Opinion
2022-1366 2022-1364
04-07-2023
PEOPLE.AI, INC., Plaintiff-Appellant v. CLARI INC., Defendant-Appellee PEOPLE.AI, INC., Plaintiff-Appellant v. SETSAIL TECHNOLOGIES, INC., Defendant-Appellee
EDWARD R. REINES, Weil, Gotshal &Manges LLP, Redwood Shores, CA, argued for plaintiff-appellant. Also represented by SARAH STERNLIEB, New York, NY; ZACHARY TRIPP, Washington, DC. JONATHAN WEINBERG, King &Spalding LLP, Washington, DC, argued for defendant-appellee SetSail Technologies, Inc. Also represented by PAUL ALESSIO MEZZINA; ALLISON H. ALTERSOHN, New York, NY; DAVID SHANE BRUN, San Francisco, CA. EUGENE NOVIKOV, Morrison &Foerster LLP, San Francisco, CA, argued for defendant-appellee Clari Inc. Also represented by DARALYN JEANNINE DURIE; ANDREW TRELOAR JONES, Washington, DC; RAGHAV KRISHNAPRIYAN, Brussels, Belgium.
This disposition is nonprecedential.
Appeal from the United States District Court for the Northern District of California in No. 3:21-cv-06314-WHA, Judge William H. Alsup.
EDWARD R. REINES, Weil, Gotshal &Manges LLP, Redwood Shores, CA, argued for plaintiff-appellant. Also represented by SARAH STERNLIEB, New York, NY; ZACHARY TRIPP, Washington, DC.
JONATHAN WEINBERG, King &Spalding LLP, Washington, DC, argued for defendant-appellee SetSail Technologies, Inc. Also represented by PAUL ALESSIO MEZZINA; ALLISON H. ALTERSOHN, New York, NY; DAVID SHANE BRUN, San Francisco, CA.
EUGENE NOVIKOV, Morrison &Foerster LLP, San Francisco, CA, argued for defendant-appellee Clari Inc. Also represented by DARALYN JEANNINE DURIE; ANDREW TRELOAR JONES, Washington, DC; RAGHAV KRISHNAPRIYAN, Brussels, Belgium.
Before NEWMAN, CHEN, and CUNNINGHAM, Circuit Judges.
CUNNINGHAM, Circuit Judge.
People.ai, Inc. appeals from the United States District Court for the Northern District of California's grant of judgment on the pleadings under Federal Rule of Civil Procedure 12(c) in favor of Defendants, Clari Inc. and SetSail Technologies, Inc. People.ai, Inc. v. SetSail Techs., Inc., 575 F.Supp.3d 1193 (N.D. Cal. 2021) (Decision). People.ai asserted a total of seven patents against Clari or SetSail. Id. at 1197. The district court held that the asserted claims of all seven patents are invalid under 35 U.S.C. § 101. Id. at 1212. People.ai appeals as to three of the asserted patents, U.S. Patent Nos. 10,922,345, 10,565,229, and 10,657,129. We affirm.
The '345 patent was asserted against Clari. Decision at 1197. The '229 and '129 patents were asserted against both Clari and SetSail. Id.
I. BACKGROUND
People.ai offers business-analytics software to optimize customer relationship management (CRM) systems. Decision at 1197. Those systems track and manage business relationships and interactions with customers and potential customers. Id. For example, CRM systems allow businesses to track customer and account information, sales leads, and communications between salespeople and customers. Appellant's Br. 6. The more data provided to a CRM system, the better the system works. Decision at 1197.
The patents at issue in this appeal are directed to the way data is added to "systems of records," which may be "customer relationship management (CRM) systems, enterprise resource planning (ERP) systems, document management systems, applicant tracking systems, among others." '345 patent col. 50 ll. 14-17, 29-34; '229 patent col. 49 ll. 39-42, 54-59 (same); '129 patent col. 64 ll. 14-17, 29-34 (same); see also Appellant's Br. 1-2 ("People.ai's claims are directed to concrete improvements to existing customer relationship management (CRM) systems, and in particular the use of an objective rules-based approach for using tailored filtering policies to intelligently derive useful business information from emails, meetings, and phone calls, matching that information with customer accounts or sales opportunities, and recording those relationships and activities." (emphasis removed)). The patents explain that, "[t]ypically, these systems of records are manually updated, which can result in multiple issues," stemming from the inherently fallible nature of any process performed manually-the data may be entered late, incorrectly, or not at all, "resulting in systems of records that include outdated, incorrect, or incomplete information." '345 patent col. 50 ll. 17-26; '229 patent col. 49 ll. 42-51 (same); '129 patent col. 64 ll. 17-26 (same).
The patents are directed to overcoming these issues with manual data entry. '345 patent col. 50 ll. 29-31; '229 patent col. 49 ll. 54-56 (same); '129 patent col. 64 ll. 29-31 (same). "In particular," the patents "describe[] systems and methods for linking electronic activities," such as "electronic mail, phone calls, [and] calendar events," "to record objects included in one or more systems of record." '345 patent col. 50 ll. 31-36; '229 patent col. 49 ll. 56-61 (same); '129 patent col. 64 ll. 31-36 (same).
A. '345 Patent
The '345 patent is entitled "Systems and Methods for Filtering Electronic Activities by Parsing Current and Historical Electronic Activities." People.ai agreed at oral argument that we could limit our analysis to the patent claims analyzed by the district court. Oral Arg. at 14:55-15:15, https://oralarguments.cafc.uscourts.gov/de-fault.aspx?fl=22-1364_01092023.mp3. Specifically, the district court focused its analysis on claim 11 and briefly addressed claim 18 of the '345 patent. Decision at 1208-09; see also Appellant's Br. 10 n.3.
