159 for more computer memory details.)
establish a first chat-based communication session between a virtual assistant software application of the server computing device and a client computing device; (In ¶0053 – 54; Fig. 1 (150, 105, 130, 140 and 115); Fig. 6 (605 and 660): teaches that the system’s “connection management system 150 can determine whether any connections are established between network device 105 and an endpoint associated with the client (or remote server 140)” and upon “selecting an endpoint to communicate with network device 105, connection management system 150 can establish connections between the network device 105 and the endpoint”. The “endpoint” can be the “client device 130” of “a company that sells products online” or the “terminal device 115” from the “agents” such as “a sales associate” (see ¶0049). Moreover, the “connection management system 150” can “perform automated actions (e.g., rule-based actions, artificial intelligence originated actions, etc.) based on the live communications” (see ¶0056) and a “software agent or application” (see ¶0062) can be installed in the system wherein such “software agent or application” can also be installed to the other endpoints which is interpreted as a virtual assistant software application. Further, the system can exchange “series of messages or communication exchange being routed between two devices (e.g., a network device and endpoint; see ¶0096” wherein this message can be “natural language communication” (see ¶0098) which is directed to the first chat-based communication session.)
receive a request from a user of the client computing device to join an agent to the first chat-based communication session; (In ¶0130 – 131; Fig. 8 (805 – 810): teaches “At step 805, a communication is received from a user device” including words from natural language wherein “the communication type can be a text messaging or chat application” (see ¶0108). For example, “the communication may state, “I want to speak to a representative” that is parsed and compared to “identified operative words in a database” that may be “related to an action available to a user of the user device” directed to a user request (see ¶0132), in accordance to ¶0036 – 37 applicant specifications.)
capture a topic value and a user message from the first chat-based communication session; (In ¶0101; Fig. 7 (615); Fig. 8 (810 – 815): teaches that the system parses the communication to identify the “operative words” (see ¶0131) and the system’s “intent management engine 615 may assess the (e.g., extracted or received) message” to identify “one or more intents for the message. Examples of intents can include (for example) topic, sentiment, complexity, and urgency” wherein a “topic can include, but it not limited to, a subject, a product, a service, a technical issue, a use question, a complaint, a refund request or a purchase request, etc.” As for an intent, it can be determined “based on a semantic analysis of a message (e.g., by identifying keywords, sentence structures, repeated words, punctuation characters and/or non-article words); user input (e.g., having selected one or more categories); and/or message-associated statistics (e.g., typing speed and/or response latency)” directed to a topic value and a user message, in accordance to ¶2, p. 11 applicant specifications.)
train an intent classification model on (i) historical user messages exchanged between the virtual assistant software application and one or more other client computing devices, (ii) known intents corresponding to the historical user messages, and (iii) user context data, to predict a user intent for an input user message; (In ¶0102; Fig. 7 (615, 730 and 735); Fig. 10E: teaches the use of an AI model or ML model to “a semantic analysis of the message and determine the intent” wherein the “semantic meaning for a particular class can be defined using a training data set and the machine learning model” (directed to known intents and user context). But also, these “messages” or “communications” include content such as additional data of “online history data” (see ¶0051) directed to historical user messages and a “text passage, recording or file based on, for example, an analysis of a received communication (e.g., a semantic or mapping analysis)” directed to user context data (see ¶0058). Refer to ¶0145 – 146 wherein “training data” can be selected to train the “intent model” wherein “annotated data set may be chosen” that further includes “view of the communication, the identified intent, the confidence (i.e., quality of association), and the number of judgments” directed to another example of training an intent classification model on (i) historical user messages, (ii) known intents and (iii)user context data, respectively. But also, the system lets an administrative user (e.g. “client user”) to customize the “intent model” used to identify intents based on “a list of intents” selected to display or enter “training phrases” for a particular intent (see ¶0151).)
