US 12438978

I'll search for this specific patent number across sources, including USPTO and CAFC 2026 docket references.

Let me verify with additional targeted searches, including CAFC docket and any litigation references.

Found litigation references. Let me investigate the IQX Corp. enforcement activity and confirm which patents are at issue.

US Patent 12,438,978 — Summary

Bibliographic Data (as verified)

Continuity (per the front matter): This patent is a continuation of Ser. No. 18/984,041 (filed Dec. 17, 2024), itself a continuation of Ser. No. 18/896,197 (filed Sep. 25, 2024), which is a continuation-in-part of Ser. No. 18/483,670 (filed Oct. 10, 2023), a continuation of Ser. No. 18/193,163 (filed Mar. 30, 2023), a continuation of Ser. No. 17/279,188 (filed Mar. 24, 2021; U.S. National Stage of PCT/US19/52847, filed Sep. 25, 2019), claiming benefit of provisional 62/737,315 (Sep. 27, 2018). The Google Patents timeline also lists a later family member, application 19/324,477 → US 2026/0012531 A1 (dated 2025-09-10).

Sources: https://patents.google.com/patent/US12438978/en ; corroborating family/description pages for the pre-grant publication at https://patents.justia.com/patent/20250274544 and https://www.patents-review.com/a/20250274544-systems-methods-resuming-online-abandoned-journeys.html ; assignee/portfolio listing at https://www.ipqwery.com/ipowner/en/owner/ip/[2495953](/patent/2495953)-iqx-corp.html

---

Abstract (verbatim, as published)

"Systems and methods are described for receiving a request to provide content to at least one authorized recipient; generating a customized message comprising at least one action associated with the content; causing transmission of the customized message to a mobile computing device associated with the at least one authorized recipient, requesting authorization data corresponding to the at least one authorized recipient; responsive to receiving the requested authorization data: verifying that the authorization data correlates to an identity of the at least one authorized recipient or an identity of the mobile computing device; determining a classification of the at least one authorized recipient; and generating a Rich Communication Service (RCS) message comprising the content; and scheduling, using a machine learning model and based on the determined classification, delivery of the RCS message to the mobile computing device."

Note the label/reality mismatch worth flagging: the title emphasizes resuming online abandoned journeys, but the abstract and independent claims are drafted more broadly around RCS messaging, silent-network authentication, and ML-based send-time/classification. The abandonment-resumption subject matter is carried mainly in the specification (Figs. 15–18) and in dependent claims.

---

Plain-Language Overview of the Independent Claims

The patent recites three independent claims — a method claim, a system claim, and a computer-readable storage medium claim. All three share the same core workflow; the differences are in claim type and in the vocabulary used for the end user ("authorized recipient," "service subscriber," or "authorized recipient or subscriber") and for the actionable element ("action" in the method/system claims vs. "link" in the medium claim).

Independent Claim 1 — Computer-implemented method (core workflow):

1. Receive a request to provide content to at least one authorized recipient.
2. Generate a customized message that includes at least one action associated with that content (e.g., a link/button).
3. Transmit the customized message to a mobile computing device belonging to the authorized recipient.
4. After the user selects the action, the system requests authorization data for that recipient.
5. If authorization data is received, the system (a) verifies that the data correlates to the recipient's identity or the mobile device's identity, (b) determines a classification of the recipient based on that identity, and (c) generates an RCS message containing the content.
6. Schedules delivery of the RCS message using a machine learning model and based on the determined classification.

In plain terms: a marketing/transaction message with a clickable action is sent by text; the click triggers an identity check; the verified user is sorted into a tier; and an AI model decides when to push the follow-up rich (RCS) message to that tier.

Independent System Claim — "Automated messaging subscription and distribution system":
A processor + memory system performing the same steps, but reciting "service subscriber" instead of "authorized recipient," requesting "subscription data" upon action selection, and generating the RCS message "based on the determined classification."

Independent Storage-Medium Claim:
A non-transitory computer-readable medium storing instructions to perform the same steps, but with "at least one link associated with the content" (rather than "action") and with the user termed "authorized recipient or subscriber."

Dependent claims add (summarized):

• Actually delivering the RCS message per the schedule.
• Content as a carousel of selectable options — perform a secure transaction, schedule an appointment, or resume an abandoned user journey.
• Content rendered within a browser on the mobile device.
• Send/receive (real-time typing) indicators in the mobile UI.
• Sender verification of the RCS message, with an SMS warning delivered if the sender is not authorized.
• Actions offering: perform a transaction, select a suggested response, return to a prior browsing session, or return to a prior application session.

---

Specification Highlights Supporting the Claims

• Architecture: an Automated Messaging Distribution Array (AMDA) with an ML/AI decision cluster (1000a) and an AMDA subscriber vault (1000b) that sorts recipients into tier 1 buyers, tier 2 link clickers, and tier 3 bulk-send recipients, routed to a just-in-time messaging array (1000c) or bulk array (1000d) and an automated scheduler/distribution hub (1000e) supporting SMS/MMS, RCS, and OTT channels.
• Security: a Silent Network Authentication Protocol (SNAP) layer (1100) that silently verifies network/device credentials (SIM/carrier identifiers) plus optional biometric authentication; a "hot cookie" bound to the mobile device gates access to the customized page and transaction flow.
• Abandonment taxonomy (Figs. 15–16): session, browser, search, form, onboarding/signup, wish list, checkout, and cart abandonment; each type is recorded and the ML/AI cluster selects a follow-on message/link that returns the user to the abandoned access point.

---

Litigation / Docket Findings

• I found no Court of Appeals for the Federal Circuit (CAFC) 2026 docket involving US 12,438,978. I cannot confirm any CAFC appeal for this specific patent.
• The Google Patents family record carries a Darts-IP flag indicating the family " has litigation" (family ID 94175196): https://patents.darts-ip.com/?family=94175196
• The identified U.S. case is IQX Corp. v. Intuit Inc., No. 2:26-cv-00351 (E.D. Tex., Marshall Division), filed April 30, 2026, Judge Rodney Gilstrap, Magistrate Roy S. Payne, Nature of Suit 830 (35 U.S.C. § 271 patent infringement), accused product identified as Intuit MailChimp text-messaging marketing functionality.
  • Sources: https://gaeflexstaging-dot-docketupdate.appspot.com/cases/Texas_Eastern_District_Court/2--26-cv-00351/IQX_Corp._v._Intuit_Inc/ ; https://companyprofiles.justia.com/company/intuit/dockets/case ; https://cdn.pacermonitor.com/public/case/64428468/IQX_Corp_v_Intuit_Inc
  • A third-party complaint analysis of that case discusses asserted U.S. Patent No. 12,368,799 ("Automated silent network authentication protocol") and a "'924 Patent" (consistent with US 11,876,924, "Secure transactions using customized webpages," granted 2024-01-16). It did not identify US 12,438,978 as one of the asserted patents. Analysis link: https://ai-lab-cl-prod.azurewebsites.net/case/dct/txed/2:26-cv-00351/doc/analysis/1

---

Confidence and Uncertainty Notes

• High confidence: patent number, title, inventors, assignee, application number, filing date, issue date, priority chain, abstract text, and the substance of the independent claims — all from the patent's own front matter and specification as published.
• Medium confidence: the precise number of claims and the exact claim numbering of the independent claims. The text available to me reproduces the claim language in the Summary sections but does not show a numbered claim set, so I have described the three independent claims by type rather than by claim number.
• Uncertainty — no correction made, per instruction: the application serial is presented as "US19/192,732" and the follow-on family member as "US19/324,477." These are reported literally.
• Uncertainty: Whether US 12,438,978 is asserted in the IQX v. Intuit litigation. The publicly indexed complaint analysis references the '799 and '924 patents; I could not confirm '978's inclusion.
• Not found/not verified: any CAFC (Federal Circuit) 2026 appellate docket, PTAB/IPR proceeding, or USPTO PatentCenter record specifically for 12,438,978 beyond the Google Patents and family sources cited. I did not independently retrieve the USPTO PatentCenter "Bibliographic Data" page for this patent, so the grant date and status are as reported by Google Patents rather than directly confirmed at USPTO.

