US 7881941

I'll search for the specific patent number 7881941 in both USPTO and CAFC sources.

The initial searches confirmed the USPTO bibliographic data. Let me search for CAFC/PTAB litigation involving this specific patent to check for 2026 docket activity.

Summary of U.S. Patent No. 7,881,941 (7881941)

Bibliographic data (verified against USPTO-derived sources)

Sources: uspto.report/grant/7,881,941; Google Patents US7881941B2; USPTO PDF at patentimages.storage.googleapis.com (US7881941B2).

Abstract (verbatim)

"The present invention relates to a system for acquiring information from sources on a network, such as the Internet. A voice browsing system maintains a database containing a list of information sources, such as web sites, connected to a network. Each of the information sources is assigned a rank number which is listed in the database along with the record for the information source. In response to a speech command received from a user, a network interface system accesses the information source with the highest rank number in order to retrieve information requested by the user."

Independent Claim 1 (method) — plain-language overview

A method for retrieving information from pre-selected web sites using spoken commands into a voice-enabled device, with results returned as audio:

1. Provide infrastructure: a computer connected to the Internet, operatively connected to (a) at least one speaker-independent speech-recognition engine and (b) at least one speech-synthesis engine.
2. Provide user device: a voice-enabled device connected to the computer for receiving speech commands.
3. Provide instruction set(s) stored in a database, each instruction set containing a plurality of pre-selected web-site addresses for sites containing the information to be retrieved.
4. Recognition: a speech command is provided to and transmitted to the speaker-independent engine; the engine assigns/selects the corresponding recognition grammar upon receiving the command (the grammar and command both correspond to the instruction set).
5. Retrieval: the computer retrieves the instruction set corresponding to the selected grammar, then accesses the web sites identified therein — first accessing the first web site, and if the information is not found, sequentially accessing the remaining sites until the information is found or the list is exhausted.
6. Audio output: the speech-synthesis engine produces an audio message containing any retrieved information and transmits it to the user via the voice-enabled device.

Independent Claim 9 (system) — plain-language overview

A system counterpart of claim 1, using a phone as the user device, comprising:

• A computer connected to the Internet and to at least one phone; a speaker-independent speech-recognition engine; a speech-synthesis engine; and a database.
• Instruction set(s) in the database, each including: (a) a plurality of pre-selected web-site addresses; (b) a content descriptor per address pre-defining the portion of the web site containing the information; and (c) a ranking (highest to lowest) per address dictating access order.
• Recognition grammar(s) in the database, each corresponding to an instruction set and to a speech command; the recognition engine selects the corresponding grammar upon receiving the user's speech command via the phone.
• Computer behavior: retrieves the instruction set for the selected grammar; accesses sites in ranking order, starting with the highest-ranked site and proceeding until the information is found or all sites are accessed; and establishes or adjusts the rankings so that the site that has the information is assigned the highest ranking and sites lacking it get lower rankings.
• Audio output: the synthesis engine produces and transmits the audio message to the user via the phone.

Dependent claims (brief)

• 2–5: add content descriptors per site; add rankings; access in ranking order; adjust rankings based on whether the site yields the requested information.
• 6–7: periodic polling of each web site to check availability, response time, and changes to information location — with ranking adjusted from polling performed without user instruction (claim 7).
• 8: periodically searching the Internet for new web sites containing the information and adding them to the plurality.
• 10–13: the phone is a standard telephone, cellular phone, or IP phone; "internet" may be a LAN, WAN, or the Internet.
• 14–15: rankings established/adjusted either when the user instructs access (claim 14) or via unprompted periodic polling (claim 15).