identify a predicted user intent associated with the user message and a prediction confidence value for the predicted user intent by invoking the trained intent classification model with the user message as input; (In ¶0132; Fig. 7 (615, 720, 725 and 735); Fig. 8 (815 and 820); Figs 10E, 10F and 13: teaches that the “pre-defined intent may be identified” by “associations between operative words and intents” searched in a database. Further in ¶0145 – 146, the system lets an administrative user (e.g. “client user”) to customize the “intent model” used to identify intents based on “a list of intents” selected to display or enter “training phrases” for a particular intent (see ¶0151). Refer to ¶0102 wherein the system uses a Machine Learning (ML) model and a trained “encoder model” to assess a message, conduct “semantic analysis” and determine the intent as well as converting “the natural language into a vector representation”. As for identifying a prediction confidence value for the predicted user intent, this is directed to the example of “the annotation engine 720 may automatically calculate a quality of 66% between the operative words “pay bill” and the intent “pay current bill”, while a quality of 100% may be assigned to the operative words “pay bill” and the intent “pay bill”” (see ¶0126) and wherein “the annotation engine 720” further applies algorithms to obtain these percentages or numbers also known as “quality of an association” (see ¶0133 for step 820).)
when the prediction confidence value for the predicted user intent is above a threshold: select an agent to join the first chat-based communication session based upon the predicted user intent and a complexity value associated with the predicted user intent, and (In ¶0135; Fig. 8 (825 – 830): teaches “At step 830, an agent profile of the one or more agent profiles may be selected” based on “a correlation of the agent profile to the pre-defined intent and the quality of the association between the communication and the pre-defined intent.” The “agent profile” can be retrieved based on “information associated with an agent having knowledge in a particular intent” and in “difficulty levels” directed to complexity value (see ¶0134) and if “there is 100% confidence that the communication is associated with the pre-defined intent, a certain agent very knowledgeable with that intent may be selected” (i.e. when prediction confidence value for the predicted user intent is above a threshold), in accordance to ¶0045 from Applicant specs. But also, “agent profile may further be selected based on the quality of the association between the communication and the pre-defined intent” wherein it can be in the basis of “a correlation of the agent profile to the pre-defined intent and the quality of the association between the communication and the pre-defined intent” (i.e based on a value that is above a certain threshold) and this correlation can “indicate that the pre-defined intent matches, or is closest to, an intent of which the agent is knowledgeable or has experience, for example”. Refer to ¶0101 wherein the system’s “intent management engine 615 may assess the (e.g., extracted or received) message” by identifying “one or more intents” based on “complexity”.)
when the prediction confidence value for the predicted user intent is below the threshold: select an agent to join the first chat-based communication session based upon the topic value, (In ¶0135; Fig. 8 (825 – 830): teaches “At step 830, an agent profile of the one or more agent profiles may be selected” based on “a correlation of the agent profile to the pre-defined intent and the quality of the association between the communication and the pre-defined intent.” The “agent profile” can be retrieved based on “information associated with an agent having knowledge in a particular intent, category, subject or topic” directed to topic value (see ¶0134) and if “the confidence is relatively low, e.g., 50%, an agent less knowledgeable with that intent may be selected because it is less likely that the correct intent was identified, and the most knowledgeable agent may not be needed”, in accordance to ¶0051 from applicant specs.)
