People.ai, Inc. v. Clari Inc.

Opinion

NORTHERN DISTRICT OF CALIFORNIA

No. C 20-09148 WHA Plaintiff, No. C 21-06314 WHA v.

ORDER RE JUDGMENT ON THE Defendant. PLEADINGS

PEOPLE.AI, INC., Plaintiff, v. CLARI INC., Defendant. In two actions, patent owner has asserted seven total patents against two alleged infringers. Because all seven patents contain the fundamental “do it on a computer” flaw, all seven patents are invalid as ineligible subject matter under Section 101. To the foregoing extent, defendants’ motions for judgment on the pleadings are GRANTED. Patent owner People.ai, Inc. offers business-analytics software that optimize customer business’s relationships and interactions with customers in order to streamline sales and other opportunities. The more data input into the system — and the more accurate that data are — the more helpful the CRM. Both defendants SetSail Technologies, Inc. and Clari Inc. compete in the same burgeoning market as People.ai. People.ai filed suit against SetSail for patent infringement in December 2020. In February 2021, SetSail moved to dismiss under Rule 12(b)(6), prompting People.ai to amend. SetSail again moved to dismiss the first amended complaint, which a June 2021 order granted. During the pendency of People.ai’s motion for leave to file a second amended complaint, the Court of Appeals for the Federal Circuit clarified patent pleading requirements. An order dated August 23, 2021, thus found the most prudent course forward was to permit People.ai’s further amendment. The second amended complaint accuses SetSail of infringing U.S. Patent Nos. 10,496,634; 10,565,229; and 10,657,129. The order granting leave to amend also scheduled a “patent showdown” procedure. But SetSail has skipped over that procedure and argues here that all three patents are ineligible under Section 101 (SetSail Dkt. Nos. 58, 71, 74). People.ai filed suit against Clari in March 2021 in the United States District Court of the District of Delaware. In July 2021, the action was transferred to our district and reassigned to the undersigned in light of the asserted patents overlapping with the SetSail action. The first amended complaint alleges that Clari infringes U.S. Patent Nos. 10,496,634; 10,565,229; and 10,657,129 (all three of which are also asserted against SetSail); as well U.S. Patent Nos. 10,503,783; 10,657,132; 10,872,106; and 10,922,345. Clari quickly moved for judgment on the pleadings on the grounds that all seven patents are ineligible under Section 101 (Clari Dkt. Nos. 21, 33, 53). Six of the seven asserted patents, the ’129, ’106, ’229, ’783, ’634, and ’132 patents, all issued from the same bloc of three provisional applications — Provisional Application Nos. 62/676,187, 62/725,999, and 62/747,452. The ’345 patent issued from the latter two provisional applications. Accordingly, because the closely related patents asserted against the defendants in these actions overlap and defendants’ arguments that the patents are ineligible are substantially similar, this order will jointly address defendants’ motions. 1. THE LEGAL STANDARD. Judgment on the pleadings pursuant to Rule 12(c) is proper when the moving party establishes “on the face of the pleadings that no material issue of fact remains to be resolved and that it is entitled to judgment as a matter of law.” Hal Roach Studios, Inc. v. Richard Feiner & Co., 896 F.2d 1542, 1550 (9th Cir. 1989). Analysis under Rule 12(c) is “substantially identical” to analysis under Rule 12(b)(6). Chavez v. United States, 683 F.3d 1102, 1108 (9th Cir. 2012). District courts must accept all plausible factual allegations in the light most favorable to the non-moving party, but need not “accept as true allegations that are merely conclusory, unwarranted deductions of fact, or unreasonable inferences.” Sprewell v. Golden State Warriors, 266 F.3d 979, 988 (9th Cir. 2001). Patent eligibility can be determined on the pleadings when there are no factual allegations that, taken as true, prevent resolving the eligibility question as a matter of law. Similar to factual allegations, a district court adopts the non-moving party’s claim constructions, and need not engage in a full, formal Markman hearing. Aatrix Software, Inc. v. Green Shades Software, Inc., 882 F.3d 1121, 1124–25 (Fed. Cir. 2018); Data Engine Techs. LLC v. Google LLC, 906 F.3d 999, 1007 (Fed. Cir. 2018). Section 101 provides that 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.” 35 U.S.C. § 101. The implicit exception to Section 101 is that laws of nature, natural phenomena, and abstract ideas are not patentable. Section 101 thus addresses the preemption concerns underlying patent law. In Alice, the Supreme Court fashioned our now-familiar two-step inquiry for Section 101. The district court first evaluates whether the patent claim is directed to an abstract idea. If so, we consider at step two whether the claimed elements recite an inventive concept that transforms the otherwise abstract idea into a patent-eligible invention. Alice Corp. Pty. Ltd. v. CLS Bank Int'l, 573 U.S. 208, 217–18, 221, 223 (2014); Mayo Collab. Servs. v. Prometheus Labs, Inc., 566 U.S. 66, 70 (2012). In more detail, for Alice step one, the Court of Appeals for the Federal Circuit has explained that the district court should consider whether the claims “focus on a specific means or method that improves the relevant technology,” or are instead “directed to a result or effect that itself is the abstract idea and merely invoke generic processes and machinery.” Apple, Inc. v. Ameranth, Inc., 842 F.3d 1229, 1241 (Fed. Cir. 2016); see also McRO, Inc. v. Bandai Namco Games Am. Inc., 837 F.3d 1299, 1314 (Fed. Cir. 2016). For software to be patent eligible, it must go beyond merely organizing existing information into a new form, carrying out a longstanding commercial practice, or otherwise reciting a long prevalent, fundamental practice now accomplished with the benefit of a computer. See Return Mail, Inc. v. U.S. Postal Serv., 868 F.3d 1350, 1368 (Fed. Cir. 2017), reversed and remanded on other grounds, 139 S. Ct. 1853 (2019); Intellectual Ventures I LLC v. Capital One Financial Corp. (Capital One), 850 F.3d 1332, 1340–41 (Fed. Cir. 2017); Intellectual Ventures I LLC v. Symantec Corp. (Symantec), 838 F.3d 1307, 1313–14 (Fed. Cir. 2016). For actions “involving computer- related claims, there may be close calls about how to characterize what the claims are directed to. In such cases, an analysis of whether there are arguably concrete improvements in the recited computer technology could take place under step two.” Enfish, LLC v. Microsoft Corp., 822 F.3d 1327, 1339 (Fed. Cir. 2016). At Alice step two, we ask what else is there in the claim. The second part of the Alice inquiry searches for an inventive concept in an element or combination of elements that is sufficient to ensure that the patent