Claim 11 of the '345 patent recites:
A system comprising:
one or more processors coupled with memory and configured by machine-readable instructions to:
identify a first electronic activity and a second electronic activity
associated with a data source provider that has been transmitted by a sender of the first electronic activity and the second electronic activity and received by one or more recipients of the first electronic activity and the second electronic activity, the first electronic activity and the second electronic activity readable by the one or more recipients;
parse the first electronic activity to identify one or more electronic accounts associated with at least the sender or the one or more recipients of the first electronic activity;
determine, responsive to parsing the first electronic activity, that the first electronic activity is sent from or received by an electronic account of the one or more electronic accounts, the electronic account corresponding to the data source provider;
determine, responsive to parsing the second electronic activity, that the second electronic activity is sent from or received by the electronic account of the one or more electronic accounts;
select, based on the electronic account, one or more filtering policies associated with the data source provider to apply to the first electronic activity and the second electronic activity, the selected one or
more filtering policies including at least one of i) a keyword policy configured to identify electronic activities including a predetermined keyword; ii) a regex pattern policy configured to identify electronic activities including one or more character strings that match a predetermined regex pattern; or iii) a logic-based policy configured to identify electronic activities based on participants of the electronic activities satisfying a predetermined group of participants;
determine, by applying the selected one or more filtering policies to the first electronic activity, to restrict the first electronic activity from being matched with one or more record objects of a system of record of the data source provider based on the first electronic activity satisfying at least one of the selected one or more filtering policies, the system of record of the data source provider including a plurality of record objects;
restrict the first electronic activity from being matched with one or more record objects of the system of record;
determine, by applying the selected one or more filtering policies to the second electronic activity, to match the second electronic activity with one or more record objects of the
system of record of the data source provider based on the second electronic activity not satisfying any of the selected one or more filtering policies;
match, responsive to determining to match the second electronic activity with the one or more record objects, the second electronic activity with a first record object of the one or more record objects responsive to a match policy; and
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.'345 patent col. 192 l. 57-col. 193 l. 53.
Claim 18 depends from claim 11 and adds limitations directed to maintenance and use of "node profiles." '345 patent col. 194 ll. 34-52. Node profiles are "data profiles that store information on various entities, such as a person's name and email address." Appellant's Br. 13 (quoting Decision at 1199).
Claim 18 recites:
The system of claim 11, wherein the one or more processors are further configured to:
maintain a plurality of node profiles corresponding to a plurality of unique entities, each electronic account of the one or more electronic accounts linked to a respective node profile of the plurality of node profiles;
determine, for the first electronic activity, participants of the first electronic activity
based on respective electronic accounts for the participants included in the first electronic activity;
identify, for each of the participants, the respective node profile having an electronic account value for an electronic account field of the node profile which matches the respective electronic account of the participant included in the first electronic activity; and
apply the one or more filtering policies to the first electronic activity based on extracted field-value pairs from the node profiles for the participants of the first electronic activity.'345 patent col. 194 ll. 34-52.
B. '229 Patent
The '229 Patent is entitled "Systems and Methods for Matching Electronic Activities Directly to Record Objects of Systems of Record." As with the '345 patent, People.ai agreed at oral argument that we could limit our analysis to those claims analyzed by the district court. Oral Arg. at 14:55-15:15. The district court analyzed claim 19 and briefly addressed claims 6, 7, and 11. Decision at 1206-07; see also Appellant's Br. 10 n.3. As People.ai does not raise any arguments on appeal directed to the limitations of claims 6, 7, and 11, we focus our analysis solely on claim 19. See Ballard Med. Prods. v. Allegiance Healthcare Corp., 268 F.3d 1352, 1363 (Fed. Cir. 2001) (declining to consider arguments not briefed on appeal).
Claim 19 of the '229 patent recites:
A system comprising:
one or more processors; and
a memory coupled to the one or more processors, the one or more processors configured to:
access a plurality of electronic activities transmitted or received via electronic accounts of one or more data source providers;
access a plurality of record objects of one or more systems of record, each record object of the plurality of record objects corresponding to a record object type and comprising one or more object fields having one or more object field values, the systems of record corresponding to the one or more data source providers;
identify, an electronic activity of the plurality of electronic activities to match to one or more record objects, the electronic activity of the plurality of electronic activities identifying participants including a sender of the electronic activity and one or more recipients of the electronic activity;
determine a data source provider associated with providing the one or more processors access to the electronic activity;
identify a system of record corresponding to the determined data source provider, the system of record including a plurality of candidate record objects to which to match the electronic activity;
determine, responsive to applying a first policy including one or more filtering rules, that the electronic activity is to be matched to at least one record object of the identified system of record;
in response to determining that the electronic activity is to be matched to at least one record object of the identified system of record,
identify a first set of candidate record objects to which to match the electronic activity responsive to applying a second policy including a first set of rules for identifying one or more record objects of a first record object type based on an object field value of the record object that identifies the one or more recipients;
identify a second set of candidate record objects to which to match the electronic activity responsive to applying the second policy including a second set of rules for identifying candidate record objects based on the sender of the electronic activity, wherein the second policy includes a third set of rules for identifying candidate record objects of a second record object type;
select at least one candidate record object included in both the first set of candidate record objects and the second set of candidate record objects; and
store, in a data structure, an association between the selected at least one candidate record object and the electronic activity.'229 patent col. 144 l. 40-col. 145 l. 25. C. '129 Patent
The '129 patent is entitled "Systems and Methods for Matching Electronic Activities to Record Objects of Systems of Record with Node Profiles." Its claims are directed to the maintenance and use of node profiles. See, e.g., '129 patent col. 195 ll. 22-67. As with the other two patents at issue in this appeal, People.ai agreed at oral argument that we could limit our analysis to those claims analyzed by the district court. Oral Arg. at 14:55-15:15. The district court analyzed claim 20 and briefly addressed claims 1, 11, 12, 19, and 23. Decision at 1199-1205; see also Appellant's Br. 10 n.3. On appeal, People.ai makes arguments only as to claims 1, 11, 19, and 20. Appellant's Br. 50-55. We thus limit our discussion to those four claims. See Ballard Med. Prods., 268 F.3d at 1363.