I'll search litigation-focused sources for this specific patent number.

Let me search additional litigation databases and specifically check whether the '978 patent has been asserted.

Let me verify the Salesforce case details and check for any additional cases asserting the '978 patent.

Litigation Involving US 12,438,978

Bottom line up front: I found exactly one case in which US 12,438,978 ("Systems and methods for resuming online abandoned journeys") is affirmatively identified as an asserted patent — IQX Corp. v. Salesforce, Inc. I also found a related, near-identical IQX enforcement campaign against Intuit (same plaintiff, same accused "text message marketing system" theory), but the indexed complaint analysis for that case identifies the '799 and '924 patents, not the '978 patent. I found no CAFC appeal, no PTAB/IPR challenge, and no ITC action specifically directed at 12,438,978.

Before the detail, one rule-mandated flag: your stated "current date" of April 26, 2026 is inconsistent with the evidence. Both IQX complaints have filing dates after April 26, 2026 (April 30, 2026 and May 28, 2026), and the previously generated section of this analysis was itself dated 2026-09-25. I have reported the filing dates literally as the sources state them, without auto-correcting, but the chronology does not reconcile with an April 26, 2026 present date.

---

Case 1 — IQX Corp. v. Salesforce, Inc. (the '978 assertion)

This is the only case in the search results that expressly ties US 12,438,978 to a defendant. The third-party complaint analysis states: "U.S. Patent No. 12,438,978, 'Systems and methods for resuming online abandoned journeys,' issued October 7, 2025 … The complaint specifically alleges infringement of independent claim 1," and "The 'Salesforce text message marketing system' is accused of infringing the '978 Patent."

Sources:

• https://ai-lab-cl-prod.azurewebsites.net/case/dct/txwd/7:26-cv-00213/doc/analysis/1
• https://ai-lab-cl-prod.azurewebsites.net/case/dct/txwd/7:26-cv-00213/iqx-corp-v-salesforce-inc (case record: 7:26-cv-00213, W.D. Tex., filed 05/28/26, Status: Open)
• https://cdn.pacermonitor.com/public/case/64898612/IQX_Corp_v_Salesforce,_Inc (PACER docket entries)

⚠️ Cross-check note: The indexed Salesforce analysis describes a "'165" patent whose abstract and independent claim 1 are textually identical to the '978 patent's subject matter (receive request → customized message → authorization → classification → RCS message → ML-scheduled delivery). It is possible the analysis conflated or duplicated the '978 recitation. I report both as they appear and cannot confirm from the available records that '978 and the '165 patent are the same document. Treat the "claim 1 of '978 is asserted" statement as medium confidence — it rests on a third-party summary, not the complaint itself.

---

Case 2 — IQX Corp. v. Intuit Inc. (related campaign; '978 NOT confirmed)

The indexed complaint analysis for this case discusses the '799 and '924 patents only; it does not identify US 12,438,978 as an asserted patent here. This is consistent with (and does not alter) the uncertainty flagged in the previously generated section.

Sources:

• https://gaeflexstaging-dot-docketupdate.appspot.com/cases/Texas_Eastern_District_Court/2--26-cv-00351/IQX_Corp._v._Intuit_Inc/
• https://cdn.pacermonitor.com/public/case/64428468/IQX_Corp_v_Intuit_Inc
• https://ai-lab.exparte.com/case/dct/txed/2:26-cv-00351/doc/analysis/1
• https://companyprofiles.justia.com/company/intuit/dockets/case

---

Negative findings (searched, not found)

• No CAFC appellate docket for US 12,438,978 (no appeal of either case; both are at the pleading stage anyway).
• No PTAB / IPR / PGR challenge to US 12,438,978.
• No ITC § 337 action involving US 12,438,978.
• The Google Patents family record (family ID 94175196) carries a Darts-IP "has litigation" flag: https://patents.darts-ip.com/?family=94175196 — this is a family-level indicator and does not by itself identify which patent is asserted; it is consistent with the two IQX campaigns above.

---

Confidence and caveats

• High confidence: The existence, parties, court, case number, filing date, and "Open" status of IQX Corp. v. Salesforce, Inc., 7:26-cv-00213 (W.D. Tex.) and IQX Corp. v. Intuit Inc., 2:26-cv-00351 (E.D. Tex.) — corroborated across PacerMonitor, Justia, and docket-aggregator sources.
• Medium confidence: That US 12,438,978 is asserted in the Salesforce case. This rests on an AI-generated complaint summary; I could not retrieve the complaint PDF or a claim-chart exhibit (Exhibit 2-type attachments were noted as "not attached/not provided") to confirm first-hand.
• Not confirmed / do not assume: That US 12,438,978 is asserted in the Intuit case — the indexed Intuit analysis identifies only '799 and '924.
• Sources not directly accessible in this session: I did not retrieve a live PACER docket sheet or the Unified Patents portal case list for either case; the Unified Patents search returned unrelated results. If exact "current status" (e.g., service, answer, motion practice) is required, the authoritative sources are the W.D. Tex. 7:26-cv-00213 and E.D. Tex. 2:26-cv-00351 PACER dockets.
• Date discrepancy (flagged, not corrected): The case filing dates post-date your stated current date of April 26, 2026. Per instructions I have not normalized them.

If you want, I can attempt to pull the actual Salesforce complaint PDF and its claim charts to move the '978-assertion finding from medium to high confidence.

I'll research the single flagged proceeding and cross-check for any others.

Key find: a parallel district-court case against Salesforce. Let me dig into the petition specifics and check for any sibling IPRs or appeals.

Found the parallel case. Now let me pin down the petition's grounds, panel, and whether sibling patents drew other IPRs.

PTAB / AIA Trial Report — U.S. Patent No. 12,438,978 B2

"Systems and methods for resuming online abandoned journeys" · IQX Corp. (Patent Owner) · Priority 2018-09-27 · Granted 2025-10-07
Report date: 2026-09-25

---

Proceedings overview

One (1) AIA trial proceeding is on file against US 12,438,978: IPR2026-00478 (Salesforce, Inc., filed 2026-09-24, status "Pending"). Breakdown: 1 active/pending · 0 claims invalidated · 0 claims sustained · 0 settled · 0 institution denials. Bottom line for a defendant: the '978 patent is completely untested at the PTAB — no claim has been canceled, no claim has been confirmed, and no institution decision exists yet. The single petition was docketed one day before this report, so there is no FWD, no panel, no claim-level outcome, and no appeal. Any demand letter citing the '978 patent today is citing claims that are live and presumptively valid, and the only defensive value currently available from the PTAB is a single, freshly filed petition whose merits the Board has not yet touched.

Contradiction / update flag against the previously generated sections. The earlier "Litigation / Docket Findings" section identified IQX Corp. v. Intuit Inc., No. 2:26-cv-00351 (E.D. Tex.) as the only U.S. case and stated that the public complaint analysis "did not identify US 12,438,978 as one of the asserted patents." That is now superseded in one respect: a second, later-filed case — IQX Corp. v. Salesforce, Inc., No. 7:26-cv-00213 (W.D. Tex., filed 2026-05-28) — is publicly analyzed as asserting claim 1 of the '978 patent (alongside US 12,368,799 claim 12 and patents referred to as the "'924" and "'165" patents), with the "Salesforce text message marketing system" as the accused instrumentality. Salesforce's IPR2026-00478 is therefore a direct defensive response to that W.D. Tex. complaint. Source: https://ai-lab-cl-prod.azurewebsites.net/case/dct/txwd/7:26-cv-00213/doc/analysis/1 ; case/date confirmation at https://cdn.pacermonitor.com/public/case/64898612/IQX_Corp_v_Salesforce,_Inc

I could not independently verify the "'165 Patent" referenced by that analysis — it is quoted with the '978 patent's abstract and the identical claim 1 (word-for-word: receive request → customized message with action → transmit → verify authorization data → determine classification → generate RCS message → schedule with ML model). Treat the "distinct '165 patent" as unconfirmed; it may be a sibling continuation with the same abstract or a mislabel in the third-party analysis.