claim, in practice, amounts to significantly more than a patent upon the ineligible abstract concept itself. The recitation of generic computer hardware or generic software structures does not transform an abstract idea into a patent-eligible invention. A patent claim must do more than merely state an abstract idea and say “apply it” or “apply it with a computer.” A non-conventional arrangement of conventional components can, however, give rise to an inventive concept. Alice, 573 U.S. at 217–18, 221–23; Symantec, Mobility LLC, 827 F.3d 1341, 1349–50 (Fed. Cir. 2016). A district court may not rely on “technological details set forth in the patent’s specification and not set forth in the claims to find an inventive concept.” Symantec, 838 F.3d at 1322. While the specification can illuminate the true focus of a claim, unclaimed features are irrelevant to the Alice analysis. ChargePoint, Inc. v. SemaConnect, Inc., 920 F.3d 759, 766, 769 (Fed. Cir. 2019), cert. denied, 140 S. Ct. 983, (2020); Am. Axle & Mfg., Inc. v. Neapco Holdings LLC, 967 F.3d 1285, 1293 (Fed. Cir. 2020). Procedurally, it is not necessary for a district court to address each claim of the patents- in-suit. The Alice analysis may instead focus on representative claims. The parties may agree as to which claims qualify as representative or the district court may select representative claims where the other claims recited in the patent are substantially similar and linked to the same abstract idea, and the patentee does not present any meaningful argument for the distinctive significance of any claim limitations not found in the representative claim. See Berkheimer v. HP Inc., 881 F.3d 1360, 1365 (Fed. Cir. 2018); Content Extraction & Transmission LLC v. Wells Fargo Bank, Nat’l Ass’n, 776 F.3d 1343, 1348 (Fed. Cir. 2014). Upon review, this order finds the claims that People.ai discussed in its claim charts and which the parties focused on in their briefing rank as representative, but considers each of the other claims that People.ai specifically addressed in its briefing and in the hearing. This order proceeds to address each patent in turn. 2. THE ’129 PATENT (ASSERTED AGAINST SETSAIL AND CLARI). First up is the ’129 patent, entitled “Systems and Methods for Matching Electronic Activities to Record Objects of Systems of Record with Node Profiles.” Per Alice step one, defendants contend representative claim 20 is directed to a patent-ineligible abstract concept. People.ai replies that the claim is patent eligible because “it is directed to the automatic use of rules of a particular type, and [defendants have] not provided evidence that the claimed process is the same process previously used” (Opp. I at 14, Clari Dkt. No. 58; see also Opp. II at 18, SetSail Dkt. No. 100). Stripped of excess verbiage, claim 20 discloses a system with “one or more processors” configured to: maintain node profiles; access electronic activities; maintain record objects; “extract data” from the electronic activities; “match the electronic activity to at least one node profile”; “match the electronic activity to at least one record object”; and “store, in a data structure, an association between the electronic activity and the at least one record object.” To match the electronic activity with a record object, the one or more processors are configured to apply a matching policy “based on one or more recipients of the electronic activity and a sender of the electronic activity” (emphasis added). Translating the jargon, the claim consists of generic software objects. “Node profiles” are data profiles that store information on various entities, such as a person’s name and email address (e.g., ’129 patent, col. 2:1–3, fig. 6B; SetSail Dkt. No. 46 at 16). “Record objects” are another type of data profile that store information on various opportunities and accounts (’129 patent, cols. 68:5–16, 68:23–67). A prototypical “electronic activity” is an email or phone call (id. at col. 22:23–28). And “systems of record” are CRM platforms like Salesforce (id. at col. 1:18–37; Opp. I at 1). The claimed system thus matches an electronic activity like an email to an opportunity profile and a business profile by determining that the sender and recipients of the email match information contained in the profiles. People.ai’s own summary aligns: “The claims . . . include specific limitations that define how the matching is carried out. It is carried out by comparing object filed values of a record object to extracted data, and the object field values correspond to senders and recipients of the electronic activity” (Opp. II at 3, emphasis added). Recognizing a claim must not be described at too high a level of abstraction, this order finds claim 20 directed to an abstract idea analogous to “1) collecting data, 2) recognizing certain data within the collected data set, and 3) storing that recognized data in a memory.” Content Extraction, 776 F.3d at 1347; see also Return Mail, 868 F.3d at 1368; Enfish, 822 F.3d at 1337. This order finds the asserted claims of the ’129 patent parallel the activities of a prototypical corporate salesperson, similar to the corporate mailroom analogy in Symantec. correspondence related to business opportunities; the salesperson will “keep business rules defining actions to be taken regarding correspondence based on attributes of the correspondence,” rules such as maintaining contact lists and checking the sender and recipients of the communications; the salesperson will then “apply those business rules to the correspondence,” which would entail matching the incoming communications to particular contacts and particular accounts, and then filing those updates in the correct records. Symantec, 838 F.3d at 1317. People.ai finds this sort of analogy improper because, in its view, to adequately track the claims, the salesperson would need to maintain communications for multiple companies (Opp. I at 15–16). But having a salesperson manage communications from several businesses, rather than just one, does not destroy the analogy. Consulting firm and business-services companies, for example, often handle the communications for multiple businesses. The asserted claims of the ’129 patent do little else than recite a common commercial practice long performed by humans. In fact, the specification explains how the ’129 patent addresses the “challenges of manually entering data” into CRM (Opp. I at 14, citing ’129 patent, col. 1:18–37). And limiting the claims to a particular technological environment, like CRM management, renders them no less abstract. See Capital One, 850 F.3d at 1340; Return Mail, 868 F.3d at 1368; FairWarning IP, LLC v. Iatric Sys., Inc., 839 F.3d 1089, 1094–95 (Fed. Cir. 2016). The claims recite an architecture but: “Although these data structures add a degree of particularity to the claims, the underlying concept embodied by the