Claim 1 of the '129 patent recites:
A method comprising:
maintaining, by one or more processors, a plurality of node profiles corresponding to a plurality of unique entities, each node profile including a plurality of fields, each field of the plurality of fields including one or more node field values;
accessing, by the one or more processors, a plurality of electronic activities transmitted or received via electronic accounts associated with one or more data source providers, the one or more processors configured to update the plurality of node
profiles using the plurality of electronic activities;
maintaining, by the one or more processors, a plurality of record objects of one or more systems of record, each record object of the plurality of record objects comprising one or more object fields having one or more object field values;
extracting, by the one or more processors, data included in an electronic activity of the plurality of electronic activities;
matching, by the one or more processors, the electronic activity to at least one node profile of the plurality of node profiles based on determining that the extracted data of the electronic activity and the one or more values of the fields of the at least one node profile satisfy a node profile matching policy; matching, by the one or more processors, the electronic activity to at least one record object of the plurality of record objects based on the extracted data of the electronic activity and object values of the at least one record object by:
identifying, by the one or more processors, responsive to applying at least one matching policy of a plurality of matching policies for identifying record objects based on one or more recipients of the electronic activity and a sender of the electronic activity, a set of record objects with which to match the electronic activity, each record
object of the set identified based on the one or more recipients or the sender of the electronic activity; and
selecting by the one or more processors, the at least one record object included in the set of record objects; and
storing, by the one or more processors, in a data structure, an association between the electronic activity and the at least one record object.'129 patent col. 195 ll. 22-67. Dependent claim 11 depends from claim 10, which in turn depends from claim 1; similarly, claim 19 depends from claim 1. Id. col. 197 ll. 24-40, col. 198 l. 56-col. 199 l. 19. Claim 11, by way of dependent claim 10, adds limitations directed to matching an electronic activity to a record object based on information in the relevant node profile. Id. col. 197 ll. 24-40. Claim 19 adds limitations requiring matching an electronic activity to a record object based on "selecting . . . at least one candidate record object included in both the first set of candidate record objects and the second set of candidate record objects to match to the electronic activity based on the first set of rules and the second set of rules of the matching policy." Id. col. 198 l. 56-col. 199 l. 19.
D. District Court Decision
The district court applied the two-step Alice/Mayo test and held that all asserted claims of the '345, '229, and '129 patents are directed to an abstract idea, lack an inventive concept, and are therefore not patent eligible under § 101. Decision at 1205-09, 1212. First, the district court compared the asserted claims to "the activities of a prototypical corporate salesperson." Decision at 1200 (referring to '129 patent's asserted claims); id. at 1206 (identifying corporate salesperson analogy for '229 patent's asserted claims); id. at 1208 (reiterating corporate salesperson analogy for '345 patent's asserted claims). The district court explained that the corporate salesperson filters and matches communications as claimed in the '345 and '229 patents when "she discards the junk mail before updating the business files she maintains with relevant communications." Id. at 1208 ('345 patent); see also id. at 1206 ('229 patent). The corporate salesperson maintains and uses data structures analogous to the claimed node profiles in the '129 patent when she applies business rules (such as checking sender and recipient) to correspondence to match incoming communications to particular contacts and accounts and then updates the correct records. Id. at 1200 ('129 patent). The district court concluded that the asserted claims "do little else than recite a common commercial practice long performed by humans." Id. ('129 patent); see also id. at 1206 (stating that the '229 patent's asserted claims are directed to "a long common practice"); id. at 1208 (stating that the "corporate salesperson has long conducted" the activity claimed in the '345 patent's asserted claims).
With respect to step two, the district court found no inventive concept in the asserted claims of any of the patents. Id. at 1202-05 ('129 patent); id. at 1206-07 ('229 patent); id. at 1208-09 ('345 patent). In sum, it concluded that all asserted claims are invalid as patent ineligible under § 101. Id. at 1212.
People.ai appeals. We have jurisdiction under 28 U.S.C. § 1295(a)(1).
II. DISCUSSION
We review a district court's grant of judgment on the pleadings under the standard of review applied by the regional circuit, here, the Ninth Circuit. See Data Engine Techs. LLC v. Google LLC, 906 F.3d 999, 1007 (Fed. Cir. 2018). The Ninth Circuit reviews grants of motions for judgment on the pleadings de novo, accepting as true all factual allegations in the complaint and viewing those factual allegations in the light most favorable to the plaintiff. LeGras v. AETNA Life Ins. Co., 786 F.3d 1233, 1236 (9th Cir. 2015) (citations omitted).
"Patent eligibility under § 101 is a question of law that may implicate underlying issues of fact." In re Killian, 45 F.4th 1373, 1378 (Fed. Cir. 2022) (citations omitted). "Patent eligibility can be determined on the pleadings under Rule 12(c) when there are no factual allegations that, when taken as true, prevent resolving the eligibility question as a matter of law." Data Engine Techs., 906 F.3d at 1007 (citations omitted).
Under § 101, "[w]hoever 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" Title 35 of the United States Code. The Supreme Court has long held that there is an "implicit exception" in § 101-"[l]aws of nature, natural phenomena, and abstract ideas are not patentable." Alice Corp. Pty. v. CLS Bank Int'l Ltd., 573 U.S. 208, 216 (2014) (quoting Ass'n for Molecular Pathology v. Myriad Genetics, Inc., 569 U.S. 576, 589 (2013)). The Supreme Court has established a two-step test for determining whether claims fall within one of the judicial exceptions. Alice, 573 U.S. at 217-18; Mayo Collaborative Servs. v. Prometheus Lab'ys, Inc., 566 U.S. 66, 77-78 (2012). At step one, we "determine whether the claims at issue are directed to a patent-ineligible concept," such as an abstract idea. Alice, 573 U.S. at 218. If the claims are directed to a patent-ineligible concept, we "examine the elements of the claim to determine whether it contains an inventive concept sufficient to transform the claimed abstract idea into a patent-eligible application" at step two. Id. at 221 (internal quotation marks omitted) (quoting Mayo, 566 U.S. at 72, 80).
The asserted claims of the '345, '229, and '129 patents are directed to an abstract idea at Alice/Mayo step one and lack a saving inventive concept at Alice/Mayo step two. Therefore, the asserted claims are patent ineligible.
A. '345 Patent
We agree with the district court that, under Alice/Mayo step one, claim 11 of the '345 patent is "directed to the abstract idea of data processing by restricting certain data from further analysis based on various sets of generic rules." Decision at 1208. And like the district court, we can find no inventive concept to save this claim from patent ineligibility at Alice/Mayo step two. See id. at 1208-09. Similarly, we hold that claim 18, the only other '345 patent claim about which People.ai makes any specific argument in its briefing before this court, is directed to an abstract idea at Alice/Mayo step one and lacks a redeeming inventive concept at Alice/Mayo step two.
i. Alice/Mayo Step One
Claim 11 is a system claim that relies on "one or more processors" configured to perform the following steps: (1) identify a first and a second electronic activity (e.g., emails); (2) determine that the first electronic activity is sent or received by a certain electronic account by parsing the first electronic activity; (3) determine that the second electronic activity is sent or received by a certain electronic account by parsing the second electronic activity; (4) select a filtering policy that includes at least one of (i) a keyword policy, (ii) a regex pattern policy, or (iii) a logic-based policy; (5) apply the filtering policy "to restrict the first electronic activity from being matched with one or more record objects;" (6) apply the filtering policy and match the second electronic activity to a record object based on a "match policy;" and (7) transmit to a system of record (e.g., CRM) "instructions to store an association between the second electronic activity and the first record object in the system of record." '345 patent col. 192 l. 57-col. 193 l. 53. Our understanding of the claim matches, almost exactly, the district court's explanation of claim 11. Decision at 1208. And as the district court found, this claimed system accomplishes the same ends using the same steps long undertaken by a salesperson or corporate mailroom sorting correspondence and setting aside certain correspondence for further processing and filing. Id.