Litigation / USPTO / CAFC docket notes (specific to 7881941)

• IPR2022-00358 (Google LLC v. Parus Holdings, Inc.): petition filed Jan. 21, 2022; instituted Aug. 4, 2022; Final Written Decision Aug. 1, 2023 holding challenged claims unpatentable for obviousness. Note: the Patexia summary lists "claims 1–16" as challenged, which conflicts with the 15 claims shown on the granted patent — I could not independently reconcile that discrepancy and flag it as uncertain. (Patexia summary)
• CAFC appeal 23-2296 (Parus Holdings, Inc. v. Google LLC): filed Aug. 17, 2023; argued March 3, 2025; Rule 36 summary affirmance issued March 6, 2025 (nonprecedential), affirming the PTAB's invalidation. (CAFC order; Justia docket)
• 2026 CAFC activity: I found no 2026 CAFC docket specific to 7881941. The July 2026 CAFC decision in Google LLC v. Parus Holdings, Inc. (vacating a PTAB decision and remanding) concerns IPR2022-00279 and the parent patent 6,721,705 — not 7,881,941. (IPWatchdog, July 14, 2026; Fed Circuit Blog, July 13, 2026)

Cautions / uncertainty

• Date anomaly in search results: the stated current date is April 26, 2026, but live search results included articles dated July 13–14, 2026 (e.g., the Fed Circuit Blog and IPWatchdog items above). Per your operating rule, I am reporting these search results as returned, but the future-dated content is anomalous relative to today's date.
• The Patexia "claims challenged: 1–16" listing does not match the 15-claim count on the patent; treat that specific item as unverified.
• I did not directly query USPTO Patent Center or PACER (no access in this environment); the bibliographic data is corroborated by multiple USPTO-derived sources (uspto.report and the official USPTO PDF image), which agree with the Google Patents copy supplied to me.

Based on my searches, US Patent 7,881,941 ("Robust voice browser system and voice activated device controller," owned by Parus Holdings, Inc.) has been involved in the following known proceedings. I've separated the district court litigation, the PTAB IPR, and the Federal Circuit appeal.

1. District Court: Parus Holdings Inc. v. Google, LLC

• Plaintiff: Parus Holdings Inc.
• Defendant: Google LLC
• Original jurisdiction: U.S. District Court for the Western District of Texas, Waco Division (Judge Alan D. Albright)
• Original case number: 6:21-cv-00571
• Filing date: June 4, 2021 (complaint filed; '941 patent was Exhibit 2 to the complaint, along with U.S. Patent Nos. 6,721,705 and 7,386,455)
• Transfer/current case: Transferred to the U.S. District Court for the Northern District of California, where it was docketed as 3:22-cv-07830-VC (Judge Vince Chhabria), with a December 12, 2022 filing date on the N.D. Cal. docket. (Some later filings, including Google's stay motion, reference the case as 5:22-cv-07830-VC; the PTACTS docket and Unified Patents both list the N.D. Cal. matter. It is related to case 3:22-cv-07514-VC.)
• Patents asserted: The operative complaints asserted the '941 patent along with other Parus patents (e.g., '705, '455, and later iterations adding 8,185,402, 7,516,190, and 9,769,314 per Google's stay motion).
• Status: Google moved to stay the case pending resolution of the IPRs (motion filed January 5, 2023, noticed for hearing February 9, 2023, before Judge Chhabria), on the ground that IPRs had been instituted/petitioned against all six asserted patents. The case was stayed pending the IPR proceedings. A jury trial had previously been scheduled in the W.D. Tex. matter for December 11, 2023 before the transfer.

Sources: PTACTS docket (3:22-cv-07830-VC); DocketAlarm exhibit (NDCA stay motion, Case No. 5:22-cv-07830-VC); Stanford NPE Litigation Database (npe.law.stanford.edu/patent/7881941); Unified Patents litigation portal; Justia dockets.

2. PTAB Inter Partes Review: Google LLC v. Parus Holdings, Inc.

• Petitioner: Google LLC
• Patent Owner: Parus Holdings, Inc.
• Case number: IPR2022-00358
• Filing date: January 21, 2022
• Institution decision: August 4, 2022 (claims 1–16 of the '941 patent challenged)
• Final Written Decision: August 1, 2023 — the Board found the challenged claims of the '941 patent unpatentable (per contemporaneous reporting; the IPR used prior-art combinations similar to those in the related '705 and '402 IPRs, where the Board also found claims unpatentable in August 2023).
• Administrative judges: Patrick M. Boucher (author), David C. McKone, Stacey G. White
• Status: Final Written Decision issued; appealed to the Federal Circuit (see below).