limitations merely encompasses the abstract idea itself of organizing, displaying, and manipulating data.” Capital One, 850 F.3d at 1341. Indeed, this order finds it difficult to conceptualize associating emails to a business opportunity without considering the sender and recipients, triggering the preemption concern that Alice noted. This order pauses for a moment to address People.ai’s related accusation that SetSail has presented contradictory theories on the scope of the patent claims in this action, which undermines its arguments for judgment on the pleadings (Opp. II at 9–12). People.ai argues [were] specific and People.ai’s allegations need to be highly detailed to honor their limitations.” Now, in contrast, SetSail “takes the contrary position and asks this Court to find the claims to be so broad as to be meaningless” (ibid.). People.ai’s argument here is specious. The order granting SetSail’s motion to dismiss dealt with the adequacy of People.ai’s pleadings. Here, we review the adequacy of People.ai’s patents under Section 101. Turning back to our Alice evaluation, People.ai seeks to align its patent claims with those considered in McRO, where the Federal Circuit found the claims it reviewed patent eligible at Alice step one (Opp. I at 6; Opp. II at 15). But McRO is largely inapposite. First, McRO dealt with patent claims directed to automatically animating facial expressions for animated characters. Here, in contrast, all the asserted claims concern data analytics optimization for CRM. A cascade of other Federal Circuit cases are more apt comparisons. The claims in Symantec concerned filtering methods for emails and other data files. 838 F.3d at 1313. Return Mail addressed claims that recited methods for processing undeliverable mail and relaying mailing address data. 868 F.3d at 1367–68. The claims in Capital One recited systems and methods for editing XML documents by extracting data for user manipulation and then saving the modifications in the underlying XML document. 850 F.3d at 1339–40. Content Extraction considered claims directed to extracting data from documents, recognizing specific information from the data, and storing that data in a memory. 776 F.3d at 1345, 1347. All of these decisions found the claims they reviewed patent- ineligible subject matter under Alice. Our claims are much more comparable to the claims in those actions than the claims in McRO. Second, and relatedly, the claims in McRO employed “unconventional rules” that replaced subjective, artistic actions performed by humans with specific, objective, and mathematical rules executed by computer. McRO, 837 F.3d at 1303, 1313–14. Here, there is simply no artistic, subjective element in the commercial process of recognizing the senders and recipients of emails and associating those emails with various accounts and business profiles. See also FairWarning, 839 F.3d at 1094. Third, People.ai misreads McRO when it argues that its patent claims are not directed to a patent-ineligible abstract idea because they recite specific limitations (e.g., Opp. I at 2, 6–7, 14; Opp. II at 3, 8, 12–13). McRO upheld the claims at Alice step one because they set out “meaningful requirements” and were limited to “rules with specific characteristics.” McRO, 837 F.3d at 1313 (emphasis added). Nowhere in McRO did the Federal Circuit suggest that, without more, a claim directed to a patent-ineligible abstract idea could pass muster because it recites many, detailed claim limitations. That runs afoul of Alice, which warned against making an eligibility determination that depends simply on the draftsman’s art. 573 U.S. at 224 (citing Parker v. Flook, 473 U.S. 584 (1978)). None of the limitations recited in claim 20, despite their numerosity and verbosity, are meaningful requirements. Rather, the claim elements recite only generic, conventional ways of sorting and organizing data. People.ai also cites Koninklijke KPN N.V. v. Gemalto M2M GmbH, 942 F.3d 1143, 1149–50 (Fed. Cir. 2019), which likewise focused on specificity at step one. But Koninklijke emphasized specificity to distinguish patent-eligible improvements in computer capabilities from abstract ideas that merely invoke computers as a tool. A highly specific invocation of a computer as a tool remains ineligible. We consider the technical improvements People.ai proffers for the asserted claims of the ’129 patent at step two. See Enfish, 822 F.3d at 1339. Lastly, the Federal Circuit has certainly not hesitated in finding lengthy, detailed claims patent ineligible. See, e.g., In re Bd. of Trustees of Leland Stanford Junior Univ., 989 F.3d 1367, 1370–72 (Fed. Cir. 2021). In sum, this order finds that the asserted claims of the ’129 patent are directed to an abstract idea. This order proceeds to Alice step two. * * * At Alice step two, defendants contend that the asserted claims of the ’129 patent contain no inventive concept and recite merely generic computer components. People.ai makes two arguments that the claims embody an inventive concept. First, it says the asserted claims of the ’129 patent represent a technical improvement because they are directed to storing associations between electronic activities and record objects separate from the CRM. Second, People.ai states that the claimed invention improves efficiency. This order finds defendants’ arguments persuasive here. We start with People.ai’s centerpiece inventive-concept theory. People.ai contends the system recited in all the asserted claims of the patents-in-suit represent a technical improvement because they claim “the storing of associations between electronic activities and record objects separate from the system of record such as the CRM” (Opp. I at 17; Opp. II at 19). People.ai explains that “storing associations and carrying out matching between electronic activities and record objects outside of the CRM allows for more efficient syncing to the CRM and solves a technological problem related to API request limits imposed by CRMs” (Opp. I at 8; Opp. II at 2). Remember, an inventive concept cannot be concocted from the pleadings or the specification, it must be firmly rooted in the language of the claim. In other words, an unclaimed feature cannot constitute an inventive concept. See, e.g., ChargePoint, 920 F.3d at 766, 769; Am. Axle & Mfg., 967 F.3d at 1293. At first blush, the asserted claims of the ’129 patent do not appear to recite the technical solution that People.ai describes. The ’129 patent describes itself as being directed towards automatically associating electronic activities with record objects, thereby addressing the problem of entering data manually, which can be challenging, time consuming, and error prone (’129 patent at Abstract, Background). People.ai offers three explanations for how the asserted claims capture the technical improvement they have put forward here. First, in its