The Supreme Court has held that "fundamental . . . practice[s] long prevalent in our system of commerce" are abstract ideas. Alice, 573 U.S. at 219-20 (quoting Bilski v. Kappos, 561 U.S. 593, 611 (2010)). As we have found in other cases, "[a]utomation or digitization of a conventional method of organizing human activity . . . does not bring the claims out of the realm of abstractness." Weisner v. Google LLC, 51 F.4th 1073, 1083 (Fed. Cir. 2022); see also Credit Acceptance Corp. v. Westlake Servs., 859 F.3d 1044, 1055 (Fed. Cir. 2017) ("Our prior cases have made clear that mere automation of manual processes using generic computers does not constitute a patentable improvement in computer technology."). The '345 patent confirms that the claimed invention is directed to replacement of an already existing manual process of updating systems of record with an automated process, and the benefits of its claims are improvements to accuracy, speed, and efficiency-benefits inherent in automation. See, e.g., '345 patent col. 50 ll. 12-37. The asserted claims of the '345 patent are similar to those we have found patent ineligible in other cases.
In Intellectual Ventures I LLC v. Symantec Corp., we held that claims directed to a "method of filtering emails" "to address the problems of spam e-mail and the use of email to deliver computer viruses" were directed to an abstract idea. 838 F.3d 1307, 1313 (Fed. Cir. 2016) (Symantec). We explained that "it was long-prevalent practice for people receiving paper mail to look at an envelope and discard certain letters, without opening them, from sources from which they did not wish to receive mail based on characteristics of the mail." Id. at 1314. And we held that applying that "well-known idea using generic computers to the particular technological environment of the Internet" is directed to an abstract idea. Id. (internal quotation marks omitted) (citation omitted).
In University of Florida Research Foundation, Inc. v. General Electric Co., we addressed claims for receiving and converting physiologic treatment data from bedside machines from a machine-specific format into a machine-independent format, processing that data, and displaying the results. 916 F.3d 1363, 1366 (Fed. Cir. 2019). Previously, this treatment data had been entered into information systems manually, which was error-prone, "time-consuming and expensive." Id. at 1367 (citation omitted). The patent at issue proposed "replacing the 'pen and paper methodologies' with 'data synthesis technology' in the form of 'device drivers written for the various bedside machines' that allow the bedside device to present data from the various bedside machines 'in a configurable fashion within a single interface.'" Id. (citation omitted). We found that the patent "acknowledges that data from bedside machines was previously collected, analyzed, manipulated, and displayed manually, and it simply proposes doing so with a com-puter"-"a quintessential 'do it on a computer' patent." Id. (citations omitted). Automation of the previously manual process "conserve[d] human resources and minimize[d] errors." Id. But we held that even though the automation might "result in life altering consequences," a "laudable" outcome, that improvement "does not render it any less abstract." Id. (citation omitted).
The asserted claims of the '345 patent, like the claims in Symantec and University of Florida Research Foundation, are directed to automation of a long prevalent manual process. Symantec is particularly relevant, as that case dealt with claims directed to a method of "filtering emails" akin to the '345 patent's claims to filtering a first and a second "electronic activity." See Symantec, 838 F.3d at 1313. In Symantec, the goal of the claims was to filter out harmful or spam e-mail. Id. In the '345 patent, the goal of the claims is to filter out certain "electronic activities," including emails, leaving only useful electronic activity for inclusion in a system of record. See, e.g., '345 patent abstract. Like Symantec, it was a long prevalent practice for salespeople receiving correspondence to set aside certain correspondence based on its characteristics and to file or further process other correspondence, as called for by the claims of the '345 patent. See 838 F.3d at 1314 ("[I]t was long-prevalent practice for people receiving paper mail to . . . discard certain letters . . . based on characteristics of the mail."). That the claimed automation leads to expected improvements in speed, accuracy, and completeness is laudable, but as we explained in University of Florida Research Foundation, the inherent benefits of automation "do[] not render it any less abstract." See 916 F.3d at 1367.
People.ai seeks to distinguish Symantec by arguing that the Symantec claims used computers to filter emails in the same way that the manual process had long been performed, whereas the asserted claims of the '345 patent "recite a specific series of steps with specific kinds of rulebased filtering policies (e.g., keyword policy, regex pattern, or specific kinds of logic-based policies), with the filtering policies selected at a particular level of granularity, that together differs from the judgment-based process previously used by humans." Appellant's Reply Br. 17. Peo-ple.ai asserts that the '345 patent's claims are more like those we found patent eligible in McRO or Finjan. Appellant's Br. 31-36 (first citing McRO, Inc. v. Bandai Namco Games Am. Inc., 837 F.3d 1299 (Fed. Cir. 2016); and then citing Finjan, Inc. v. Blue Coat Sys., Inc., 879 F.3d 1299 (Fed. Cir. 2018)); Appellant's Reply Br. 16-18. It argues that the claimed system replaces a manual "subjective process with an automated objective and rules-based process," filtering electronic activities in a fundamentally different way than people did manually, which improves the functionality of existing CRM systems. Appellant's Br. 29-30. We disagree.
Automation of a manual process may not be an abstract idea if the automated process differs from the manual process and provides "a specific means or method that improves the relevant technology." See McRO, 837 F.3d at 1314-15 (citing Enfish, LLC v. Microsoft Corp., 822 F.3d 1327, 1336 (Fed. Cir. 2016)). In McRO, we held that using unconventional rules in the ordered combination of claimed steps of patents related to "automating part of a preexisting 3-D animation method" were not directed to an abstract idea at Alice/Mayo step one. Id. at 1302-03. The fact that the steps employed by the claims in McRO differed from those previously employed in the manual process was critical to our conclusion. See id. at 1302-03 ("We hold that the ordered combination of claimed steps, using unconventional rules that relate sub-sequences of phonemes, timings, and morph weight sets, is not directed to an abstract idea and is therefore patent-eligible subject matter under § 101." (emphasis added)); see also FairWarning IP, LLC v. Iatric Sys., Inc., 839 F.3d 1089, 1094 (Fed. Cir. 2016) (explaining that "the traditional process and newly claimed method [at issue in McRO] stood in contrast: while both produced a similar result, i.e., realistic animations of facial movements accompanying speech, the two practices produced those results in fundamentally different ways"). Moreover, the claims incorporating the unconventional rules provided "a specific asserted improvement in computer animation" by "allowing computers to produce 'accurate and realistic lip synchronization and facial expressions in animated characters.'" McRO, 837 F.3d at 1313-14 (citation omitted).