Dunn teaches that communications between “network device 105” and “one or more endpoint(s) 112” can be routed through “connection management system 150” (see ¶0055; Dunn) wherein such communications could be new or modified (see ¶0121; Dunn) through a “continuous channel between two devices” that could either be “stablished, used or terminated” (see ¶0109; Dunn). However, Dunn does not explicitly teach the abilities of establishing a second chat-based communication session specifically between the virtual assistant software application and an agent computing device and join this second chat with the first chat-based communication session to simultaneously connect the client computing device, the agent computing device, and the virtual assistant software application, either when a the prediction confidence value for the predicted user intent is above or below the threshold. However, Hill teaches:
establish a second chat-based communication session between the virtual assistant software application and an agent computing device associated with the selected agent, and (In ¶0065; Fig. 2 (1, 3 and 5); Fig. 3 (2 – 4): teaches that the system “provides three types or levels of conversation management and the system may switch between these during a given conversation” wherein the conversation that can either be in audio or text as “modes of conversation”, interchangeably (see ¶0062) that is directed to the second chat-based communication. Specifically, one of the levels is the “Blended Agent Assist” level or “mode” wherein the system “involves a human agent by presenting him with the customer inquiry and a number of suggested responses ranked by confidence/similarity (“match score”)” which is directed to the second chat-based communication session between the virtual assistant software application and an agent computing device. Refer to ¶0067, ¶0284 and ¶0299 – 300 wherein “confidence thresholds” are considered and adjusted (i.e. directed to the prediction confidence value for the predicted user intent being above or below a threshold) to switch between the “three modes of conversation management”.)
join the second chat-based communication session to the first chat-based communication session so that each of the client computing device, the agent computing device, and the virtual assistant software application are simultaneously connected to each other via the first chat-based communication session. (In ¶0065; Fig. 2 (1, 3 and 5); Fig. 3 (10, 12, 18, 20 and 22 – 16): teaches the joining chat-based communication sessions into one is taught as the “system places a call to the next available agent” while “the customer is waiting, the system connects to an available human agent and plays a whisper of customer B's question” (e.g. the second chat-based communication) in order to select the “an appropriate suggested answer and hits ‘respond,’ enabling the system to complete the interaction” without letting the customer know that “a human agent selected any of the responses”, although, the virtual agent, the human agent and the customer were simultaneously all together in the joined conversation, as illustrated in Fig. 2 for this prior art, and in accordance to ¶0049 from Applicant disclosure. Refer to ¶0070 and ¶0074 for more details. Refer to ¶0090 wherein “software agents may hold a conversation with at least one customer while conversing with a human agent during resolution of a response” wherein such capability “may be extended to have agents talking to customers over multiple channels at once” which reflects the joining of the second with the first chat-based communication sessions.)
It would have been obvious to one of ordinary skill in the art before the earliest effective filing date of the claimed invention to modify Dunn to provide the abilities of establishing a second chat-based communication session specifically between the virtual assistant software application and an agent computing device and join this second chat with the first chat-based communication session to simultaneously connect the client computing device, the agent computing device, and the virtual assistant software application, either when a the prediction confidence value for the predicted user intent is above or below the threshold, as taught by Hill in order to “reduce agent time on a call by enabling him to quickly ‘direct’ the system to the correct resolution” while letting the customer end their conversation “without knowing that a human agent selected any of the responses” (¶0065; Hill).
Regarding claims 2 and 16:
The combination of Dunn and Hill, as shown in the rejection above, discloses the limitations of claims 1 and 15, respectively.
Dunn further teaches:
wherein capturing a topic value comprises: presenting a plurality of selectable topic values in the first chat-based communication session; (In ¶0097: Under the broadest reasonable interpretation (BRI), this limitation is satisfied as the system can connect with a “network device” to present an “app page or webpage” that provides a “particular input command” for an “automatically generated message” that can be based on user inputs such as “button or key presses or recorded speech signals, or speech to text software”. Moreover, the system can receive different requests from the user (e.g. “operative words”) such as “I want to pay my bill” and “I want to speak to a representative” statements (see ¶0124 and ¶0131) and/or receive “user input (e.g., having selected one or more categories)” which are assessed by the system to identify a topic (see ¶0101) as a message which are directed a plurality of selectable topic values.)
detecting user input that corresponds to one of the selectable topic values in the first chat- based communication session; and (In ¶0101; Fig. 8 (805 – 810): teaches that the system’s “intent management engine 615” may assess and identify “the (e.g., extracted or received) message” and its intent including a “topic” by determining “user input (e.g., having selected one or more categories)”.)