briefing, People.ai explains: The claim recites “one or more processors” that are configured to access electronic activities (such as emails) from electronic accounts “associated with one or more data source providers” and the same one or more processors maintain “record objects of one or more systems of record” (CRM systems) and maintain node profiles and store associations between the electronic activities and record objects. Dkt 21-1 at 199:20-61. Therefore, the “one or more processors” are separate from the processors of the CRM system and the associations between the record objects and the electronic activities are stored separate from the CRM system (Opp. I at 17). This explanation fails to demonstrate a separation between the processors of the CRM and the processors of the claimed system. While claim 20 of the ’129 patent notes the number of processors — “one or more” — it does not define the location or relationship of the claimed system’s processors relative to the processors of the CRM. The claim language only recites that the association between the electronic activity and the record object is stored “in a data structure,” which does not preclude storage of the association in the CRM system itself (’129 patent, col. 199:61). Indeed, despite People.ai generally asserting this theory, claim 11 of the ’345 patent expressly mandates storing the associations in the CRM. More on that issue later. Second, at the hearing, People.ai offered a further explanation for its separation theory. Counsel argued the hook lies in the use of (plural) “systems of record”: So our point in the briefing is, you can’t maintain a record object of one or more systems of record if you are the system of record. You must be a system separate from the system of record in order to maintain record objects of multiple systems of record. So the idea here is the data processing system is maintaining record objects for many different customers, many different systems. And it’s doing this matching, it’s doing this, you know, maintaining the node profiles and performance prediction. And it’s able to push this information back to the data source providers, like we talked about in the briefing, you know, through sort of a bulk sync or through bulk transmissions. (Tr. 15–16). As this order understands it, on the one hand, you have the claimed system that maintains data from systems of record. On the other hand, you have several, individual systems of record — the systems whose data is ingested and maintained by the claimed system. Counsel would have us find a logical division between these two systems; that logical division resulting in associations between electronic activities and record objects being stored separate from the system of record. But the internal logic of counsel’s premise does not track the language of the claim. Counsel acknowledges this theory requires the claimed system maintain record objects from multiple systems of record, otherwise nothing prevents the claimed system from being maintained within the CRM itself. But the claim has no such limitation, and recites a system configured to “maintain a plurality of record objects of one or 1 Third, People.ai presented figure three at the hearing for support of its inventive concept 2 theory, which is a common figure in all seven patents-in-suit: ee ee ee ee eee ee ee eee ee eee ieemenpalnnel 9302 i PERFORMANCE) | = 4 wide SS FEATRIATON He rae we $320 mA 5 yATCHNG ‘Geapa [> PROCESGING) | 6 ! Ls ! 7 : iwoesnow] «= [SGOT | = F-SaDaW)... [S00 : 8 AZ ss to NN ?

v 14 argued the figure demonstrates how the data in the systems of record (9360) goes into the

15 ingestion box (9307) of the data processing system (9300), where electronic activities are then

Q 16 matched to record objects (Tr. 14-15). Counsel pointed to the box demarcating the data

= 17 processing system (9300), arguing defendants’ characterizations “just remove box 9300

. . . . . Z 18 altogether” (Tr. 15). Figure three depicts an embodiment of the invention, the search for an 19 inventive concept focuses on the language of the claim. And as explained, the language of the 20 claim does not preclude storage of the association in the CRM system itself. Because the 21 asserted claims do not require stroing associations separately from the systems of record, the 22 asserted claims do not recite the inventive concept People.ai has proffered. See Am. Axle & 23 Mfg., 967 F.3d at 1293; Interval Licensing LLC v. AOL, Inc., 869 F.3d 1335, 1348 (Fed. Cir. 24 2018). 25 Even if People.ai’s separation theory did make an appearance in the claim language, it 26 still falls short of qualifying as a transformative inventive concept. “Making associations 27 between electronic activities and record objects and storing them separate from the CRM” 28 describes, in substance, implementation of generic computer functionality akin to caching or

processing and storing data on a remote server. This fails to qualify as an inventive concept. See, e.g., Customedia Techs., LLC v. Dish Network Corp., 951 F.3d 1359, 1365–66 (Fed. Cir. 2020); Smart Sys. Innovations, LLC v. Chicago Transit Auth., 873 F.3d 1364, 1374–75 (Fed. Cir. 2017). Indeed, the stated benefits the claimed technical improvement address — avoiding numerous API requests and permitting functionality without being connected to CRM — reflect conventional improvements expected upon application of caching or remote-server functionality. People.ai submits the declaration of Oleg Rogynskyy, its CEO and a named inventor of the seven patents-in-suit. The declaration fleshes out People.ai’s separation theory and its alleged benefits, but at our procedural posture, review is limited to the contents of the pleading. This order accordingly declines to consider Mr. Rogynskyy’s declaration. See Clegg v. Cult Awareness Network, 18 F.3d 752, 754-55 (9th Cir. 1994); FRCP 12(d). In any event, this order questions whether the declaration would provide much support since People.ai has not found a hook for the inventive concept in the language of the claims. Further factual allegations would not seem to rectify this flaw. People.ai next argues that the ’129 patent embodies other, specific improvements, such as “increasing efficiency of the system, reducing resource consumption, and eliminating the need to run multiple searches across databases” (Opp. I at 16). People.ai notes that the specification for the ’129 patent describes “the improvement over manual methods of matching, stating that ‘due to the large volume of heterogenous electronic communications transmitted between devices and the challenges of manually entering data, inputting the information regarding each electronic communication into a system of record can be challenging, time consuming, and error prone’” (Opp. I at 14, citing ’129 patent, col. 1:18–37). People.ai asserts that its allegations regarding efficiency