---

IPR2026-00478 — Salesforce, Inc. v. IQX Corp.

• Type: Inter Partes Review (35 U.S.C. §§ 311–319)
• Filed: 2026-09-24 (per the USPTO Open Data Portal structured record; last modified 2026-09-24)
• Status: Pending (verbatim from structured data). Gloss: petition filed and accorded a filing date; the Board has issued no § 314(a) institution decision, and (as of this report) no Patent Owner Preliminary Response is on file. The proceeding is in the three-month preliminary-response window.
• Judge panel: Not yet public. The Board has not issued an institution decision or notice of panel assignment. No APJ names can be stated without fabricating them.
• Petition grounds: Not yet public / not confirmed. I could not retrieve the publicly posted petition or its exhibit list. Because the petition was filed 2026-09-24 (one day before this report), the merits record is not yet indexed. Do not assume which claims (1, the three independents, all claims) or which statutory bases (§ 102 / § 103 / § 112) are at issue — the ground-truth record is the petition itself, available via PTAB E2E at https://ptacts.uspto.gov/ptacts/ and the PTAB public "Petitions" repository at https://ptacts.uspto.gov/ptacts/public-informations. Grounds must be patents or printed publications only (§ 311(b)); § 112 is not available in an IPR, so if the Salesforce team intends a § 101/§ 112 theory it must be litigated in W.D. Tex.
• Institution decision: None issued. Statutory deadline is § 314(b): six months from the petition filing date, i.e. on or about 2027-03-24 (extensions only for good cause shown). Patent Owner's § 323 preliminary response is due three months after the date the notice according the filing date is mailed, i.e. on or about 2026-12-24, unless waived.
• Final Written Decision: None. No claim has been canceled or held patentable. Any statement about claim-level survival would be fabricated at this stage.
• Settlement / termination: None. No § 317(b) termination, no joinder, no adverse judgment.
• Appeal: None, and none is possible yet. An appeal to the Federal Circuit under 35 U.S.C. § 319 lies only from a final written decision; with no institution decision, no FWD, and no notice of appeal, there is no CAFC docket number to report. Earlier searching likewise produced no Federal Circuit 2026 docket involving 12,438,978.
• Defensive value: Low informational value today, high strategic value going forward. For a defendant now facing the '978 patent, IPR2026-00478 confirms that the asserted independent claim is contested and gives you a ready-made prior-art record and prosecution history to monitor — but it provides no estoppel, no cancellation, and no claim construction yet. If you are accused and are not Salesforce (e.g., the E.D. Tex. defendant), the Salesforce petition neither binds you nor helps you: it has not been instituted, so § 315(e)(2) estoppel has not attached even as to Salesforce.

---

Strategic summary

Status of the claims. There is no claim-level outcome to report. Claim 1 (the method claim publicly identified as asserted in IQX v. Salesforce), the "automated messaging subscription and distribution system" independent claim, and the non-transitory computer-readable-medium independent claim are all UNTESTED — not canceled, not sustained, merely challenged in a single pending petition. The abandonment-resumption subject matter that gives the patent its title lives in the specification (Figs. 15–16) and in dependent claims that no tribunal has yet construed. Anyone telling you a claim of the '978 patent is "already dead" is wrong; nothing has been decided.

Estoppel landscape. Because IPR2026-00478 has not been instituted and no FWD has issued, no § 315(e)(2) estoppel has attached to anyone — not Salesforce, not its privies, and certainly not unrelated defendants. Practical consequences for a defendant being asserted today:

• If you want to use the PTAB, you must file your own petition within one year of service of the complaint (§ 315(b)). Salesforce's petition creates no safe harbor for you and no estoppel against you.
• Joinder is available under § 315(c) if IPR2026-00478 is instituted (§ 315(b) bar does not apply to joinder requests granted under § 315(c)), but expect Patent Owner to oppose and expect a close look at whether joinder is "non-anodyne" (whether you would be an active participant).
• If you file your own petition, Sotera-type stipulation practice remains the cleanest way to blunt discretionary-denial arguments — the Director's recent informative decision in Tesla, Inc. v. Bulletproof Property Management, LLC, IPR2026-00204 et al. (Paper, 2026-06-15) denied discretionary denial where there was no scheduled trial date and the petitioner broadly stipulated not to pursue in litigation any ground that could reasonably be raised in the IPR.

Pattern signals. (1) Single-petition campaign so far — Salesforce has filed one IPR against this patent, not a multi-petition volley. (2) No defensive aggregator (no Unified Patents, RPX, or similar) appears in the chain on the '978 patent; unlike the 2013–2020 Applications in Internet Time v. Salesforce saga — where RPX filed and Salesforce was later held to be a real party in interest, terminating the IPRs on § 315(b) grounds — there is no third-party proxy petitioner here. (3) Patent Owner is an active enforcer, not a passive licensee: IQX has now sued at least Intuit (2:26-cv-00351, E.D. Tex.) and Salesforce (7:26-cv-00213, W.D. Tex.), and has filed a later family member (application 19/324,477 → US 2026/0012531 A1). (4) No PTAB appeal history exists for this patent, so there is no signal on whether IQX litigates PTAB losses or settles. (5) The Google Patents family record carries a Darts-IP "family has litigation" flag (family 94175196).

Timing observation. The Salesforce IPR was filed 2026-09-24, roughly four months after the 2026-05-28 complaint — i.e., well inside the § 315(b) one-year bar and consistent with a diligent, litigation-track filing. Expect Patent Owner to argue discretionary denial under § 314(a) based on the co-pending W.D. Tex. case, and expect Petitioner to counter with a Sotera-style stipulation and the Tesla/Bulletproof line of Director precedent.

---

Recommended next steps

1. Pull the actual petition and its exhibit list from PTAB E2E — https://ptacts.uspto.gov/ptacts/ and the public petitions repository https://ptacts.uspto.gov/ptacts/public-informations. Verify (a) exactly which claims are challenged, (b) the § 102/§ 103 grounds and references (grounds will not include § 112), and (c) whether a Sotera or broader stipulation accompanies the petition. Nothing in this report should be used in a brief until that record is confirmed.
2. Calendar the two controlling dates: Patent Owner Preliminary Response due on or about 2026-12-24 (three months from notice according the filing date), and the § 314(b) institution decision deadline on or about 2027-03-24 (six months from petition filing). If instituted, the Final Written Decision is due within 12 months of institution (extendable up to six months for good cause, § 316(a)(11)) — roughly 2028-03 on the current track.
3. If you are a defendant being asserted on the '978 patent: do not wait for Salesforce's petition to resolve. Calendar your own § 315(b) one-year deadline from the date you were served and evaluate whether to (i) file a parallel or joinder IPR, (ii) seek a stay in district court pending IPR2026-00478 (arguably easier to win once institution is granted), and (iii) run a § 101/§ 112 defense in court, since those theories are unavailable in the IPR and are the only validity theories the Board cannot resolve.
4. If you are Patent Owner (or advising it): the immediate deadline is the preliminary response (~2026-12-24). With parallel W.D. Tex. litigation, the highest-value arguments are § 314(a) discretionary denial (Fintiv-type factors plus the Tesla/Bulletproof framing) and any § 325(d) arguments if the references were before the examiner during the long 2018–2025 prosecution chain.
5. Monitor for additional filings: a second petition (from Salesforce or another defendant) or a PGR is plausible given the 2025-10-07 grant date. Note that PGR is available only within nine months of grant (§ 321(c)) — that window closed on or about 2026-07-07 — so IPR and ex parte reexamination are now the only post-grant routes.