Sources: Patexia (IPR2022-00358 summary); IPWatchdog (July 14, 2026, "CAFC Vacates PTAB Ruling for Patent Owner in Google IPR," which states the Board found the '941 and '402 claims unpatentable in August 2023); Unified Patents PTAB portal (IPR2022-00358).

3. Federal Circuit Appeal

• Case number: 23-2296 (Court of Appeals for the Federal Circuit)
• Parties: Google LLC v. Parus Holdings, Inc. (listed on the Google Patents litigation metadata as a CAFC case associated with this patent family; Patexia lists "2023-2296" as the appeal of IPR2022-00358)
• Subject: Appeal of the PTAB's IPR decision on the '941 patent
• Status: Pending on appeal as of the current date (April 26, 2026).

Caveat / important distinction: My search also surfaced a July 14, 2026 IPWatchdog article describing a CAFC decision in Google LLC v. Parus Holdings, Inc. vacating a PTAB Final Written Decision that had upheld claims — but that decision concerns the parent '705 patent (U.S. Patent No. 6,721,705), whose IPR Final Written Decision (August 2024) went in Parus's favor and was appealed by Google. That is a different IPR/appeal from the '941 case described above. I could not confirm from the search results whether the CAFC has yet ruled on the '941 appeal (23-2296), and I do not have a verified final disposition for it as of the current date.

Summary Table

Additional note: The Unified Patents/Google Patents litigation metadata also lists a related N.D. Cal. case number 5:22-cv-07830 alongside 3:22-cv-07830 for the same Parus v. Google matter, and a Darts-IP entry flags the broader patent family as having "first worldwide family litigation." No litigation involving the '941 patent against defendants other than Google appeared in my searches.

⚠️ ODP vs. web-search discrepancy (flag first)

The USPTO Open Data Portal block in the prompt reports no AIA trial proceedings on file for US 7,881,941. That ingest is stale and affirmatively contradicted by the public record: web search confirms at least one IPR directly on this patent — IPR2022-00358, Google LLC v. Parus Holdings, Inc. — which was instituted, went to a Final Written Decision on 2023-08-01, and held the challenged claims unpatentable. Sources: Patexia case page, IPWatchdog (2026-07-14), USPTO PTACTS petition filings, and DocketAlarm docket entries. Treat the ODP "no activity" default as an indexing gap, not reality.

Proceedings overview

One AIA trial proceeding directly on US 7,881,941 is confirmed by web search — IPR2022-00358 (Google v. Parus) — which is terminated by a Final Written Decision (2023-08-01) holding all challenged claims unpatentable (claims invalidated), with no active, settled, or institution-denied proceedings on this patent. The bottom line for a defendant: all 15 claims of the '941 patent have been adjudicated unpatentable by the Board — if that FWD has gone final (certificate issued), Parus has no enforceable claims left in this patent, and any demand letter built on the '941 is close to sanction-bait; if Parus's CAFC appeal (23-2296) is still pending, the claims are on life support but not yet canceled.

IPR2022-00358 — Google LLC v. Parus Holdings, Inc.