capturing the selected topic value from the first chat-based communication session. (In ¶0101; Fig. 7 (615); Fig. 8 (810 – 815): teaches that the system parses the communication to identify the “operative words” (see ¶0131) and the system’s “intent management engine 615 may assess the (e.g., extracted or received) message” to identify “one or more intents for the message” such as a “topic” which include “a subject, a product, a service, a technical issue, a use question, a complaint, a refund request or a purchase request, etc.” As for an intent, it can be determined “based on a semantic analysis of a message (e.g., by identifying keywords, sentence structures, repeated words, punctuation characters and/or non-article words); user input (e.g., having selected one or more categories); and/or message-associated statistics (e.g., typing speed and/or response latency)” directed to a selected topic value.)
Regarding claims 3 and 17:
Dunn, as shown in the rejection above, discloses the limitations of claims 2 and 16, respectively.
Dunn further teaches:
wherein capturing a user message comprises: presenting an information request message in the first chat-based communication session; (In ¶0097: teaches receiving a “message generated based on inputs” (e.g. “an instruction or request”; see ¶0101 for more request details) at a “user interface” which is directed to presenting an information message request in a chat-based communication session.)
detecting user input in response to the information request message in the first chat-based communication session; and capturing the user input as the user message. (In ¶0099: teaches that “a network device may be presenting an app page of a particular client, which may offer an option to transmit a communication to an agent” and upon “receiving user input corresponding to a message, a communication may be generated to include the message and an identifier of the particular client”.)
Regarding claims 4 and 18:
Dunn, as shown in the rejection above, discloses the limitations of claims 3 and 17, respectively.
Dunn further teaches:
wherein the detected user input comprises a chat message or a spoken utterance. (In ¶0098: teaches the received “message can be a natural language communication, whether spoken or typed”.)
Regarding claims 5 and 19:
The combination of Dunn and Hill, as shown in the rejection above, discloses the limitations of claims 1 and 15, respectively.
Dunn further teaches:
wherein identifying a predicted user intent associated with the user message by applying a trained intent classification model to the user message and a prediction confidence value for the predicted user intent comprises: converting the user message into a machine-readable format suitable for input to the trained intent classification model; invoking the trained intent classification model using the converted user message as input to predict an intent classification for the converted user message and generate a prediction confidence value for the predicted intent classification; and (In ¶0101 – 102; Fig. 7 (615, 720, 725 and 735); Fig. 8 (815 and 820): teaches that the “intent management engine 615 may assess the (e.g., extracted or received) message” and includes a ML model that is “configured to conduct a semantic analysis of the message and determine the intent” wherein “training data set” can be used along with the model to define the “semantic meaning for a particular class”. This “classification is based upon semantic similarity” while the “message is encoded by a “encoder model that is trained on data that converts the natural language into a vector representation” which is directed to converting the user message into a machine-readable format as input to execute it in the trained intent classification model. As for the generation of a prediction confidence value for the predicted intent classification, this is interpreted as the example wherein “the annotation engine 720 may automatically calculate a quality of 66% between the operative words “pay bill” and the intent “pay current bill”, while a quality of 100% may be assigned to the operative words “pay bill” and the intent “pay bill”” (see ¶0126) and “the annotation engine 720” applies algorithms to obtain these percentages or numbers (see ¶0133 for step 820). Further refer to ¶0128 wherein the “intent management engine 615” includes an “artificial intelligence engine 735” that can “apply artificial intelligence to the intent model to aggregate intent-related data and draw conclusions about actions that may be taken based on the results and analysis”.)
identifying the predicted user intent based upon the predicted intent classification. (In ¶0102: teaches that that system uses a “machine learning model configured to conduct a semantic analysis of the message and determine the intent”. But also, a “neural network may be used to classify the vectors” of the converted message “into different semantic classes” during the semantic analysis. Refer to ¶0135 wherein at “step 815, a pre-defined intent associated with the one or more operative words may be identified”, for example “pre-defined intent may be identified, in some embodiments, through stored associations between operative words and intents, as stored in a database” which is directed to identifying predicted user intents based on predicted intent classifications.)