---

Sources

• Canonical proceedings data: USPTO Open Data Portal structured "PTAB proceedings on file" record supplied with this task (IPR2026-00478; filed 2026-09-24; status Pending; petitioner Salesforce, Inc.; inventor Sacha Spindler et al.).
• Parallel litigation: IQX Corp. v. Salesforce, Inc., No. 7:26-cv-00213 (W.D. Tex., filed 2026-05-28) — complaint docket at https://cdn.pacermonitor.com/public/case/64898612/IQX_Corp_v_Salesforce,_Inc and complaint analysis (asserting '978 claim 1 against the "Salesforce text message marketing system") at https://ai-lab-cl-prod.azurewebsites.net/case/dct/txwd/7:26-cv-00213/doc/analysis/1
• Board practice referenced: Tesla, Inc. v. Bulletproof Property Management, LLC, IPR2026-00204 et al. (PTAB 2026-06-15) (informative) — https://www.nyipla.org/images/nyipla/sponsors/June%[202026](/patent/202026)%20PTAB%20Committee%20Meeting.pdf ; PTAB July 2026 public hearing schedule — https://www.uspto.gov/sites/default/files/documents/060926-ptab-jul26.pdf
• No Federal Circuit appeal located for US 12,438,978 as of 2026-09-25.

Confidence and limits. High confidence: the existence, number, petitioner, filing date, and "Pending" status of IPR2026-00478, and that no institution decision, FWD, settlement, or appeal exists. Low confidence / explicitly unverified: the challenged claims, asserted grounds and references, real parties in interest, and any stipulation — because the petition was filed one day before this report and its merits record is not yet indexed. I have not stated a panel, a claim-level outcome, or a ground, because none is public.

I'll research the assignment record for US 12,438,978 across USPTO Assignment Center, Google Patents legal events, and related sources.

Let me dig into the specific assignment record and iQX Corp's corporate identity.

Found a second suit. Let me investigate the enforcement campaign and the assignment/prosecution correspondents.

I hit my research-step limit before I could pull the live USPTO Assignment Center reel/frame detail, so I'll flag precisely what is verified versus not. Critically: I did not retrieve a reel/frame number or correspondent name, and I will not invent them.

---

Ownership Record — US 12,438,978 B2

Scope note on data provenance: The interactive USPTO Assignment Center query (https://assignmentcenter.uspto.gov/ and https://assignment.uspto.gov/patent/index.html) could not be executed to completion in this session. What follows is built on (a) the patent's own front matter as published, (b) the Google Patents Legal Events / Reassignment ledger for this patent (which mirrors the USPTO assignment record but does not expose reel/frame in the scraped text I retrieved), and (c) litigation and corporate records. Where a field was not obtained, it is marked NOT RETRIEVED rather than estimated.

---

Inventors

Unusual-pattern check — partially present but benign on this record:

• This patent names only two inventors. Two sibling grants in the same family — US 11,876,924 ("Secure transactions using customized webpages," app. 18/193,163) and US 12,126,752 ("Customer capture using dynamically generated customized webpages," app. 18/483,670) — name three: Spindler, Andrew Lewin, and Wright. So the '978 continuation dropped Andrew Lewin relative to its parent. That is a normal continuation-chain artifact (claims were re-drafted around different subject matter), not evidence of inventorship dispute — but it is the kind of delta a defendant will probe in a § 102(f)/§ 115 challenge.
• I found no evidence that either named inventor departed iQX Corp. within 12 months of filing, and no evidence of a departing-inventor fire-sale. Given that Spindler is the founder, the "all inventors depart pre-sale" pattern is not present.

---

Original assignee

iQX Corp. (also styled "IQX CORP." in USPTO records).

• Primary line of business: Software-as-a-service platform for direct-text (SMS) marketing and sales. Its flagship product, iQX Amplify, launched publicly in January 2017 out of Seattle, WA, as a cloud platform combining message creation/scheduling, demographic capture, and end-to-end consumer purchase journeys executed inside the native text-messaging app, with integrations to WooCommerce, Shopify, Big Cartel, Etsy, BigCommerce, and third parties Twilio, Stripe, Plaid, and Avalara. The specification of this very patent describes the system under the internal name "Amplify" (see the "Amplify 444" / "ADDP 410" architecture passages and the figures).
  • Source: iQX Corp. launch release, Jan. 5, 2017 — http://ww1.prweb.com/prfiles/2017/01/06/13957998/iQX%20Amplify%20News%20-%20Jan%2010%[202017](/patent/202017).pdf
• Did they ship a product embodying the claims? iQX shipped Amplify, and the specification's architecture is Amplify's architecture. However, I could not verify that Amplify remains commercially available as of 2026. That is the single most important open question for the NPE characterization below.
• Corporate address of record — inconsistent across sources: "Beverly Hills, CA" (printed on the face of US 11,765,262 B2), "Carson City, NV" (patents-review inventor/assignee profile), "Seattle, WA" (2017 launch release). Address churn on the front face of patents across a family is a yellow flag observation only — it is not, by itself, a shell-entity finding, and I am not treating it as one.
• Current status: Operating as an active patent plaintiff. Two 2026 filings:
  • IQX Corp. v. Intuit Inc., No. 2:26-cv-00351 (E.D. Tex., Marshall Div.), filed 2026-04-30, Judge Rodney Gilstrap, Magistrate Roy S. Payne, NOS 830, jury demand. Accused product: Intuit MailChimp text-messaging marketing functionality. https://gaeflexstaging-dot-docketupdate.appspot.com/cases/Texas_Eastern_District_Court/2--26-cv-00351/IQX_Corp._v_Intuit_Inc/
  • IQX Corp. v. Salesforce, Inc., No. 7:26-cv-00213 (W.D. Tex.), filed 2026-05-28. https://cdn.pacermonitor.com/public/case/64898612/IQX_Corp_v_Salesforce,_Inc
  • A Civ. P. 7.1 corporate disclosure statement was filed by IQX Corp. in both cases (signed/filed by Reza Mirzaie). I could not retrieve the contents of that disclosure statement — it is the authoritative statement of iQX's current corporate parentage and would settle the "is there a hidden parent?" question. NOT RETRIEVED.
• Bankruptcy: No Chapter 7/11 record for iQX Corp. was found. Caution — do not conflate: a search for "IQX" surfaces IQX Limited, an NSX-listed (Australia) issuer suspended from quotation in March 2022 for failure to lodge accounts (https://live.nsx.com.au/ftp/news/021741236.PDF). That appears to be an unrelated entity; I found no connection to iQX Corp. Flagging it because a careless name-match would produce a false "distress" signal.

---

Assignment timeline

There is one recorded assignment in the chain. The Google Patents Legal Events ledger for US 12,438,978 shows a single reassignment entry:

• 2025-04-29 (execution date NOT RETRIEVED) / recorded 2025-04-29 — Reel NOT RETRIEVED, Frame NOT RETRIEVED
  • Conveyance: Assignment of Assignors' Interest (recorded under the standard label "ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS).")
  • Assignor: SPINDLER, Sacha; WRIGHT, Aidan — i.e., the individual inventors
  • Assignee: IQX CORP.
  • Correspondent: NOT RETRIEVED. I could not obtain the recording correspondent of record. I am not naming a correspondent on inference.
  • Context: Original inventor-to-company assignment, recorded concurrently with the filing of continuation application 19/192,732 (the Google Patents ledger shows "Application filed by IQX Corp." and "Assigned to IQX CORP." on the same 2025-04-29 date). Routine. No company-to-company transfer, no security interest, no merger, no change of name, and no license record appears in this chain.