improvements “must be taken as true” and create fact issues that “prevent resolving the subject matter inquiry” (Opp. I at 16; Opp. II at 19). But the law of Federal Circuit is clear on this point: “[T]he improved speed or efficiency inherent with applying the abstract idea on a computer does not provide a sufficient inventive concept.” improvements expected when you incorporate a computer that amalgamates data from a variety of sources, it has not generated a factual dispute. People.ai also argues that claim 12 is specific and thus does “not preempt all automated matching of electronic activities to record objects in CRM systems using rules” (Opp. I at 16). For support, People.ai cites the Federal Circuit’s BASCOM opinion. BASCOM did not hold that the scope of preemption dictated the outcome of the Alice analysis. Instead, it recognized quite the opposite, stating that “simply because some of the claims narrowed the scope of protection through additional ‘conventional’ steps for performing the abstract idea, they did not make those claims any less abstract.” 827 F.3d at 1352. That reasoning applies here. Moreover, the asserted claims have considerable breadth and the limitations they do recite merely employ generic rules for accomplishing the abstract matching process contemplated by the claim. In sum, the elements of representative claim 20 fail to add something more and transform the claim into a patent-eligible invention. The other claims of the ’129 patent People.ai cites in its briefing or referenced at the hearing — claims 1, 11, 19, and 23 — are substantially similar to claim 20 and linked to the same abstract idea. People.ai does not meaningfully distinguish their limitations from those of claim 20. Because the asserted claims of the ’129 patent are directed to an abstract idea and contain no transformative inventive concept, they run afoul of Section 101 and do not qualify as patent-eligible subject matter. 3. THE ’106 PATENT (ASSERTED AGAINST CLARI). The ’106 patent is entitled “Systems and Methods for Matching Electronic Activities Directly to Record Objects of Systems of Record with Node Profiles.” After removing excess jargon, representative claim 19 discloses a system with “one or more processors” configured to: access electronic activities; access record objects stored in systems of record; extract data in an electronic activity; “match the electronic activity to at least one record object . . . based on the extracted data of the electronic activity and object field values” of the record object using a matching policy “based on the one or more recipients or the sender of the electronic People.ai admits “[f]or purposes of this analysis, claim 19 of the ’106 patent is substantially the same as the claim 20 of the ’129 patent” (Opp. I at 18). Upon review, this order agrees, and accordingly finds representative claim 19 of the ’106 patent directed to a patent-ineligible abstract idea for the reasons previously stated for the claims of the ’129 patent. The other claims of the ’106 patent People.ai cites in its briefing or referenced at the hearing — claims 1, 14, and 20 — are substantially similar to representative claim 19 and linked to the same abstract idea. Further, People.ai does not meaningfully distinguish these claims’ limitations from those of claim 19. The asserted claims of the ’106 patent are not patent eligible under Section 101. 4. THE ’229 PATENT (ASSERTED AGAINST SETSAIL AND CLARI). Next up is the ’229 patent, entitled “Systems and Methods for Matching Electronic Activities Directly to Record Objects of Systems of Record.” Defendants argue that the claims of the ’229 patent are ineligible under Section 101 for generally the same reasons they asserted for the claims of the ’129 patent. People.ai acknowledges the similarities between the patents, stating that its eligibility arguments for the ’129 patent claims apply “with equal force” to representative claim 19 of the ’229 patent (Opp. I at 19; Opp. II at 21). Stripped of excess verbiage, claim 19 of the ’229 patent recites a system comprising “one or more processors” configured to: determine with a first policy that includes “one or more filtering rules” that an electronic activity is to be “matched to at least one record object”; identify a “first set of candidate record objects . . . based on . . . one or more recipients” of an electronic activity; identify a “second set of candidate record objects . . . based on the sender of the electronic activity”; and associate an electronic activity with a record object based on “select[ing] at least one candidate record object in both the first . . . and the second set[s] of candidate record objects”; and store the association in a data structure. For Alice step one, People.ai contends that claim 19 of the ’229 patent “is further removed [than the ’129 patent claims] from any alleged human activity or conventional system because it identifies candidate record objects based on a second policy that includes a set of based on senders” (Opp. II at 21–22; see also Opp. I at 19). Using the prototypical examples of the patent terms discussed above reveals that claim 19 merely discloses the matching of an email to a business record by cross-referencing two sets of records: one set generated by rules based on the sender of the email, and a second set generated by a different batch of rules based on the recipients of the email. This reflects a common business practice readily accomplished in the human mind. Turning to our corporate salesperson analogy, the claim is directed to the common practice of discarding junk mail, then filing a relevant communication in the correct business file based on cross-referencing two sets of potential files, one based on the sender and the other on the recipients. The analogy does not break down, as People.ai asserts, just because the salesperson must apply several different sets of rules for filtering and compiling the sets of records that are cross-referenced. This is a long common practice. The claim is directed to an abstract idea. See Symantec, 838 F.3d at 1317–18. People.ai also contends several dependent claims of the ’229 patent provide “further specificity” and recite “specific techniques.” (see Opp. I at 3; Opp. II at 21). As an example, People.ai states that claim 11 recites the additional limitation that a specific object field be used in matching — the relevant team. This additional limitation does not merit a different result. Considering subgroups such as teams and working groups is a common business practice. Claim 11 is also directed to an abstract idea. For Alice step two, People.ai asserts the same inventive concept theory for the claims of the ’229 patent as it did for the claims of the ’129 patent (Opp. I at 19–20; Opp. II at 23). As explained in more detail above, the concept of storing associations between electronic activities and record objects separately from the CRM is not recited in the claims of the ’229 patent, nor is it an inventive concept. In the hearing, counsel stated: “You wouldn't need to identify a system of record . . . if there was only one and you were operating within it” (Tr. 32). But again, the claim recites “one or more systems of record” (emphasis added). The other claims of the ’229 patent People.ai cites in its briefing or referenced at the hearing — claims 6, 7, and 11 — are substantially similar to representative claim 19 and linked limitations from those of claim 19. The asserted claims of the ’229 patent are not patent eligible under Section 101. 