Regarding claims 6 and 20:
Dunn, as shown in the rejection above, discloses the limitations of claims 5 and 19, respectively.
Dunn further teaches:
wherein selecting an agent to join the first chat-based communication session based upon the predicted user intent and a complexity value associated with the predicted user intent comprises: identifying one or more agents as candidates for joining the first chat-based communication session based upon the predicted user intent; (In ¶0134 – 135; Fig. 8 (825 - 835): teaches at “step 825, one or more agent profiles are retrieved” and associated based on information of “an agent having knowledge in a particular intent” that is further selected based on the correlation “agent profile to the pre-defined intent and the quality of the association between the communication and the pre-defined intent” at “step 830” to further route the communication for the “terminal device associated with the agent profile” to join a “communication channel” with a user at “step 835” (see ¶0136). Refer to ¶0052 and ¶0059 for general details for identifying “a profile of each of a plurality of agents (or experts or delegates)”, and to ¶0106 for more “endpoint selection” (e.g. “agents”) details.)
determining a confidence level for each of the identified agents using the complexity value; and (In ¶0133; Fig. 8 (820 – 830): the determination of a confidence level per each identified agent using complexity value is satisfied as the system considers the “complexity of a received message,” when determining the most suitable endpoint(s) or agents (see ¶0107) and “at step 820” an algorithm is used by the system to facilitate an “annotation” that “define a quality of an association between the communication and the pre-defined intent” in any suitable form such as by words, percentages or scale numbers directed to determining a confidence level as later refined by the “correlation of the agent profile to the pre-defined intent and the quality of the association between the communication and the pre-defined intent” (see ¶0135), in accordance to ¶2, p. 14 from applicant specs.)
selecting one of the identified agents to join the first chat-based communication session based upon the confidence level. (In ¶0135; Fig. 8 (825 - 835): teaches at “step 830” wherein an “agent profile may be selected based on a correlation of the agent profile to the pre-defined intent and the quality of the association between the communication and the pre-defined intent”. For example, “if there is 100% confidence that the communication is associated with the pre-defined intent, a certain agent very knowledgeable with that intent may be selected.”)
Regarding claims 7 and 21:
Dunn, as shown in the rejection above, discloses the limitations of claims 6 and 20, respectively.
Dunn does not explicitly teach the ability of not selecting any of the identified agents due to low confidence levels falling below a threshold. However, Hill teaches:
wherein the server computing device does not select any of the identified agents when the confidence level for each of the identified agents falls below a threshold. (In ¶0067: under BRI, this conditional limitation is directed to the system adjustment of “the threshold of confidence in interpreting the customer's communication based on how busy the human agents are. This may give customers the option to try automated responses rather than waiting on busy human agents” which implies that no agent was selected due to confidence levels and thresholds.)
It would have been obvious to one of ordinary skill in the art before the earliest effective filing date of the claimed invention to modify Dunn to provide the ability of not selecting any of the identified agents due to low confidence levels falling below a threshold, as taught by Hill as it would be “obvious to try” to measure the confidence of agent’s expertise to only provide an appropriate response coming from an expert agent in the matter requested. Therefore, by not selecting any agent with satisfactory confidence levels, the business can avoid customer dissatisfaction due to an agent’s inaccurate responses, and thus, can provide “good customer service” while “ensuring cost-effective customer service” by utilizing other automatic responses assisted technology that can substitute the agents’ availability shortage while it is resolved (¶0004 and ¶0036; Hill).
Regarding claims 9 and 23:
Dunn, as shown in the rejection above, discloses the limitations of claims 1 and 22, respectively.