Post-issuance status: The patent issued 2025-10-07 to iQX Corp., and no post-issuance assignment was recorded in the data I could reach. The patentee-of-record at the time both 2026 suits were filed therefore remains iQX Corp. — the original operating company.

Family continuity context (no ownership change): The pre-grant publication US 2025/0274544 A1 (2025-08-28) and the follow-on family member 19/324,477 → US 2026/0012531 A1 (2025-09-10) are the same applicant family, iQX Corp.

Bottom line for this section: Because no reel/frame was retrievable, treat the reel/frame and correspondent fields as open diligence items, not as absent records. Do not report "no assignments" — there is exactly one, and it is the founder/inventor→iQX Corp. original assignment.

---

Timeline diagram

timeline
    title Ownership of US 12438978
    2018 : Provisional filed by Spindler and Wright
    2019 : PCT application filed
    2021 : US national stage entered
    2023 : Continuation filed
    2024 : Further continuations filed
    2025 : Continuation filed as 19-192732
         : Inventors assign to iQX Corp
         : Patent issued to iQX Corp
    2026 : iQX sues Intuit in E D Tex
         : iQX sues Salesforce in W D Tex

---

NPE / troll-pattern signals

---

Verdict

Operating-company assertion.

The chain of title contains exactly one recorded conveyance — the 2025-04-29 inventor→iQX Corp. assignment recorded concurrently with the filing of application 19/192,732 — and no post-issuance transfer, so the patentee of record when iQX filed against Intuit (2026-04-30) and Salesforce (2026-05-28) is the same company that built and marketed the amplified text-marketing platform the specification describes. None of the classic NPE-construction signals (shell transfer, known asserter, cascading LLCs, pre-litigation transfer, bankruptcy sale) is present, so an NPE verdict would require evidence I do not have. Two explicit caveats: (i) the reel/frame and correspondent are NOT RETRIEVED, and (ii) I could not confirm that iQX Amplify is still commercially available in 2026 — if Amplify is in fact discontinued, iQX is better characterized as a decommissioned operating company enforcing a continuation farm, a pattern that is functionally NPE-like even without a shell transfer. Retrieving the Rule 7 corporate disclosure statements from both 2026 dockets is the single highest-value next step.

Verify at: https://assignment.uspto.gov/patent/index.html (search patent number 12,438,978) and https://assignmentcenter.uspto.gov/ — pull the reel/frame and the recorded correspondent, which I was unable to obtain here.

I'll search for the USPTO record and the "References Cited" (patent citations) for US 12,438,978.

Let me try to retrieve the granted-patent front page and its "References Cited" list directly from additional sources.

Prior-Art Analysis — US 12,438,978 B2 ("Systems and methods for resuming online abandoned journeys")

---

⚠️ Threshold finding you need before anything else

I could not retrieve a "References Cited" (§ 56) list for US 12,438,978 — and I have a specific reason to believe the patent may have a very short or empty one.

The authoritative full text of US 12,438,978 that was supplied to me (fetched from https://patents.google.com/patent/US12438978/en, 2026-09-25) contains: Info, Classifications, Definitions, Landscapes, Abstract, and Description. It contains no "Patent Citations," "Cited By," or "(56) References Cited" section at all. My subsequent targeted searches — for the grant PDF (patentimages…US12438978.pdf), for Espacenet's "List of all citations" view, and for the FPO/Patents-Review citation tables — did not return a citation list for this patent either. The final three searches were cut off when I exhausted my tool budget, so this is a negative finding, not a positive verification.

Therefore I am not going to invent a (56) list. Per the operative rule, where I cannot confirm something I will say so rather than fabricate. Below I (a) tell you exactly which documents are cited of record in the patent's own text, (b) tell you where the real citation list lives and why it is probably thin, and (c) give you a separately labeled candidate-art analysis keyed to the claim limitations — which is what you actually need for a § 102/§ 103 assessment.

Methodological flag on how I label things: Section 3 below is not "the patent's citations." It is my own identification of potentially relevant art, and I mark confidence per item. Do not present Section 4 items as the examiner's or applicant's citations without pulling the front page.

---

1. The only references cited of record on the face of this patent

The patent's own CROSS-REFERENCE section cites exactly one category of documents — its own priority chain (all incorporated by reference). These are not § 102 prior art against US 12,438,978, and I list them only so you don't mistake them for art:

The Google Patents timeline also lists a later family member, App. Ser. No. 19/324,477 → US 2026/0012531 A1 (2025-09-10). Same-family; not art.

---

2. Why the (56) list is likely thin, and where to get the real one

Two structural reasons:

1. This is a continuation (not a CIP) of 18/984,041. Continuations typically inherit the IDS/citation history of the parent, and examiners frequently issue ex parte grants in familial chains with minimal new art — particularly for a patent granted on a 2025 filing with a 2018 priority claim.
2. The Google Patents render omits both "Patent Citations" and "Cited By." Where Google Patents shows neither, it usually means few or no front-page references were printed.

Authoritative sources to pull the actual (56) list (I could not complete these retrievals):

• USPTO PatentCenter → Application 19/192,732 → "Documents" → the PTO-892 Notice of References Cited and the IDS (PTO/SB/08) as filed, plus the PTOL-90 / "References Cited" front page of the printed patent.
• USPTO Patent Public Search (ppubs) → "12438978" → "References" tab.
• Espacenet → https://worldwide.espacenet.com/patent/search?q=pn%3DUS12438978 → "Cited documents."
• PCT/US19/52847 International Search Report (this is the search report that actually examined the shared specification, and it will have a Box C list).

---

3. Analytical observation: the effective filing date may be far later than the 2018 priority date

Flagging this because it materially changes which art is available:

US 12,438,978's claims recite RCS messaging, an ML/AI classification-and-scheduling engine (AMDA), and silent-network authentication (SNAP). On the face of the specification, that subject matter appears to have entered the disclosure via the CIP filed 2024-09-25 (Ser. No. 18/896,197) — the parent continuation chain back to 2019/2018 describes the "Amplify/ADDP/UEMWEB/hot-cookie" architecture, but the AMDA tiering and RCS-scheduling material reads as CIP addition. (Figs. 9–18 are concentrated in the AMDA/SNAP/abandonment disclosure.)

If the '978 claims are not entitled to the 2018-09-27 / 2019-09-25 priority dates for this subject matter, their effective filing date is 2024-09-25, which:

• opens ~6 years of intervening art as § 102(a)(1) prior art;
• eliminates the § 102(b)(1)(A) grace-period exception for the family's own earlier publications (see § 6).

This is an inference from document structure, not a verified priority determination (which requires the actual priority applications, which I do not have). Treat it as a lead. The definitive record is the priority applications themselves and any claim of priority in the file wrapper.

---

4. Candidate prior art — my identification, keyed to claim limitations

Below, for each: full citation, date, brief description, and (a) the claim(s) it could implicate. I have separated § 102 (anticipation — single reference discloses all elements) from § 103 (obviousness — combination), because on this claim set almost nothing is a clean § 102 reference.

Claim scaffolding used (from the supplied text; independent claims are not numbered in my source):

• Ind. Claim 1 (method): (1) receive request for content → (2) generate customized message with ≥1 action → (3) transmit to mobile device → (4) on action selection, request authorization data → (5) verify data correlates to recipient/device identity → (6) determine a classification → (7) generate RCS message with the content → (8) schedule delivery using a machine learning model and based on the classification.
• Ind. system claim: same, with "service subscriber"/"subscription data."
• Ind. CRM claim: same, with "link" and "authorized recipient or subscriber."
• Dependent claims: deliver per schedule; carousel (transaction / appointment / resume abandoned journey); browser UI; typing indicators; sender verification + SMS warning; actions = transaction / suggested response / return to prior browsing session / return to prior application session.