5. THE ’783 PATENT (ASSERTED AGAINST CLARI). We next consider the ’783 patent, entitled “Systems and Methods for Generating New Record Objects based on Electronic Activities.” Representative claim 12 of the ’783 patent discloses, after removing excess jargon, a system comprising “one or more processors” configured to: “determine . . . that an electronic activity is to be matched” to a record object; “determine for each candidate record object” a “match score indicating a likelihood of the electronic activity matching the candidate record object” by “comparing the activity field-value pairs to the object-field value pairs” of the candidate record objects; generate a new record object with its type “based on one or more participants of the electronic activity” if the match score does not satisfy a threshold; and store the association “in one or more data structures.” Under Alice step one, Clari contends the claims of the ’783 patent are directed to an abstract idea for the same reasons it asserted for the claims of the ’129 patent. People.ai disagrees, arguing that representative claim 12 is more specific because it “specifically require determining ‘a match score’” (Opp. I at 3–4, 21). To begin, this specificity argument fails for the same reasons previously laid out in the analysis of McRO above. Claim 12 addresses whether to associate a communication with an existing record or to create a new record if it is unlikely any of the existing records are a good match (’783 patent at Abstract). This is a long prevalent, fundamental human practice readily performed by our corporate salesperson. See Return Mail, 868 F.3d at 1368. Considering the match score in more detail, the asserted claims describe the limitation in functional terms. And the specification provides only generic, abstract instructions on the actual calculation of a match score. The specification does not elaborate on the calculation of a match score between an electronic activity and a record object. It does however, provide some detail for a match score between an electronic activity and a node profile, which is still insightful for our purposes: a match score between the electronic activity and a candidate node profile by comparing the strings or values of the electronic activity match corresponding values of the candidate node profile. The match score can be based on a number of fields of the node profile including a value that matches a value or string in the electronic activity. The match score can also be based on different weights applied to different fields. The weights may be based on the uniqueness of values of the field (’783 patent, col. 21:48–58). The specification merely explains that the system should compare the data in the communication with the data in a data profile, and that comparison could weigh different datapoints differently. The specification does not describe how the match score should be constructed. The claim’s disclosure of a “match score” is thus a black box for performing the desired abstract function. In other words, the claim invokes a structure but, in substance, is directed to a particular end result. See Alice, 573 U.S. at 223; Two-Way Medial Ltd. v. Comcast Cable Comms., LLC, 874 F.3d 1329, 1337 (Fed. Cir. 2017); see also Visual Memory LLC v. Nvidia Corp., 867 F.3d 1253, 1263 (Fed. Cir. 2017) (Hughes, J., dissenting). The claim is directed to an abstract idea. Under Alice step two, the parties’ arguments align with those previously discussed. People.ai’s contends that storing associations outside the CRM is an inventive concept here because it “improve[s] the . . . quality and health of the system of record,” and “improve[s] the quality of analytics that can be derived from the system.” (Opp. I at 21). This theory fails for the same reasons previously addressed, and for the additional reason that these highlighted improvements amount to mere efficiency gains. The improvements that come with the incorporation of a computer fail to qualify as an inventive concept. See Symantec, 838 F.3d at 1315. In sum, representative claim 12 of the ’783 patent does not disclose an inventive concept. The other claim of the ’783 patent People.ai cites in its briefing or referenced at the hearing — claim 13 — is substantially similar to representative claim 12 and linked to the same abstract idea. Further, People.ai does not meaningfully distinguish that claim’s limitations from those of claim 12. The asserted claims of the ’783 patent are not patent eligible under Section 101. 6. THE ’345 PATENT (ASSERTED AGAINST CLARI). Up next, the ’345 patent, entitled “Systems and Methods for Filtering Electronic Activities by Parsing Current and Historical Electronic Activities.” Stripped of excess verbiage, representative claim 11 of the ’345 patent recites a system comprising “one or more processors” configured to: identity a first electronic activity and a second electronic activity; parse the first electronic activity to identify the sender or the recipient(s); select one or more filtering policies to apply including at least one of (i) a keyword policy, (ii) a regex pattern policy, or (iii) a logic-based policy; apply the filtering policies “to restrict the first electronic activity from being matched with one or more record objects”; apply the filtering policies and match the second electronic activity to a record object based on a “match policy”; “transmit, to the system of record, instructions to store an association between the second electronic activity and the first record object in the system of record.” The parties make substantially the same arguments for Alice step one as those discussed previously (Opp. I at 4, 22). For the same reasoning outlined above, this order finds claim 11 directed to the abstract idea of data processing by restricting certain data from further analysis based on various sets of generic rules. Our corporate salesperson has long conducted this activity every time she discards the junk mail before updating the business files she maintains with relevant communications. Under Alice step two, we first consider People.ai’s argument that claim 11 is patent eligible because it reduces computing resources and the amount