In Finjan, the claims involved generating a "security profile" of a downloadable item via a "behavior-based" virus scan with "information about potentially hostile operations." 879 F.3d at 1304. As in McRO, we found it important that the claimed method differed from the traditional method. Id. at 1304-06 (finding eligible at Al-ice/Mayo step one claims requiring a virus scanning security profile including "information about potentially hostile operations produced by a 'behavior-based' virus scan" in contrast to "traditional, 'code-matching' virus scans that are limited to recognizing the presence of previously-identified viruses, typically by comparing the code in a downloadable to a database of known suspicious code"). We also found that the claims in Finjan were "directed to a nonabstract improvement in computer functionality" by improving the virus scanning capability of computer security systems. Id. at 1305.
Here, the steps claimed in '345 patent claim 11 do not differ from those previously used in the long-prevalent manual practice of selecting certain communications for further processing and filing in a CRM or other system of records- despite People.ai's contrary argument.
Claim 11 of the '345 patent allows for three possible types of filtering rules:
i) a keyword policy configured to identify electronic activities including a predetermined keyword;
ii) a regex pattern policy configured to identify electronic activities including one or more character strings that match a predetermined regex pattern; or
iii) a logic-based policy configured to identify electronic activities based on participants of the electronic activities satisfying a predetermined group of participants.'345 patent col. 193 ll. 17-26. Salespeople have long filtered their correspondence according to rules falling within these broad categories of "filtering policies." For example, the "logic-based policy" would be used by a salesperson choosing not to send e-mails from his or her spouse to a CRM (or before the advent of CRMs, choosing not to file them in his or her business records). The '345 patent confirms that this type of long-practiced filtering is contemplated by the claims. Id. col. 88 ll. 25-44 ("The filtering engine 270 can maintain user-specific filtering policies that include one or more rules defined for specific users.... In another example, the user may define a rule to restrict emails sent to the user's spouse at a given company to be linked to record objects of the company."). People.ai's assertion that this automated objective rule differs from the subjective filtering traditionally used by salespeople is unavailing. A salesperson seeking to not save personal correspondence in his or her business records would use the same rule (excluding from records emails from a spouse's email address) and do so either manually or by using an automated rule. The claims of the '345 patent, unlike those addressed in McRO, do not claim a different method than that traditionally used long before the application of computer technology to the problem of sorting correspondence.
The claims of the '345 patent, also unlike those addressed in McRO and Finjan, do not improve computer functionality. Although the claimed automation of sorting correspondence may improve speed and accuracy, this improvement comes from replacing a human with a computer in that sorting procedure. In such cases, "the focus of the claims is not on . . . an improvement in computers as tools, but on certain independently abstract ideas that use computers as tools." FairWarning IP, 839 F.3d at 1095 (first citing Bancorp Servs., L.L.C. v. Sun Life Assurance Co. of Can. (U.S.), 687 F.3d 1266, 1278 (Fed. Cir. 2012); and then citing Elec. Power Grp., LLC v. Alstom S.A., 830 F.3d 1350, 1354 (Fed. Cir. 2016)). That principle applies here.
People.ai asserts that the asserted claims of the '345 patent differ from the traditional process because manually practicing all of the steps of these claims:
would require opening and reading all incoming and outgoing communications at a company (as
well as monitoring all phone calls and meetings); determining the sender and recipients of each communication; selecting different rules-based filtering policies depending on the account associated with the particular sender or recipient; reading the content of all of those communications in order to identify relevant business information and apply the selected filtering policies; applying the rules-based filtering policies to assess whether the communication should be logged as relevant to particular sales opportunities or instead excluded; then logging the appropriate information in the correct account or opportunity record in a place where salespeople, company leadership, and other corporate systems could then make use of that information.
Appellant's Br. 46 (emphases in original).
People.ai's arguments are not tethered to the asserted claims. For example, claim 11 requires analysis of only two communications ("identify a first electronic activity and a second electronic activity"), not analysis of every communication into and out of a company. Compare '345 patent col. 192 ll. 60-61 with Appellant's Br. 46. Similarly, claim 11 does not require the claimed system to "read[] the content of all of those communications in order to identify relevant business information." Compare '345 patent col. 193 ll. 113 with Appellant's Br. 46. Rather, claim 11 requires that the processor(s) "parse the first electronic activity to identify one or more electronic accounts associated with at least the sender or the one or more recipients of the first electronic activity"-not necessarily read the content of all communications. '345 patent col. 193 ll. 1-4. Finally, claim 11 does not require the claimed system to "log[] the appropriate information in the correct account or opportunity record in a place where salespeople, company leadership, and other corporate systems could then make use of that information." Compare '345 patent col. 193 ll. 51-53 with Appellant's Br. 46. Rather, it requires the system to "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"-not necessarily in a place where specific people can access the information. '345 patent col. 193 ll. 51-53.
After weeding out the steps not required by the claims from People.ai's argument, three steps remain: "determining the sender and recipients of each communication; selecting different rules-based filtering policies depending on the account associated with the particular sender or recipient;" and "applying the rules-based filtering policies to assess whether the communication should be logged as relevant to particular sales opportunities or instead excluded." Appellant's Br. 46. As we discussed above, we see no difference between these limitations and the steps employed by the prototypical salesperson in the manual process.
Lastly, People.ai argues that claim 18's "node profile[]" limitations provide an additional concrete limitation rendering it patent eligible at Alice/Mayo step one. Appellant's Br. 30; '345 patent col. 194 ll. 34-52. People.ai argues that the claimed node profile is a specific data structure that allows for "matching and filtering based on information that is not in the CRM." Appellant's Br. 30; see also id. at 34-35. People.ai argues that by doing some analysis outside the CRM, the claims permit bulk updating of a CRM and solve a network traffic limitation imposed by certain CRMs. Id. at 31.