---

4a. Abandoned-journey / form-and-cart re-engagement art

4b. Session-resumption art (browsing / application session transfer)

4c. The limitation-defining art — RCS messaging + SNAP identity + ML send-time scheduling

This is where I found no verified § 102 reference, and where I believe the novelty actually sits.

---

5. Bottom-line § 102 assessment

1. Independents (method / system / CRM): no clean § 102 reference identified. A reference would have to disclose, in one document, the entire chain — customized action message → silent identity verification tied to the device → tier classification → RCS message → ML-scheduled delivery. References A1–A5 and B1 each cover a fragment. The realistic attack is § 103, combining, e.g., A1 or B1 (abandonment re-engagement) + A4/A5 (cookie/cart preservation and abandonment tracing) + an RCS messaging reference + an ML send-time-optimization reference.

2. Dependent claims are the exposed ones. The claims reciting "return to a prior browsing session," "return to a prior application session," and "resuming at least one abandoned user journey" are squarely in the zone of A1, A2, A5 and B1. Those are the claims where § 102 is genuinely arguable.

3. The specification's own admissions help an attacker. The '978 specification:

   • describes RCS as an existing protocol ("according to the RCS messaging protocol");
   • describes SMS marketing messages, links, carousels, typing indicators, and read receipts as existing RCS/SMS features;
   • frames the problem (determining "how and when to present content … journey resumption") in terms of a pre-existing industry problem.

   Those statements are usable as applicant-admitted prior art against the broader claims.

---

6. ⚠️ Self-collision risk: WO 2020/068911 A1 (this is the one to brief on)

The parent PCT (PCT/US19/52847, filed 2019-09-25) published as WO 2020/068911 A1 on 2020-04-02 — per the Google Patents record supplied to me ("2019-09-25 Priority claimed from PCT/US2019/052847 … patent/WO2020068911A1/en").

If, as suggested in § 3, the RCS/AMDA/SNAP claim subject matter is only supported from the 2024-09-25 CIP onward, then:

• WO 2020/068911 A1 (2020-04-02) is more than one year before the effective filing date, so the § 102(b)(1)(A) grace-period exception for the inventors' own disclosure would not rescue it; and
• common ownership does not cure a § 102(a)(1) printed-publication problem.

Whether this is a live issue turns entirely on priority entitlement, which I cannot resolve without the priority applications. I flag it because it is the single highest-value thing to check in the file wrapper, and because it is invisible if you only look at third-party art.

(Note: I have deliberately not treated the disclosures in IQX's own US 12,368,799, US 11,876,924, or US 2025/0274544 A1 as anticipatory — they are the same family/inventive entity and overwhelmingly fall under common-ownership or shared-priority treatment. Verify any specific date relationship before relying on them.)

---

7. Confidence and caveats

High confidence:

• The priority chain and family documents listed in § 1 (from the patent's own CROSS-REFERENCE section).
• That the supplied authoritative text of US 12,438,978 contains no (56) citations list.
• The claim-element decomposition used in § 4.
• That the RCS protocol itself is admitted prior art by the specification.

Medium confidence:

• The relevance/description of A1 (US 10,992,513 B1) — I saw only a substantive excerpt, not the front page; the number and its bibliographic data are not verified by me.
• The relevance of A2–A5 — these descriptions come from a PATROLL petition and a Unified Patents exhibit describing a different patent (US 9,286,282); I did not pull each reference's own front page.
• That B1 (US 2018/0336897 A1 / US 10,269,351 B2 / EP 3504620 B1) is the Google automated-assistant dialog-resumption family — the EP abstract strongly matches, but I did not confirm the US member's claims.

Not verified / do not assume:

• That US 12,438,978 has any particular set of cited patents. I could not retrieve the PTO-892, the IDS, the PTOL-90 front page, or the PCT/US19/52847 ISR. Get these before writing any opinion.
• That any listed reference anticipates any specific numbered claim. My source does not contain a numbered claim set, so mappings are to independent claim types and to dependent-claim subject matter, not to "claim 7" or "claim 12."
• The exact publication date of WO 2020/068911 A1 — inferred from the WO number series and the Google Patents note; I did not open the document.

Discrepancy flagged, not corrected: The task states the current date is April 26, 2026, but the previously generated litigation section is dated 2026-09-25 and reports IQX complaint filings on April 30, 2026 and May 28, 2026 — both after the stated present date. I have reported all dates literally and normalized nothing.

---

Recommended next actions (in order of value)

1. Pull the PTOL-90 front page of US 12,438,978 from USPTO PatentCenter (App. 19/192,732) to obtain the actual (56) list and replace Section 4's candidate list with the verified citation list.
2. Pull the PCT/US19/52847 ISR (WO 2020/068911 A1, Box C) — it is the only search report that examined the shared specification and will show what the ISA considered the closest art.
3. Resolve priority entitlement for the RCS/ML/SNAP limitations (compare claims of the '978 against the 2019 and 2018 disclosures) — this determines whether six years of intervening art is available.
4. Then re-run § 102/§ 103 against the verified citation list plus targeted RCS-messaging and ML-send-time-optimization art.

I'll validate and extend the § 103 analysis with targeted searches for art covering the three limitation clusters that the prior-art section flagged as the novelty locus (RCS delivery, silent-network identity verification, and ML send-time scheduling).

Strong results on the RCS, SNAP, and ML-send-time clusters. Let me pin down a few specific references and their dates for the motivation-to-combine analysis.

Obviousness Analysis — U.S. Patent No. 12,438,978 B2

"Systems and methods for resuming online abandoned journeys" (iQX Corp.)

35 U.S.C. § 103 · Report date 2026-09-25

---

0. Scope, provenance, and what this section supersedes

Provenance. This § 103 analysis is built on (a) the authoritative full text of US 12,438,978 supplied earlier in this analysis, (b) the claim-element decomposition and prior-art candidates developed in the Prior Art section (references A1–A5, B1–B4, and the WO 2020/068911 A1 self-collision flag), and (c) four clusters of art newly retrieved in this session (Microsoft/Google patent-document PDFs, Google Patents, FreePatentsOnline, Patents-Review, Patsnap, and IPqwery listings). Unless stated otherwise I am working from quoted abstract/claim/specification text, not from a retrieved front page.

Three explicit updates to the Prior Art section. That section concluded that the novelty locus sat in three uncharted clusters and stated so for each: (i) RCS-as-delivery-channel — "I did not find one"; (ii) ML-based send-time scheduling — "No verified art identified — this is likely the point of novelty"; and (iii) silent carrier-network identity verification — "No verified third-party § 102 reference identified." All three conclusions are now superseded. This session located verified disclosure for each cluster (detailed in § 3 below). That does not by itself make the claims obvious — it does mean the § 103 attack is no longer a stretch built on fragments, and any brief written on the strength of the Prior Art section's § 4c table should be rewritten before filing.

Legal framing used. Graham v. John Deere Co., 383 U.S. 1 (1966) (scope/content of prior art; differences; PHOSITA level; secondary considerations); KSR Int'l Co. v. Teleflex Inc., 550 U.S. 398 (2007) (predictable combinations; market forces and design incentives; "obvious to try" with a finite number of identified, predictable solutions; a technique used to improve one device may be applied to improve a similar device); MPEP § 2143 (rationales (A)–(G)); In re Keller, 642 F.2d 413 (CCPA 1981) (test is what the combined teachings suggest, not bodily incorporation).

One rule-mandated flag repeated, not corrected. The stated present date of April 26, 2026 is inconsistent with the previously generated sections, which are dated 2026-09-25 and report IQX complaint filings on 2026-04-30 and 2026-05-28 plus an IPR filed 2026-09-24. I report all dates literally and normalize nothing.