Dunn further teaches:
wherein the user context data comprises one or more of: profile information associated with the user of the client computing device, device information associated with the client computing device, or a contact channel associated with the first chat-based communication session. (In ¶0051: teaches that the system’s “connection management system 150 can facilitate strategic routing of communications” that can “include a message with content (e.g., defined based on input from an entity, such as typed or spoken input)” and “additional data, such as data about a transmitting device (e.g., an IP address, account identifier, device type and/or operating system); a destination address; an identifier of a client; an identifier of a webpage or webpage element (e.g., a webpage or webpage element being visited when the communication was generated or otherwise associated with the communication) or online history data; a time (e.g., time of day and/or date); and/or destination address”. Also, the system can identify “account data associated with a particular account identifier or device” directed to profile information associated with the user.)
Regarding claims 10 and 24:
The combination of Dunn and Hill, as shown in the rejection above, discloses the limitations of claims 1 and 15, respectively.
Dunn further teaches:
wherein the server computing device determines an agent availability based upon the captured topic. (In ¶0072 – 73: teaches that the “connection management system 150 can be configured to dynamically, and in real-time, evaluate communications, agent availability, capabilities of terminal devices or agents, and so on, to influence routing determinations”. More specifically, the system can identify “a terminal device 215 selected to participate in a communication exchange with network device 205” which can be identified “based on one or more factors disclosed herein (e.g., availability, matching between a communication's topic/level of detail with agents' or terminal devices' knowledge bases, predicted latency, channel-type availability, and so on).” (see ¶0106 for general details). Refer to ¶0111 – 112 for an example of how the system considers agent’s availability based on weights/scores related to different factors or rules such as “particular communication” matches, previous user satisfaction feedback, immediate availability of a particular agent, “a degree to which an agent (associated with the endpoint) is knowledgeable about a topic in the communication”, etc.)
Regarding claims 11 and 25:
Dunn, as shown in the rejection above, discloses the limitations of claims 10 and 24, respectively.
Dunn further teaches:
wherein determining an agent availability based upon the captured topic comprises: determining an agent service window associated with the captured topic; (In ¶0112: teaches an example wherein the system can evaluate “three endpoints” (e.g. “agents”) for potential communication routing based on a score that “pertains to a match for the particular communication” in which the system further determines that “only the third endpoint is immediately available”. See ¶0106 wherein endpoints’ scores are compared to select a suitable endpoint.)
comparing a current time to the agent service window; and determining that no agents are available for the captured topic when the current time is outside the agent service window. (In ¶0112: teaches that in the example wherein the system can evaluate “three endpoints” (e.g. “agents”) for potential communication routing, a “second endpoint will be available for responding within 15 minutes, but that the first endpoint will not be available for responding until the next day” which is directed to the agent service window being compared with a current time. But also, under BRI this system can determine when no agents are available due to the current time being outside the agent service window.)
Regarding claims 29 and 30:
The combination of Dunn and Hill, as shown in the rejection above, discloses the limitations of claims 1 and 15, respectively.
Dunn teaches that communications between “network device 105” and “one or more endpoint(s) 112” can be routed through “connection management system 150” (see ¶0055; Dunn) wherein such communications could be new or modified (see ¶0121; Dunn) through a “continuous channel between two devices” that could either be “stablished, used or terminated” (see ¶0109; Dunn). However, Dunn does not explicitly teach the ability of disconnecting after a period of time the second chat-based communication session from the first one to disconnect the agent computing device from the original or first conversation between virtual assistant software application and the client computing device. However, Hill further teaches:
further comprising disconnecting, after a period of time, the second chat-based communication session from the first chat-based communication session such that the agent computing device is no longer connected to the virtual assistant software application and the client computing device via the first chat-based communication session, and wherein the virtual assistant software application and the client computing device remain connected to each other via the first chat-based communication session. (In ¶0065; Fig. 2 (1, 3 and 5); Fig. 3 (10, 12, 18, 20 and 22 – 16): teaches an example in the “Blended Agent Assist” mode during a text conversation regarding a “Customer B” asking for an “overnight address” for a “payment of services” wherein the conversation starts with an “automated system” and then a “call” is placed to an available “human agent” to assist the customer. Once the “human assistant” responds to the customer in a seamless manner by selecting “an appropriate suggested answer and hits ‘respond’ (directed to an after period of time), the “system resumes its interaction with customer B by providing an overnight address” and “customer B finishes the call without knowing that a human agent selected any of the responses”. Thus, this disclosure is directed to disconnecting the second chat-based communication session from the first chat-based communication session in order to disconnect the agent computing device and return to the original conversation between the virtual assistant software application and the client computing device.)