The district court found that People.ai did not meaningfully distinguish the limitations of claim 18 from claim 11 and concluded that claim 18 was directed to the same abstract idea: "data processing by restricting certain data from further analysis based on various sets of generic rules." Decision at 1208-09. We agree. People.ai's arguments fail because the node profile, as described by the specification and discussed by the district court, is merely a computerized version of a rolodex entry or file corresponding to an individual, such as a specific customer, in a filing cabinet. The district court stated that node profiles are "data profiles that store information on various entities, such as a person's name and email address." Id. at 1199. On appeal, People.ai agrees with that definition and points to it in connection with claim 18 of the '345 patent. Appellant's Br. 13 ("The specification explains that 'node profiles' are 'data profiles that store information on various entities, such as a person's name and email address.'") (citation omitted). Moreover, People.ai asserts that the node profiles solve a specific technological problem created by some CRMs, such as Salesforce, that limit the number of daily interactions a user may have with the CRM, Appellant's Br. 31, but the claims are not limited to CRMs. That is, there is no evidence of a technological problem with the claimed systems of record, nor does claim 18 present a technological solution; rather it presents a conventional solution to a conventional problem of data organization.
In conclusion, the asserted claims of the '345 patent are directed to the abstract idea of "data processing by restricting certain data from further analysis based on various sets of generic rules." See Decision at 1208. This is a longstanding process. As we have held in several prior cases, automation of a longstanding manual process is not patent eligible at Alice/Mayo step one. See, e.g., Symantec, 838 F.3d at 1313-16; Univ. of Fla. Rsch. Found., 916 F.3d at 1367. Nor do the benefits of that automation, such as increased accuracy or efficiency, render the automated process patent eligible at Alice/Mayo step one. Thus, the asserted claims of the '345 patent, as exemplified by claims 11 and 18, are directed to an abstract idea at Alice/Mayo step one.
We find no inventive concept in either claim 11 or claim 18 of the '345 patent sufficient to render those claims patent eligible at Alice/Mayo step two.
We have repeatedly held that "[t]he abstract idea itself cannot supply the inventive concept." Am. Axle &Mfg., Inc. v. Neapco Holdings LLC, 967 F.3d 1285, 1299 (Fed. Cir. 2020) (collecting citations) (citations omitted); see also Trading Techs. Int'l, Inc. v. IBG LLC, 921 F.3d 1378, 1385 (Fed. Cir. 2019). Here, as already discussed above, claim 11 of the '345 patent requires a system to (1) identify a first and a second electronic activity (e.g., emails); (2) determine that the first electronic activity is sent or received by a certain electronic account by parsing the first electronic activity; (3) determine that the second electronic activity is sent or received by a certain electronic account by parsing the second electronic activity; (4) select a filtering policy that includes at least one of (i) a keyword policy, (ii) a regex pattern policy, or (iii) a logic-based policy; (5) apply the filtering policy "to restrict the first electronic activity from being matched with one or more record objects;" (6) apply the filtering policy and match the second electronic activity to a record object based on a "match policy;" and (7) transmit to a system of record (e.g., CRM) "instructions to store an association between the second electronic activity and the first record object in the system of record." '345 patent col. 192 l. 57-193 l. 53. Those steps are all necessary parts of the abstract idea of "data processing by restricting certain data from further analysis based on various sets of generic rules." They cannot supply the inventive concept.
The features of claim 11 not recited in the above paragraph are generic computer features: "A system comprising: one or more processors coupled with memory and configured by machine-readable instructions." '345 patent col 192 ll. 57-59. "[T]he mere recitation of a generic computer cannot transform a patent-ineligible abstract idea into a patent-eligible invention." Alice, 573 U.S. at 223. Claims directed to performance of a longstanding manual process on a generic computer may lead to laudable increases in efficiency or accuracy (the exact kinds of improvements computerization is expected to yield in all cases), but lack an inventive concept which might render them patent eligible at Alice/Mayo step two. Intell. Ventures I LLC v. Cap. One Bank (USA), 792 F.3d 1363, 1367 (Fed. Cir. 2015) (Cap. One) ("Nor, in addressing the second step of Alice, does claiming the improved speed or efficiency inherent with applying the abstract idea on a computer provide a sufficient inventive concept.").
People.ai asserts that the inventive concept can be found in (1) the ordered combination of steps recited "to extract data from bulk communications activities," Appellant's Br. 37; (2) the filtering rules of claim 11, id. at 38-39; and (3) the node profiles of claim 18, id. at 41, 44. People.ai argues that the district court oversimplified the claims by neglecting the aforementioned limitations. Id. at 44-50. It contends that the claims of the '345 patent are more specific than-and thus distinguishable from-the claims that this court held patent ineligible in Symantec. Id. at 41-42. It argues that the claims of the '345 patent are like those found patent eligible at Alice/Mayo step two in BASCOM and Amdocs. Id. at 38-41 (first citing BASCOM Glob. Internet Servs., Inc. v. AT&T Mobility LLC, 827 F.3d 1341 (Fed. Cir. 2016); and then citing Amdocs (Israel) Ltd. v. Openet Telecom, Inc., 841 F.3d 1288 (Fed. Cir. 2016)). Finally, People.ai argues that the Defendants had to produce evidence that the claimed system was in routine or conventional use. Appellant's Br. 38, 46. We do not find these arguments persuasive.
As to People.ai's argument that an inventive concept can be found in the asserted claims' ordered combination of steps, the ordered combination of steps are exactly the same steps that a salesperson would have traditionally undertaken to filter and sort his or her correspondence by hand. See Two-Way Media Ltd. v. Comcast Cable Commc'ns, LLC, 874 F.3d 1329, 1341 (Fed. Cir. 2017) ("The steps are organized in a completely conventional way- data are first processed, sent, and once sent, information about the transmission is recorded."). The ordered combination of steps, which matches the ordered combination of steps traditionally practiced by people manually, are themselves part of the abstract idea and "cannot supply the inventive concept." See Am. Axle, 967 F.3d at 1299 (citation omitted). Our conclusion is confirmed by People.ai's identification of the purported benefit of the asserted claims- avoiding the pitfalls of manual data entry by using a computer to implement "tailored, objective selection of relevant business activities to identify relevant communications and their relationships to particular accounts and sales opportunities, particularly with the nuance and accuracy that the People.ai system's architecture allows for." Appellant's Br. 37-38. These improvements in speed, cost, and accuracy are benefits of using computers for automation generally and do not result from some other inventive concept. The ordered combination of steps in the '345 patent's claims 11 and 18 do not provide an inventive concept.
People.ai's argument that the filtering rules recited in claim 11 provide an inventive concept leads to the same conclusion. Because these filtering rules are part of the abstract idea itself, they cannot provide an inventive concept.