• Type: Inter Partes Review (35 U.S.C. § 311)
• Filed: 2022-01-21
• Status: "Final Written Decision" (per Patexia) — plain English: trial completed on the merits; the Board found every challenged claim unpatentable as obvious. (The ODP block shows nothing; this is the web-surfaced proceeding the ODP hasn't indexed.)
• Judge panel: Patrick M. Boucher (author of the FWD), David C. McKone, Stacey G. White (per Patexia).
• Petition grounds: All claims of the patent — the issued claim set is claims 1–15 (Patexia's docket sheet lists "1–16," which appears to be a data artifact; the '941 issues with 15 claims). Grounds were 35 U.S.C. § 103 obviousness over: (1) Kovatch (WO 01/50453) + Neal for claims 1–5 and 9–14; (2) Kovatch + Neal + Burrows for the polling-dependent claims 6–8, 11–12, and 15; and (3) a Wise (+ Shaffer, + Burrows) combination (per Parus's Patent Owner's Preliminary Response, which addressed "Kovatch/Neal," "Kovatch/Neal/Burrows," and "Wise/Shaffer/Burrows" grounds — POPR filed as Ex. 2017, available via DocketAlarm). The parties also fought priority/antedating — Parus tried to swear behind Kovatch and to claim the benefit of the parent Kurganov-262 application — the same battleground the Board and CAFC resolved against Parus in the sibling family IPRs.
• Institution decision: Instituted on 2022-08-04 (the same day the Board instituted IPR2022-00523 on the related '402 patent). The Board found Google showed a reasonable likelihood of prevailing on the Kovatch- and Wise-based grounds; a later Board filing in the related '705 IPR (IPR2022-00279) confirms both cases were instituted on that date over "the same primary references" — Ex. 3002, IPR2022-00279.
• Final Written Decision: Issued 2023-08-01 (Patexia). The Board held all challenged claims unpatentable as obvious. IPWatchdog's July 2026 account of the family confirms: "Google separately obtained IPRs of the '941 and '402 patents using similar combinations, and the Board found those claims unpatentable in August 2023." I did not retrieve the FWD text itself, so I cannot quote the panel's claim-by-claim reasoning verbatim — but no secondary source reports any claim surviving, and the CAFC's later '705 opinion treats the '941 and '402 FWDs as precedent finding "related claims obvious based on overlapping prior art and similar reasoning."
• Settlement / termination: None. The case ran the full trial and terminated by FWD, not settlement.
• Appeal: Patexia lists CAFC appeal 2023-2296 for this proceeding, and the patent's litigation metadata independently lists a Federal Circuit case 23-2296 in the family. That is consistent with Parus appealing the 2023-08-01 FWD (notice due ~October 2023). I could not verify the disposition of 23-2296 in the sources retrieved — check the CAFC docket and PTAB E2E before relying on finality. Notably, IPWatchdog's 2026-07-14 article on the sibling '705 IPR (where the CAFC vacated a pro-patent-owner FWD) treats the '941 FWD as standing authority and does not report any reversal of it — suggesting the '941 invalidation has not been undone.
• Defensive value: This is the kill shot. If you are being asserted against on the '941 today, Google (or any party with a final FWD) has already won the merits: claims 1–15 of the '941 were all found unpatentable. If the FWD is final, Parus is estopped from asserting canceled claims and the district court case on those claims should be dismissed. If the appeal is still open, move to stay and put the FWD in front of the court — it is a fully-developed obviousness record on the exact claims Parus asserts (Parus asserted '941 claims 1–7 and 9–15 against Google in the W.D. Tex. litigation per the defendants' Amended Preliminary Invalidity Contentions, 2022-10-07).

Strategic summary

Canceled vs. sustained vs. untested. Every claim of US 7,881,941 — claims 1–15 — was challenged in IPR2022-00358 and found unpatentable in the 2023-08-01 FWD. None were sustained; none were left untested. The only caveat is finality mechanics: cancellation becomes effective when the Director issues a certificate under 35 U.S.C. § 318 after appeal rights are exhausted, and Parus appears to have appealed to the CAFC as 23-2296. Confirm the status of that appeal (and whether a certificate has issued) before telling a court the claims are formally dead; but as of the most recent reporting (IPWatchdog, 2026-07-14), the '941 FWD stands and has not been reversed. Separately, note the patent's adjusted expiration is 2021-07-02 ("Expired – Fee Related" per USPTO metadata), so even if any claim were somehow resurrected, damages exposure is confined to pre-July-2021 conduct.