It would have been obvious to one of ordinary skill in the art before the earliest effective filing date of the claimed invention to modify Dunn to provide the ability of disconnecting after a period of time the second chat-based communication session from the first one to disconnect the agent computing device from the original or first conversation between virtual assistant software application and the client computing device, as taught by Hill in order to “reduce agent time on a call by enabling him to quickly ‘direct’ the system to the correct resolution” while letting the customer end their conversation “without knowing that a human agent selected any of the responses” (¶0065; Hill).
Claims 14 and 28 are rejected under 35 U.S.C. 103 as being unpatentable over Dunn (U.S. Pub No. 20220060580 A1) in view of Hill (U.S. Pub No. 20040162724 A1) in further view of Booher (U.S. Pub No. 20220207538 A1).
Regarding claims 14 and 28:
The combination of Dunn and Hill, as shown in the rejection above, discloses the limitations of claims 1 and 15, respectively.
Neither Dunn or Hill does not explicitly teach the ability of concurrently connect agent computing devices when complexity value associated with the user intent is below a threshold. However, Booher teaches:
wherein in addition to the first chat-based communication session the agent computing device is concurrently joined to a plurality of other chat-based communication sessions with different client computing devices when a complexity value associated with a predicted user intent of each of the plurality of other chat-based communication sessions is below the threshold. (In ¶0034: this conditional limitation is satisfied and directed to the selection of an agent that can be connected to a “customer device 110” via the “the issue coordination server 120”. The agent is selected based on “information within an agent profile 150, such as an issue capacity 151 of the agent (i.e., indicating a number of simultaneous issues that the agent can handle)”, “a connectivity of the agent 153 (e.g., whether the agent is logged in or otherwise connected to issue coordination server 120)” and “an effectiveness score 156 (e.g., AES) for the agent” (see ¶0032) wherein this process of connecting agents may “be performed concurrently, simultaneously, and/or asynchronously for thousands or even millions of issues and/or conversations at once, and may take real-time measurements of such conversations to levels of detail within a hundredth of a second or less” (see ¶0039), in accordance to ¶0050 from applicant specs. As for the AES score of the agent, includes a complexity factor which is directed to the complexity value being below a threshold (refer to ¶0062 – 65 for more details of the “complexity” criteria and thresholds considered). See ¶0103 for effectiveness scores to indicate the agent capacity to take simultaneous conversations.)
It would have been obvious to one of ordinary skill in the art before the earliest effective filing date of the claimed invention to modify Dunn and Hill to provide the ability of concurrently connect agent computing devices when complexity value associated with the user intent is below a threshold, as taught by Booher in order to “adequately determine the responsiveness of actions taken by a CSA [Customer Service Agent] in addressing the dynamic and diverse range of issues encountered by customers” (¶0004; Booher).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
Beaver (U.S. Pub No. 20200106881 A1) is pertinent because it allows “the human customer service agents to specialize in the instances where human service is preferred, but to scale to the volume of large call centers, systems and methods are provided in which human agents and intelligent virtual assistants (IVAs) co-handle a conversation with a customer. IVAs handle simple or moderate tasks, and human agents are used for those tasks that require or would benefit from human compassion or special handling. Instead of starting the conversation with an IVA and then escalating or passing control of the conversation to a human to complete, the IVAs and human agents work together on a conversation.”
Yeracaris (U.S. Patent No. US 12003668 B1) is pertinent because it “generally relates to the field of human-computer interactions, and more specifically, to virtual assistants that facilitate human-computer interactions through processing of natural languages and other forms of human expression.”