People.ai's arguments about claim 18's "node profiles" are not tethered to claim 18. People.ai argues that there is no conventional "brick-and-mortar" mailroom that maintained a "node graph of node profiles to enable the identification, storage, and analysis of data and relationships that would otherwise be unrecorded." Appellant's Br. 46. But as People.ai admits in its reply brief, none of the asserted claims require a node graph. Appellant's Reply Br. 24 ("[T]he claims themselves do not require a 'node graph[.]'"). And we see no requirement in claim 18 requiring storing "relationships" between node profiles. '345 patent col. 194 ll. 34-52. "[W]e have repeatedly held that features that are not claimed are irrelevant as to step 1 or step 2 of the Mayo/Alice analysis." Am. Axle, 967 F.3d at 1293 (citations omitted).
The '345 patent's specification explains that a node graph includes "a plurality of nodes and a plurality of edges between the nodes indicating activity or relationships." '345 patent col. 6 ll. 40-44. Each field in a node profile can include "one or more value data structures," including a value, an "occurrence metric" indicating "a level of certainty" that the recorded value is correct and can record the specific data source and electronic activity from which each value was derived. Id. col. 15 ll. 15-37, col. 18 ll. 9-17. As more data is added, the node graph can use that additional data "to populate missing fields or add new values to existing fields, reinforce field values that have low confidence scores and further increase the confidence score of field values, adjust confident scores of certain data points, and identify patterns or make deductions based on the values of various fields of node profiles of nodes included in the graph." Id. col. 12 ll. 55-65.
Claim 18 actually requires a "plurality of node profiles corresponding to a plurality of unique entities, each electronic account of the one or more electronic accounts linked to a respective node profile of the plurality of node profiles." '345 patent col. 194 ll. 36-39. Claim 18 further requires a filtering policy to be applied "based on extracted field-value pairs from the node profiles for the participants of the first electronic activity." Id. col. 194 ll. 49-52. The node profiles, as actually claimed, do not describe an unconventional architecture or unconventional assemblage of generic parts that might convey an inventive concept. Peo-ple.ai agrees with the district court's definition of "node profiles": "data profiles that store information on various entities, such as a person's name and email address." Appellant's Br. 13 (quoting Decision at 1199 when describing the claimed node profiles of claim 18). This is not an unconventional architecture. Instead, it is an electronic rolodex or an electronic filing cabinet used to store business correspondence and records with files for each customer.
People.ai argues that the asserted claims of the '345 patent are more detailed than and thus distinguishable at Alice/Mayo step two from those that this court found patent ineligible in Symantec. Id. at 41-42. People.ai argues that the asserted claims' requirements of "selection of specific kinds of objective rules, using a particular architecture, to . . . replac[e] subjective human judgment and data entry" distinguish the '345 patent's asserted claims. Id. at 42. This argument fails for the same reason discussed above- the "objective" rules permitted by the asserted claims include those used by a person manually filtering his or her correspondence and entering data. See Symantec, 838 F.3d at 1314-16 (claiming a "long-prevalent practice" without "improv[ing] the functioning of the computer itself" insufficient to render claims patentable) (citations omitted).
People.ai's argument that the asserted claims of the '345 patent are analogous to those we found patent eligible at Alice/Mayo step two in BASCOM also fails. In BASCOM, we found that claims were directed to the abstract idea of "filtering content on the Internet." 827 F.3d at 1348. At Alice/Mayo step two, we explained that "an inventive concept can be found in the non-conventional and non-generic arrangement of known, conventional pieces." Id. at 1350. And we found that the claims in BASCOM had such an inventive concept: "the installation of a filtering tool at a specific location, remote from the end-users, with customizable filtering features specific to each end user," which "gives the filtering tool both the benefits of a filter on a local computer and the benefits of a filter on the ISP server." Id. We emphasized that "[t]he claims do not merely recite the abstract idea of filtering content along with the requirement to perform it on the Internet, or to perform it on a set of generic computer components." Id. "Such claims would not contain an inventive concept." Id. (citation omitted). The claims of the '345 patent are such claims. Unlike the claims in BASCOM that contained the technological improvement of permitting customizable filtering at a specific location coupled with the benefits of remote filtering at the ISP server, 827 F.3d at 1350, the claims of the '345 patent do not require installation of the filtering tool at a specific location yielding technologically unique benefits. Rather, the claims mirror the manual process performed in corporate mailrooms long before the '345 patent's proposed automation, and the cited benefits are only those expected of any automation-increased speed and accuracy-benefits which we explained did not provide an inventive concept in Symantec. See 838 F.3d at 1315 ("'[C]laiming the improved speed or efficiency inherent with applying the abstract idea on a computer' does not 'provide a sufficient inventive concept.'" (quoting Cap. One, 792 F.3d at 1367)).
People.ai's comparison to Amdocs is similarly unavailing. In Amdocs, one of the claims at issue "entail[ed] an unconventional technological solution (enhancing data in a distributed fashion) to a technological problem (massive record flows which previously required massive databases)." 841 F.3d at 1300. Although the solution required generic components, it "necessarily require[d] that these generic components operate in an unconventional manner to achieve an improvement in computer functionality." Id. at 1300-01. Unlike the claim in Amdocs, the claims of the '345 patent seek to solve a conventional problem (slow and error-prone manual data entry) with a conventional solution (automation of manual data entry). Like the claim in Amdocs, the claims of the '345 patent require generic computer components, but unlike the claim in Amdocs, they do not require those components to operate in an unconventional manner. Thus, Amdocs does not compel us to find that the claims of the '345 patent have an inventive concept at step two.
Finally, People.ai argues that the asserted claims of the '345 patent cannot be held patent ineligible because the Defendants cited no evidence showing that "the claims recite an invention that is [] merely the routine or conventional use" of generic computer components and therefore failed to bear their burden of showing invalidity by clear and convincing evidence. Appellant's Br. 38 (quoting DDR Holdings, LLC v. Hotels.com, L.P., 773 F.3d 1245, 1259 (Fed. Cir. 2014)). The Defendants argue that People.ai's evidentiary argument is a red herring because the claimed steps merely spell out the abstract idea of filtering and filing emails and the abstract idea cannot provide the inventive concept. Appellees' Br. 50. We agree with the Defendants.