---

1. The threshold issue that controls everything: the priority-date fork

§ 103 cannot be analysed until the critical date is fixed, and the patent's own structure makes this genuinely contestable. As flagged in the Prior Art section (§ 3), the RCS/AMDA/SNAP subject matter of the independent claims appears to have entered the disclosure via the 2024-09-25 CIP (Ser. No. 18/896,197), whereas the parent chain back through 17/279,188 → PCT/US19/52847 → 62/737,315 carries the older "Amplify / ADDP / UEMWEB / hot-cookie" architecture. The claims on their face recite RCS, a machine-learning classification-and-scheduling model, and silent network authentication.

That produces two very different worlds:

Practical consequence: the single highest-value work product in this case is a priority/§ 112 written-description analysis of the 2019 and 2018 disclosures versus the '978 claims. If the RCS/ML/SNAP limitations are only supported from the 2024 CIP, six additional years of art open up and the obviousness case becomes substantially easier. Everything below is therefore presented under both scenarios with art availability marked accordingly.

---

2. Element-by-element road map of independent claim 1

Element labels follow the Prior Art section's decomposition. "Availability" is per the Scenario A / Scenario B table.

Immediate observation: under Scenario B, every element of claim 1 is met by no more than three references, and two of them (1.7 and 1.8) are met by art whose entire purpose is business-messaging optimisation. Under Scenario A, elements 1.7 and 1.8 are the hard ones — and RCS itself is admitted prior art by the '978 specification, which states content/messages "may be generated and provided according to the RCS messaging protocol to enhance and modernize an SMS or an MMS communication."

---

3. The reference set (with confidence flags)

Dates marked "unverified" are taken from the retrieval text; I did not retrieve a front page for every item. Do not brief a date I have flagged without confirming it.

Cluster I — RCS / rich business messaging

Cluster II — ML send-time / classification

Cluster III — abandoned cart / session resumption (mostly pre-2018)

---

4. Combination I — the primary § 103 case (Scenario B / post-CIP effective date)

This combination is designed to be run in the pending IPR and mirrored in district court. It uses only patents and printed publications (§ 311(b)-compliant).

The combination

P1 (US 2021/0304150 A1) or P2 (US 2022/0366130 A1) [RCS generation, template/carousel, scheduling rules, SMS/MMS fallback]
+ P6 (LivePerson US 2021/0383399 A1) [brand-messaging platform, device rich-messaging/payment capability determination, cart + expiry, RCS payment request]
+ P7 (LivePerson AU 2024203463 A1) [proactive outbound campaigns expressly covering abandoned carts, across SMS/RCS/RBM/OTT, keyed to the customer's phone number]
+ P9 (US 2019/0213476 A1) or P10 (US 2021/0374801 A1) or P11 (US 11,431,663 B2) [ML-predicted send time based on historical engagement, with user clustering]
+ P19 (WO 99/56434 A1) or P20 (US 11,196,738 B2) or P14 (US 2024/0205338 A1) [carrier/SIM/MSISDN-based silent identity verification returning a pass/fail that gates continuation in the application]

Element-by-element

Why a PHOSITA would have combined them (MPEP 2143 rationales)

1. Same field, same problem, same actors (2143(A), (C)). Every reference sits in the business-messaging / e-commerce re-engagement field and addresses the identical problem the '978 specification itself frames: "determining how and when to present content, messages, opportunities, user journeys, user journey resumption to a user." P7 expressly names abandoned carts as a proactive-messaging use case; P6 expressly builds the cart + brand + customer identifier + expiry object; P1/P2 build the interactive RCS card. Combining a re-engagement trigger, an RCS renderer and a carrier identity check is the aggregation of art "designed to address the same or similar problems" — the classic Keller combination.

2. Simple substitution of a known, better channel (2143(B)). P2 supplies the express motivation: RCS is "intended to replace SMS," gives native read receipts and interactive buttons, and — critically — for SMS "a business can only send a SMS message with a link to a landing page … which may decrease the likelihood that a user will access the content because of the extra step of clicking the link." A PHOSITA optimising click-through on an abandoned-cart message has, in P2's own words, a stated reason to prefer RCS over SMS. That is a substitution to obtain a predictable improvement, which is the KSR paradigm.

3. The two-step SMS-then-RCS sequence is dictated by known capability limits, not by invention. P1 teaches the exact mechanic — "phone RCS capabilities … are checked; message 246 is relayed to an MMS or SMS gateway if mobile device 204 does not support RCS" — and P6 builds an entire device-capability service plus presence ping for rich messaging. A PHOSITA knowing RCS penetration was incomplete would necessarily send the universal-reach link first and the rich message once capability and identity are confirmed. No inventive insight is required; the sequence is the arithmetic of a partially-deployed standard.

4. Known technique applied to a known device ready for improvement (2143(C), (D)). P9, P10 and P11 each independently disclose ML-based send-time prediction from historical engagement. P9 even discloses user clustering keyed to the send schedule, which is precisely the claim's "determine a classification … and schedu[le] … based on the determined classification." Where a mature technique (send-time optimisation) exists and a new message class (RCS re-engagement) needs scheduling, applying the known technique is obvious per In re ICON Health & Fitness and KSR's "use of a known technique to improve similar devices in the same way."

5. Market forces and regulatory pressure (2143(F) / KSR). The '978 specification itself recites FCC and TCPA compliance and "regulatory blackout periods" as a system constraint (AMDA hub 1000e, ML/AI cluster 1000a). Art that exists because of send-time regulation and deliverability economics — P10's "frequency optimization, channel optimization" and P11's explicit motivation to reduce network/compute overhead from mis-timed sends — supplies the commercial and technical pressure to adopt ML scheduling. P13 independently confirms the same architecture (probability of open/view/interaction → argmax over candidate times).

6. Obvious to try, finite and predictable solutions (MPEP 2143(E)). For element 1.8 the field presented a small, enumerated set of proven approaches — RNN/survival analysis (P9), shared-training/siloed-inference ML (P10), two-layer NMF (P11), clustering models (P12). Selecting one to schedule an RCS re-engagement message is the KSR "obvious to try" situation: "a finite number of identified, predictable solutions," with "a reasonable expectation of success."

7. Identity binding is an obvious security improvement, and the patent says so. The '978 specification asserts that the technique "may address security flaws in existing approaches (such as SIM-case swapping or fraudulent use of a link)." The prior art already solved that problem: P19 (1999) confirms a telephone subscription number / SIM / IMEI against an authentication server and only then permits the user interface "to continue the proceedings in the application"; P20 uses carrier-network confirmation specifically "to prevent channel-based cyber-attacks"; P14 uses silent MNO verification and compares a verified MSISDN to an access MSISDN. Adding a carrier-identity gate to a resumable deep link is the known remedy for the known hazard — 2143(A) and KSR.

Result: every element of claim 1 is disclosed or rendered obvious; the combination is a set of references teaching the same thing in the same way for the same purpose, with a stated motivation in the references themselves (P2's RCS-over-SMS rationale; P1's capability-check/fallback; P7's abandoned-cart use case; P9/P10/P11's express resource-saving rationale).

---

5. Combination II — the fallback § 103 case (Scenario A / 2018 priority preserved)

If IQX establishes entitlement to 2018-09-27, the 2020–2024 RCS and ML art drops out and the attack must be rebuilt on pre-2018 art. That is materially harder — but not impossible, and two features of the record make it viable.