of noise in the CRM systems, and thus improves the operation of the CRM systems (Opp. I at 22–23). See Enfish, 822 F.3d at 1339. Again, all People.ai claims is the beneficial result of increased efficiency that flows from filtering with a computer prior to inputting the data in CRM, not a specific improvement in computing. The filtering policies that the claim recites rank as simple, generic filtering methods, such as looking for certain keywords. This does not qualify as an inventive concept. See Two-Way Media, 874 F.3d at 1337; Symantec, 838 F.3d at 1314. But we have a twist at step two for the ’345 patent. As explained, People.ai contends the “are directed to storing associations outside of the CRM” (Opp. II at 1; see also Opp. I at 1). But unlike the other asserted claims of the other patents-in-suit, claim 11 of the ’345 patent explicitly discloses a system that “transmit[s], to the system of record, instructions to store an association between” the electronic activity and the record object “in the system of record” (’345 patent, col. 193:51–53). People.ai tries to explain away the difference: “By performing the steps locally and then transmitting to the system of record, the invention is able to avoid carrying out the filtering and matching steps with the CRM and it is able to avoid issues with syncing and connection that are outlined” for the other claims. (Opp. I at 23). But this does not follow. First, and foremost, claim 11, similar to the claims of the ’129 patent, recited “one or more processors” and does not define the location or relationship of the claimed system to the CRM. Second, it is unclear how the claimed system avoids issues with syncing and connection without being able to control when the transmission to the CRM takes place. This timing issue would seem to require the claimed system to be able to store the association in the claimed system. But the claim only discloses storage in the CRM. Unclaimed features are irrelevant to the Alice analysis, and the inventive concept People.ai describes is not recited in the language of the claim. See ChargePoint, 920 F.3d at 766, 769; Am. Axle & Mfg., 967 F.3d at 1293. Moreover, as explained above in the analysis of the ’129 patent, even if this order found a hook for People.ai’s proffered inventive concept in the claims of the ’345 patent, the notion of having the filtering and matching taking place in the claimed system and outside the CRM amounts to caching or data processing via a remote server, concepts that the Federal Circuit has found to be generic computer functionality. See, e.g., Customedia, 951 F.3d at 1365–66; Smart Sys. Innovations, 873 F.3d at 1374–75. In short, claim 11 does not contain an inventive concept. The other claim of the ’345 patent People.ai referenced in its briefing or at the hearing — claim 18 — is substantially similar to representative claim 11 and linked to the same abstract idea. Further, People.ai does not meaningfully distinguish that claim’s limitations from those 7. THE ’634 PATENT (ASSERTED AGAINST SETSAIL AND CLARI). We next consider the ’634 patent, entitled “Systems and Methods for Determining a Completion Score of a Record Object from Electronic Activities.” Representative claim 10 of the ’634 patent discloses a system comprising “one or more processors” configured to: select a record object from one or more systems of record; identify “electronic activities . . . associated with the first record object”; determine “at least one participant of each of the . . . electronic activities”; determine, for each participant “at least one of a role, a title, or a department corresponding to the . . . participant”; “determine a completion score indicating a likelihood of completing an event associated with the first record object, the completion score based on the timestamp of each of the plurality of electronic activities and at least one of the role, the title, or the department of the at least one participant of each of the plurality of electronic activities”; and store “in one or more data structures” the association between the record object and the completion score. Under Alice step one, similar to the asserted claims of the ’129 patent, claim 10 of the ’634 patent is directed to a long standing economic practice that is readily performed by a corporate salesperson. Assessing the likelihood that a deal will close based on who the salesperson is negotiating with — e.g., with a junior project manager or with the CEO — and when those communications occurred — e.g., yesterday or two years ago — is an elementary business concept. The specification acknowledges this, stating “enterprises rely on the data included in their systems of records to make projections or predictions on deals” (’634 patent, col. 177:32-33). People.ai disagrees with this assessment, saying that the determination of a completion score is not human conduct. The specification describes a “completion score module” that calculates the completion score, but provides only general, functional guidance on how the module would calculate the score based on the timestamp of the electronic activity, and at least one of the role, title, or the department of the participant of the electronic activity (’634 patent fig. 28, cols. 179:43–180:46). Similar to the “match score” discussed previously, the the likelihood of completing an event. In short, the claim invokes a structure but is merely directed to an end result. See, e.g., Dropbox, Inc. v. Synchronoss Techs., Inc., 815 Fed. App’x 529, 533 (Fed. Cir. 2020). People.ai next reiterates its specificity arguments, stating claim 10 provides a specific way of calculating a completion score (Opp. II at 23). For the same reasons laid out previously this theory misstates the reasoning of McRO. The claims here do not replace artists with rules. Rather, they are directed to applying conventional rules with generic characteristics to ensure efficient data processing (e.g., ’634 patent, col. 177:36–55). The claims of the ’634 patent are directed to an abstract idea. Moving to Alice step two, People.ai argues the claims of the ’634 patent contain an inventive concept because they “are directed to calculation of a completion score based on constantly changing variables that may be aggregated and compared for the purpose of predicting a likelihood of the completion of a certain event” (Opp. II at 25). People.ai further explains that “prior systems did not allow for the prediction of revenue generating events based on real-time data or the use of matched emails to generate completion scores” (Opp. I at 25). In substance, People.ai has simply highlighted the advantages of using a computer to quickly crunch the numbers. The claim only recites generic hardware and software. It does not progress beyond reciting the abstract idea of determine the likelihood of a given event completing and then