"[W]hether a claim recites patent eligible subject matter is a question of law[,] which may contain underlying facts." Berkheimer v. HP Inc., 881 F.3d 1360, 1368 (Fed. Cir. 2018) (citations omitted). "Any fact . . . pertinent to the invalidity conclusion must be proven by clear and convincing evidence." Id. (citing Microsoft Corp. v. i4i Ltd. P'ship, 564 U.S. 91, 95 (2011)). But "not every § 101 determination contains genuine disputes over the underlying facts material to the § 101 inquiry." Id. (citations omitted). This is one such case where there is no genuine dispute over the underlying material facts. It is undisputed that the computer components recited by the asserted claims of the '345 patent ("one or more processors coupled with memory and configured by machine-readable instructions," '345 patent col. 192 ll. 58-59) are generic. And the method being performed on those generic components matches the manual process conventionally performed by a salesperson, i.e., the abstract idea itself under Alice/Mayo step one.
We conclude that there is no inventive concept in claims 11 or 18 of the '345 patent. And as People.ai made no specific arguments about any other claim, we conclude that the asserted claims of the '345 patent lack an inventive concept. We affirm the district court's decision that the asserted claims of the '345 patent are patent ineligible under § 101. See Decision at 1208-09.
B. '229 Patent
We reach the same result for the asserted claims of the '229 patent. At Alice/Mayo step one, claim 19 of the '229 patent is directed to the same abstract idea as the asserted claims of the '345 patent. And we, like the district court, can find no saving inventive concept at Alice/Mayo step two. Decision at 1205-07.
Claim 19, like claim 11 of the '345 patent, requires a filtering policy. Unlike claim 11 of the '345 patent, claim 19 does not restrict the filtering policy to a type or types of rules. See '229 patent col. 144 l. 40-col. 145 l. 25. Rather, with respect to the filtering policy, it says only: "determine, responsive to applying a first policy including one or more filtering rules, that the electronic activity is to be matched to at least one record object of the identified system of record." Id. col. 144 l. 66-col. 145 l. 2. Claim 19 requires various other rules, but it similarly does not provide any specificity as to those rules: "identify a first set of candidate record objects to which to match the electronic activity responsive to applying a second policy including a first set of rules for identifying one or more record objects of a first record object type based on an object field value of the record object that identifies the one or more recipients;" and "identify a second set of candidate record objects to which to match the electronic activity responsive to applying the second policy including a second set of rules for identifying candidate record objects based on the sender of the electronic activity, wherein the second policy includes a third set of rules for identifying candidate record objects of a second record object type." Id. col. 145 ll. 6-18 (emphases added). These limitations do nothing to distinguish the claims of the '229 patent from those of the '345 patent already found to be abstract at Alice/Mayo step one above.
People.ai makes the same arguments with respect to the '229 patent as it makes with respect to the '345 patent, see Appellant's Br. 28-50, and those arguments fail for the same reasons. The only new argument People.ai advances as to the '229 patent is that its claims require data storage on a local processor outside of the CRM and that this additional requirement, which the '345 patent lacks, makes the claims of the '229 patent not abstract. Id. at 31; Appellant's Reply Br. 13. People.ai also argues that this local storage requirement renders the asserted claims of the '229 patent eligible at Alice/Mayo step two. Appellant's Br. 41. Peo-ple.ai's argument is not persuasive.
Even assuming the '229 patent's claims require the use of local storage-an argument the district court rejected, Decision at 1206, local storage of information is not sufficient in this case to render the asserted claims patent eligible. Wherever the associations are stored, the idea underlying the '229 patent's claims is abstract. Storing associations locally for bulk upload to a system of record, such as a CRM, is itself an abstract idea. It is similar to a corporate mailroom, which might sort mail according to filtering policies, match mail to certain filing locations or recipients, and then store that mail in the mailroom until delivering it in bulk once a day. This local storage requirement cannot provide the inventive concept.
C. '129 Patent
The asserted claims of the '129 patent are similar to those of the '345 and '229 patents, and they fail both steps of the Alice/Mayo test for many of the same reasons. In People.ai's own words: "Claim 1 of the '129 Patent recites a method for constructing and maintaining a node graph based on data extracted from communications activities and matched to a CRM." Appellant's Br. 50 (citing '129 patent col. 195 ll. 22-67). However, as People.ai admits in its reply brief, the claims do not require a node graph but merely require a plurality of node profiles. Appellant's Reply Br. 24. And as already explained, "node profiles" are "data profiles that store information on various entities, such as a person's name and email address." Decision at 1199. Construction and maintenance of these node profiles fails both steps of Alice/Mayo. As already explained, node profiles are merely an electronic rolodex or an electronic filing cabinet used to store business correspondence and records with files for each customer.
The matching policies of claim 1 of the '129 patent do not help. They are generic matching policies that would be used by any person to manually associate correspondence with an entry in a rolodex or filing cabinet. For example, the claimed matching policies are fulfilled by matching an electronic activity to a node profile "based on determining that the extracted data of the electronic activity and the one or more values of the fields of the at least one node profile satisfy a node profile matching policy" and by matching an electronic activity to a record object based on recipient or sender. '129 patent col. 195 ll. 42-61. In other words, the matching policies can be fulfilled by filing an electronic activity in a digital file folder according to the identity of the sender. Claim 1, even when considering the matching policies, fails both Alice/Mayo steps for the same reasons provided for the '345 and '229 patents above.
Dependent claim 11, in People.ai's own words, "provides for matching an electronic activity to a record object based on information that is not in the electronic activity itself or stored in the system of record, but instead is stored exclusively in a node profile of the node graph." Appellant's Br. 51. But the fields of the node profile include information such as a "person's name and email address." Decision at 1199. The claimed searching for emails using an email address is an abstract idea. Furthermore, it does not provide an inventive concept because, as we explained in our analysis of the '345 patent, "[t]he abstract idea itself cannot supply the inventive concept." Am. Axle, 967 F.3d at 1299 (citation omitted).
Dependent claim 19 (also dependent from claim 1) requires, in People.ai's own words, "a specific kind of matching policy-a policy including two sets of rules (one for identifying record objects based on the recipients and the other based on the sender)-and that identifies the relevant records in a particular way (where the record object matches both sets of policies)." Appellant's Br. 51. These sets of rules merely organize data according to the sender and recipient of a communication. Notably, corporate mailrooms and salespeople have long organized correspondence by sender and recipient and filed said correspondence in the correct file based on that information. This claim too fails both steps of the Alice/Mayo inquiry.
CONCLUSION
We have considered People.ai's remaining arguments and find them unpersuasive. For the reasons discussed above, we affirm the district court's decision.
AFFIRMED.