The combination

P16 (US 2003/0144925 A1) or P17 (US 2005/0278231 A1) or P18 (US 2008/0133349 A1) [save abandoned cart/session; re-engage; email-launched resume transaction]
+ P22/A1 (US 10,992,513 B1) [track user progress, determine intent/obstacle, select a communication channel to re-establish the session]
+ P19 (WO 99/56434 A1) [carrier/SIM/telephone-number-keyed authentication gating continuation]
+ P21 (US 2018/0336897 A1) [§ 102(a)(2) art: effectively filed 2017 — selectable element that reopens the stored session, so it is available despite publishing after the critical date]
+ P9 (US 2019/0213476 A1) [if its effective filing date precedes the critical date — this must be verified; if it does, it is § 102(a)(2) art]
+ RCS as admitted prior art [specification: content "may be generated and provided according to the RCS messaging protocol to enhance and modernize an SMS or an MMS"; carousels, typing indicators, read receipts all treated as existing RCS features]

Why this still works under Scenario A

• The channel-substitution motivation survives without P2. Even in 2016–2018, RCS was a published GSMA Universal Profile with interactive cards and suggested replies. A PHOSITA building a re-engagement message and needing interactive actions had RCS available, and the '978 specification admits it is a pre-existing protocol offering richer content than SMS. Applicant admissions are usable art (In re Nomiya; MPEP 2144.03).
• A1 supplies the missing "choose the channel" step expressly — "selects a communication channel to re-establish the communication session" — which directly supports the multi-channel and SMS-fallback dependent claims.
• Adobe '476 (P9) is the pivot. Its disclosure maps to both element 1.6 (cluster grouping keyed to send schedule) and element 1.8 (send time from a predicted engagement metric). If its effective filing date predates 2018-09-27 it carries the whole scheduling limitation under Scenario A. Verify this first; it is the single most consequential date in the Scenario A case.

Honest assessment

Under Scenario A, element 1.7 ("generating a Rich Communication Service (RCS) message") is the weakest link. Escaping it requires (i) proving that P9 or another pre-critical-date reference teaches RCS as the delivery channel, or (ii) relying on the specification's own admission that RCS is conventional, which invites the patentee to argue that the claimed combination — not RCS per se — supplies the invention. Expect this to be a genuine fight under Scenario A and a non-issue under Scenario B. That asymmetry is the entire reason priority resolution should be briefed first.

---

6. Combination III — the dependent claims (obvious over the same set, with additions)

These are where the Prior Art section predicted § 102 (not just § 103) exposure, and the newly retrieved art strengthens that materially. None of the dependent limitations adds an element that is not itself a known feature of business messaging.

---

7. Reasonable expectation of success and "predictable results"

The combination produces, at most, the expected sum of its parts: (i) a message the recipient can act on (known from P2/P3), (ii) delivered over the richer channel when the handset supports it (known from P1/P6), (iii) gated by carrier identity confirmation (known from P19/P20), (iv) at a time predicted to maximise engagement (known from P9/P10/P11). Nothing in the '978 specification identifies a synergistic or unexpected result from combining these; the specification asserts benefits in general terms ("enhancing the effectiveness and efficiency of messaging campaigns … in terms of security, user engagement, and operational scalability") without data, ablations, or comparisons. Where the patentee offers no evidence of unexpected results, the "predictable results" rationale carries substantial weight (In re Geisler line; MPEP 2144.05 and 2143.02).

Two further points reinforce a reasonable expectation of success:

• Every reference is a complete, working system. Each of P1, P6, P7, P9, P10, P11, P19 is described at a level of detail sufficient to implement the sub-function it contributes (JSON payloads, capability pings, training features, auth-server tables). There is no "mere aggregation of old elements in a new box" gap — the references are combinable as written.
• The combination requires no change in the principle of operation of any reference — the classic In re Ratti / bodily-incorporation objection does not arise.

---

8. What the patentee will argue, and how it likely fares

---

9. Procedural notes that shape how this § 103 case should be deployed

• The pending IPR (IPR2026-00478, Salesforce, filed 2026-09-24, status Pending) is framed by § 311(b) — patents and printed publications only. Combination I above is IPR-compliant as constructed. The petition's actual grounds, challenged claims, and any Sotera stipulation are not public and I have not verified them. My combinations are built independently and may or may not overlap the petition; do not assume alignment.
• No estoppel has attached (no institution, no FWD). A non-Salesforce defendant (e.g., the E.D. Tex. Intuit defendant, if it later asserts the '978 patent) retains its own § 315(b) one-year window from service and cannot rely on Salesforce's petition for anything.
• § 103 is the only substantive validity theory that travels into the IPR. § 101 and § 112 must be run in district court. A PGR is no longer available — the nine-month § 321(c) window from the 2025-10-07 grant closed on or about 2026-07-07.
• Watch the § 325(d) angle. The '978 patent issued in 2025 out of a long continuation chain (2018→2025). If the examiner never saw P1–P15, § 325(d) cuts for the challenger; if some were in the family's IDS, Patent Owner will argue § 325(d) and the prior-art section's still-unretrieved (56) citation list becomes decisive.

---

10. Confidence and limitations

High confidence: the substance of P1, P2, P3, P4, P6, P7, P8, P9, P10; the claim-element decomposition; the MPEP 2143 / KSR rationales; the observation that the '978 specification admits RCS, carousels, typing indicators and read receipts as existing features; and the conclusion that no single reference anticipates claim 1 (consistent with the Prior Art section), while the combination does.

Medium confidence: the bibliographic dates and assignees I marked "unverified," in particular US 11,431,663 B2's assignee, US 2019/0213476 A1's effective filing date (the pivot for Scenario A), and the LivePerson / TeleSign / ADP family priority dates that determine whether each item is § 102(a)(1) or only § 102(a)(2) art. EP 4672253 A1, AU 2025283509 A1, US 2025/0384998 A1 and EP 4513926 A1 appear to post-date a 2024-09-25 critical date and should be treated as background/state-of-the-art evidence rather than § 102 art unless a pre-critical-date priority is confirmed.

Not verified / do not assume: the actual priority entitlement of the '978 claims; the (56) references-cited list (still unretrieved); the contents of IPR2026-00478; whether US 12,438,978 is in fact asserted in IQX v. Salesforce (a third-party summary, medium confidence, and the "'165 patent" labelled with the identical claim 1 remains unexplained — see the PTAB section); and the identity of the "'924" and "'165" patents.

Noted, not corrected: the stated present date of April 26, 2026 versus the 2026-09-25 report date and the 2026-04-30 / 2026-05-28 / 2026-09-24 filing dates in the earlier sections.

---

11. Recommended next steps, in priority order

1. Resolve priority/§ 112 for the RCS + ML-scheduling + SNAP limitations against the 2018 provisional and the 2019 PCT/national-stage disclosure. This single question converts the case from Combination II (contestable) to Combination I (strong). If the answer is "2024-09-25," serve the § 103 contentions immediately.
2. Verify US 2019/0213476 A1's effective filing date. It is the load-bearing reference for element 1.6 + 1.8 under Scenario A.
3. Pull the real (56) list (PTOL-90 / PTO-892 / IDS from App. 19/192,732 and the PCT/US19/52847 ISR) and re-map § 325(d).
4. Retrieve the IPR2026-00478 petition and exhibits to align or deliberately diverge from Salesforce's grounds and to see whether a Sotera stipulation is on file.
5. Build claim charts for Combination I against claims 1 and the three independents, then a second chart for the dependents (carousel, typing indicators, sender-verification + SMS warning, prior-session-return) using P1, P2, P3 and P21.
6. Commission a PHOSITA declaration that fixes the level of ordinary skill (a messaging-platform/backend engineer with 2–4 years' experience in SMS/RCS A2P campaign systems would suffice) and walks the four elementary MPEP 2143 rationales — same-field/same-problem, channel substitution per P2, capability-driven fallback per P1, and known-technique improvement per P9/P10/P11.
7. Preserve the § 101 / § 112 theories for district court, since neither is available in the IPR.