saying “apply it” with computer software. As this order has emphasized, “merely adding computer functionality to increase the speed or efficiency of the process does not confer patent eligibility on an otherwise abstract idea.” Intellectual Ventures I LLC v. Capital One Bank (Capital One II), 850 F.3d, 1363, 1370 (Fed. Cir. 2015). Additionally, People.ai also reiterates its separation theory: [The ’634 patent’s] concrete improvements include the determination of accurate predictions of revenue generating events, and the presentation of such predictions to business decision makers, making it possible for the enterprise to obtain the benefit of the data within its systems of record without having to access servers on which the electronic activities matched to the record objects are stored (Opp. II at 25, citing SetSail Sec. Amd. Compl. ¶ 70). People.ai reminds we must accept these factual allegations as true (ibid.). But the problem is, as explained in detail in the review of the ’129 patent claims, this inventive concept is not found in the language of the claim. People.ai here focuses on the separation between the claimed system and the system that stores the electronic activities. But claim 10 only discloses “one or more hardware processors,” not the location of the claimed system relative to the system that stores the electronic activities. Moreover, as Clari notes, the claim also recites “identify[ing] a plurality of electronic activities transmitted or received via electronic accounts and associated with the first record object” (’634 patent, col. 196:51–53). This order finds the system architecture asserted by People.ai is not found in the claim, and hence cannot qualify as the claim’s inventive concept. And as explained, even if the separation theory was found in the claim, it does not qualify as an inventive concept. See Interval Licensing, 869 F.3d at 1348; ChargePoint, 920 F.3d at 769. The other claim of the ’634 patent People.ai cites in its briefing or referenced at the hearing — claim 17— is substantially similar to representative claim 10 and linked to the same abstract idea. Claim 17, dependent on claim 10, recites a “stage value,” which serves as another generic input comparable to the recited timestamps (stage values will be discussed further for the ’132 patent). People.ai does not meaningfully distinguish the limitations of claim 17 from those of claim 11. In sum, the asserted claims of the ’634 patent are not patent eligible under Section 101. 8. THE ’132 PATENT (ASSERTED AGAINST CLARI). Finally, the ’132 patent is entitled “Systems and Methods for Forecasting Record Object Completions.” Stripped of draftsmanship, representative claim 12 recites a system comprising “one or more processors” configured to: access data from electronic activities associated with a given record object that includes “a first object field-value pair identifying a stage of a process”; parse the data of the electronic activities; determine the role of the participant based on their seniority, department, or role; determine for the record object “a likelihood that the process . . . will be completed within a predetermined time . . . based on . . . the stage of the structures.” A field-value pair is a standard data entry, such as “associating a value of John to the first name field” (’132 patent, cols. 149:16–21). As for stages of the process, the “stages can include, but are not limited to: prospecting, developing, negotiation, review, closed/won, or closed/lost” (id. at col. 68:35–37). Clari contends that the ’132 patent is directed towards an abstract concept for the same reasons as the ’634 patent. People.ai acknowledges that exemplary claim 12 of the ’132 patent “is similar to claim 10 of the ’634 patent” (Opp. I at 24). Under Alice step one, this order finds that determining the likelihood that a given process will be completed in a predetermined amount of time is a longstanding commercial practice readily performed by our corporate salesperson. A corporate salesperson uses information such as the role of the participant on the other side of the deal (e.g., CEO or junior project manager), the stage of the process (e.g., prospecting or closing), and the information compiled from emails to estimate the likelihood of a given event occurring on a particular timetable. The ’132 patent recognizes that businesses often make these types of predictions and projections (’132 patent, col. 50:13–17). People.ai contends that claim 12 of the ’132 patent “is more specific [than claim 10 of the ’634 patent] and includes additional inputs making it even more removed from activities that might be carried out by a human being” (Opp. I at 24). But for the same reasons laid out above, none of the limitations that claim 12 recites are meaningful or recite a specific improvement in computing. The stage of the process is the new input here, and according to the specification, it can either be defined by the user or determined by the system using a “stage classification engine” (’132 patent, col. 68:29–37). Either way, the stage value is another black box for performing the desired abstract function: either the user makes the determination the same way our corporate salesperson had always evaluated the stage of the deal, or the patent invokes the stage classification engine and applies generic computer functionality to make the determination. To that end, the specification explains: The stage classification engine (325) can be any script, file, program, application, set of instructions, or computer-executable code, that is configured to enable a computing device on which the 1 (id. at col. 68:19-24). The stage value limitation is thus analogous to the “match score” and 2 “completion score” previously addressed. Claim 12 of the ’132 patent is directed to an abstract 3 idea. 4 Under Alice step two, People.ai asserts the same arguments it gave for the claims of the 5 °634 patent. They fail for the same reasons. In addition, the “calculation of a completion score 6 based on constantly changing variables” and “dynamically determining a completion score for 7 a business opportunity” equate to adding computer functionality to increase speed and 8 efficiency (Opp. I at 25). This does not amount to an inventive concept. See Capital One II, 9 792 F.3d at 1370. 10 People.ai did not address any further claims of the °132 patent beyond claim 12 in its 11 briefing or at the hearing. In sum, the asserted claims of the 132 patent are not patent eligible 12 under Section 101. 14 For the reasons stated, the motions for judgment on the pleadings are GRANTED. The 15 asserted claims of the °129, °106, °229, ’783, °345, 634, and the ’132 patents are invalid as a 16 patent ineligible under Section 101. IT IS SO ORDERED. 18 19 Dated: December 13, 2021. Ls Pee 21 ~ WILLIAM ALSUP 22 UNITED STATES DISTRICT JUDGE 23 24 25 26 27 28