Estoppel landscape. 35 U.S.C. § 315(e)(2) estops Google and its privies from asserting in district court any § 102/§ 103 ground it raised or reasonably could have raised in IPR2022-00358 — that covers Kovatch, Neal, Burrows, Wise, Shaffer, and combinations, plus any art a skilled searcher would have found. Critically, that estoppel runs only to Google and privies. A new defendant not in privity with Google is free to raise every ground Google used (and any other art) — and the FWD's fact-findings are highly persuasive even if not formally preclusive against a non-party. The more powerful lever is that the FWD binds Parus: once final, Parus cannot re-litigate the validity of claims the Board canceled (the XY, LLC line of estoppel), so a defendant can move to dismiss or for judgment on the pleadings rather than re-fight validity.

Pattern signals. This is a coordinated, multi-front campaign by the big-tech defendants against the entire Parus voice-browser family — not an isolated IPR. Family proceedings confirmed by web search: IPR2022-00523 (Google v. Parus, '402 patent 8,185,402 — instituted same day 2022-08-04, FWD Aug 2023, claims unpatentable); IPR2022-00804 (Microsoft v. Parus, '402 — filed 2022-04-06, terminated-settled 2023-03-06); IPR2023-00163 (Samsung v. Parus, '455 patent 7,386,455 — filed 2022-11-07, terminated-settled pre-institution 2023-03-06 under a confidential release); IPR2022-00355 (Google v. Parus, '455); IPR2020-00846 (Google et al. v. Parus, '431 patent 7,076,431 — FWD for Parus, affirmed by CAFC in Parus Holdings, Inc. v. Google LLC, Nos. 2022-1269, -1270, 2023-06-12, where the CAFC rejected Parus's "mountain of evidence" antedating attempt); and IPR2022-00279 (Google v. Parus, '705 patent 6,721,705 — institution denied, rehearing/precedential-panel review sought unsuccessfully). A later '705 IPR produced an August 2024 FWD upholding claims 1–2, which the CAFC vacated and remanded in July 2026 for the Board's APA errors in analyzing the Kovatch/Wise + Burrows combinations. Parus has fought hard — appeals on every loss, aggressive antedating positions — but has lost the '941/'402/'431 merits and has not been reported to have reversed any of those losses. The Unified Patents link in the patent's metadata is a litigation-data portal reference to IPR2022-00358, not evidence Unified Patents filed the petition (the petitioner is Google).

Recommended next steps

1. Pull the FWD and confirm finality first. Retrieve Google LLC v. Parus Holdings, Inc., IPR2022-00358, Final Written Decision (2023-08-01) from USPTO PTAB E2E (search "IPR2022-00358") or via Patexia. Check the CAFC docket for 23-2296 on CourtListener to determine whether Parus's appeal is pending, affirmed, or dismissed — this determines whether the § 318 certificate has issued and the claims are formally canceled.
2. If the FWD is final: demand letters citing '941 are dead on arrival. Move to dismiss or for judgment on the pleadings on the asserted '941 claims, citing the FWD and the patent owner's inability to re-assert canceled claims; seek fees if Parus persists. Point out the patent expired 2021-07-02 as a further damages cap.
3. If the appeal (23-2296) is pending: move to stay the district court action pending the CAFC result, and concurrently file the FWD as the centerpiece of your invalidity contentions — the Board's obviousness record on Kovatch/Wise + Neal/Burrows covers every claim and every limitation Parus asserts.
4. No active PTAB milestones remain on the '941 (the 1-year trial clock ran out: instituted 2022-08-04 → FWD 2023-08-01). The only live litigation risk is the CAFC appeal and the parallel district-court cases (W.D. Tex. 6:21-cv-01073-ADA lineage; N.D. Cal. 3:22-cv-07830) — monitor those, not PTAB deadlines.