Misiaszek (U.S. Patent No. 11862165 B1) is pertinent because it “relates generally to managing automated communication systems and, more particularly (although not necessarily exclusively), to an optimized virtual assistant for connecting a user to a live agent.”
Clodore (U.S. Pub No. 20230123022 A1) is pertinent because it “relates generally to communication processing using structured frameworks to deliver favorable conversational responses. More specifically, techniques are provided to deploy a framework to assist virtual and/or human agents to provide conversational responses to customers based on the intent communicated by these customers.”
Zhang (U.S. Pub No. 20180054464 A1) is pertinent because it “generally relates to online services. More specifically, the present teaching relates to methods, systems, and programming for virtual agents.”
Erhart (U.S. Pub No. 20210029249 A1) is pertinent because it “provide a method of processing messages received in an asynchronous communication system”
Barak (U.S. Pub No. 20160360466 A1) is pertinent because it “relates generally to intelligent routing of communications. More specifically, techniques are provided to conditionally route communications for a given client to a consistent terminal device.”
Lahav (U.S. Pub No. 20200344353 A1) is pertinent because it “relates generally to facilitating routing of communications. More ally, techniques are provided to route requests to appropriate resources with sufficient capacity in a network, with modeled load capacity and capacity routing.”
Salter (U.S. Pub No. 20220329561 A1) is pertinent because it “relates generally to facilitating routing of communications. More specifically, techniques are provided to dynamically route messages having certain intents between bots and user devices during communication sessions configured with multi-channel capabilities.”
Salter – b (U.S. Pub No. 20220182462 A1) is pertinent because it “relates generally to communication processing using artificial-intelligence (AI). More specifically, techniques are provided to deploy an AI platform to select and manage endpoints in a communication channel, which enables customers to engage with endpoints best suited to answer natural language queries.”
Erhart (U.S. Pub No. 20210029065 A1) is pertinent because it “relate generally to communication methods and specifically to communication methods performed in an asynchronous communication system.”
Cabrera-Cordon (U.S. Patent No. 10554590 B2) is pertinent because it is “a device, method, and computer-readable medium for generating an automated agent enabled to engage in a multi-turn discussion with a user in response to a received request.”
Brophy (U.S. Pub No. 20160117628 A1) is pertinent because it “relates generally to workflow in a customer support environment and, more specifically, to assigning work based on language.”
Srivastava (U.S. Pub No. 20220382997 A1) is pertinent because it “relates generally to a dialogue complexity assessment method, and more particularly but not by way of limitation, to a system, method, and computer program product for determining complexity as a data-driven, context-independent indicator to manage sets of dialogs and services operations.”
George (U.S. Pub No. 20210014356 A1) is pertinent because it “relates generally to systems and methods for connecting nodes on a network for communication and particularly to the timing of automatic connects with a selected communication node.”
Any inquiry concerning this communication or earlier communications from the examiner should be directed to Ivonnemary Rivera Gonzalez whose telephone number is (571)272-6158. The examiner can normally be reached Mon - Fri 9:00AM - 5:30PM.
Examiner interviews are available via telephone, in-person, and video conferencing using a USPTO supplied web-based collaboration tool. To schedule an interview, applicant is encouraged to use the USPTO Automated Interview Request (AIR) at http://www.uspto.gov/interviewpractice.
If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Nathan Uber can be reached at (571) 270-3923. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300.
Information regarding the status of published or unpublished applications may be obtained from Patent Center. Unpublished application information in Patent Center is available to registered users. To file and manage patent submissions in Patent Center, visit: https://patentcenter.uspto.gov. Visit https://www.uspto.gov/patents/apply/patent-center for more information about Patent Center and https://www.uspto.gov/patents/docx for information about filing in DOCX format. For additional questions, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
/IVONNEMARY RIVERA GONZALEZ/Examiner, Art Unit 3626
/NATHAN C UBER/Supervisory Patent Examiner, Art Unit 3626