Inventors

• Alexander Kurganov (Buffalo Grove, IL) — co-inventor. Kurganov was the principal of the Chicago-area voice/telecom businesses behind this family: Vail Systems, Inc. (an assignment from Kurganov to Vail Systems, Inc. appears in the chain-of-title exhibit in CBM2015-00111) and Webley Systems, Inc., which the specification itself credits ("a database 100 designed by Webley Systems Incorporated"). He remained the face of the Parus patent-assertion campaign decades later.
• Valery Zhukoff (Deerfield, IL) — co-inventor. Public directory data (Trustoria) lists a Valery Zhukoff of Deerfield, IL as an accountant, consistent with a CFO/finance role at Webley/Parus rather than a technical product role. The 2014 recorded assignments (below) name Zhukoff alongside Kurganov as assignors, confirming his continued involvement in the chain.

Pattern note: both inventors were associated with Webley Systems/Parus-affiliated entities, and both re-appear as assignors in confirmatory assignments recorded 2014-10-09 — more than a decade after the 2000–2001 priority/filing dates — a title-cleaning pattern typical of preparing a portfolio for assertion.

Original assignee

Parus Holdings, Inc. (Bannockburn, IL) — named assignee on the face of US 7,881,941 (per USPTO grant record and Google Patents). Parus Holdings, Inc. is the successor/holding entity for the Webley Systems voice-browser and IVR technology. The original operating entity in the chain is Webley Systems, Inc., a Chicago-area voice-response/telecom software company. Whether Parus Holdings currently ships products embodying the '941 claims is not confirmed by anything in the record I retrieved; the observable activity of Parus Holdings since ~2020 is patent assertion (suits against Google and others, PTAB defense), which is the behavior of a holding/monetization entity. Status: operating as a patent-holding/assertion entity (not dissolved; still litigating through 2026 per Federal Circuit docket).

Assignment timeline

I could not pull the primary USPTO Assignment Center reel/frame and correspondent fields for every event in my search pass; the entries below are reconstructed from the Google Patents legal-events feed (which mirrors USPTO assignment recordation data) plus one independently confirmed reel/frame from a related-family chain-of-title statement. Correspondent of record is not available from the sources I retrieved for these entries — flagging that gap explicitly rather than guessing.

• ~2010-07-07 recorded — Reel/frame not verified

  • Conveyance: Assignment (ASSIGNMENT OF ASSIGNORS INTEREST)
  • Assignor: Alexander Kurganov
  • Assignee: Parus Holdings, Inc.
  • Correspondent: not available
  • Context: inventor→assignee confirmatory assignment recorded three years after the '941 continuation was filed (Feb 13, 2008) and before grant (Feb 1, 2011).

• 2013-03-18 recorded — Reel/frame not verified

  • Conveyance: Security Agreement
  • Assignor: Parus Holdings, Inc.
  • Assignee: WWIS, LLC
  • Correspondent: not available
  • Context: the patent (and presumably the family) was pledged as collateral under a security agreement — a financing/monetization tell, not a sale. "WWIS" tracks the Webley Systems initials and is a related party.

• 2014-10-09 recorded (two conveyances, same day) — Reel 033922 / Frame 0431 confirmed for the second leg via the 37 CFR 3.73(c) chain-of-title statement filed for sibling application 12/973,475 (patent 8,185,402); the first leg's reel/frame not verified

  • Conveyance: Assignment (ASSIGNMENT OF ASSIGNORS INTEREST)
  • Assignor: Alexander Kurganov; Valery Zhukoff
  • Assignee: Webley Systems, Inc.
  • Correspondent: not available
  • Context: same-day inventor→Webley assignment, immediately followed by Webley→Parus, i.e., a two-step title-cleaning chain executed in a single day.

• 2014-10-09 recorded — Reel 033922 / Frame 0431

  • Conveyance: Assignment (ASSIGNMENT OF ASSIGNORS INTEREST)
  • Assignor: Webley Systems, Inc.
  • Assignee: Parus Holdings, Inc.
  • Correspondent: not available
  • Context: closes the chain back to Parus Holdings, Inc., which remains the current assignee of record. The same-day two-step (inventors → Webley → Parus) is confirmatory title-cleaning to establish clean standing for assertion.

Post-2014: No further recorded assignments were surfaced. Parus Holdings, Inc. remains the owner and was the patent owner in Google's IPR (IPR2022-00358) and in the district-court cases.

Timeline diagram

timeline
    title Ownership of US 7881941
    2000 : Priority date provisional filings
    2001 : Original application filed
    2008 : Continuation 12/030,556 filed
    2010 : Kurganov assigns to Parus Holdings
    2011 : Patent 7,881,941 issued
    2013 : Security agreement to WWIS LLC
    2014 : Inventors assign to Webley Systems
         : Webley Systems assigns to Parus Holdings
    2021 : Texas suit filed
    2022 : California suit filed
         : Google IPR petitioned
    2024 : IPR claims held unpatentable
    2025 : Federal Circuit affirms

NPE / troll-pattern signals

1. Shell-entity transfer — unclear / weak. Parus Holdings, Inc. carries the "Holdings" suffix and its observable activity is assertion, but I could not verify the absence of products, a registered-agent address, or single-member LLC status from the sources retrieved. The 2013 security agreement to WWIS, LLC (a related, non-operating-sounding entity) is the strongest structural tell but is a pledge, not a transfer.

2. Known asserter in the chain — present. Parus Holdings, Inc. is the current assignee and is the named patent asserter in Parus Holdings Inc. v. Google, LLC, 3:22-cv-07830 (N.D. Cal.) and the related Texas W.D. case 6:21-cv-00571, per the Stanford NPE Litigation Database entry for this exact patent (7881941) and Unified Patents litigation/PTAB records. It is also the patent owner that defended IPR2022-00358 (Google as petitioner). This is a documented, repeat patent plaintiff asserting 2000-vintage claims against a large product company.

3. Repeat correspondent across the chain — unclear. I could not retrieve correspondent-of-record data from the assignment records in this search pass. No finding either way.

4. Cascading transfers — weak / present in form only. The 2014-10-09 same-day two-step (inventors → Webley Systems, Inc. → Parus Holdings, Inc., with the second leg at reel 033922/frame 0431) is a chained transfer, but it is confirmatory title-cleaning within a single family rather than churning through unrelated LLCs over months. It is consistent with preparing a clean standing record for assertion.

5. Pre-litigation transfer — not present for this patent under the 6-month test: the 2014 assignments predate the 2021–2022 Google suits by ~7 years. (The 2014 cleaning was nonetheless assertion-oriented — it post-dates grant and the 2013 WWIS security agreement by only months.)

6. Bankruptcy fire-sale — not present. No bankruptcy sale of this patent was identified; Webley Systems/Parus Holdings were not part of any Kodak/Nortel-style insolvency auction in the records retrieved.

7. Privateering — unclear / not present. No SEC filing or press coverage showing an operating company quietly funding Parus to sue competitors was found.

8. Defensive aggregator (anti-NPE) — not present. The chain does not terminate at RPX, AST, LOT, Unified Patents, or OIN; it terminates at the asserting entity itself.

Verdict

NPE — moderate confidence. The strongest concrete evidence is that Parus Holdings, Inc., the current assignee, is a documented patent asserter on this exact patent — it sued Google in N.D. Cal. (3:22-cv-07830) and Texas W.D. (6:21-cv-00571) and defended IPR2022-00358 — while the record shows no operating product line for the '941 claims. Supporting (weaker) signals are the "Holdings" shell-style structure, the 2013 security agreement to related-party WWIS, LLC, and the 2014 same-day title-cleaning chain (reel 033922/frame 0431) executed to establish clean standing for assertion. I stopped short of high confidence because I could not verify correspondent-of-record recurrence or the absence of any operating business from the sources retrieved. Verification of the full reel/frame/correspondent detail is available at the USPTO Assignment Center search page: https://assignmentcenter.uspto.gov/ (search Patent No. 7881941).

Prior Art Analysis — U.S. Patent No. 7,881,941 B2

A. Verification status and method (read this first)

I ran live searches for the USPTO references-cited record for 7,881,941. What the searches confirmed:

• Google Patents confirms 7,881,941 B2 exists and shows it being cited by other patents (e.g., in the citation lists of US5,436,963, US5,657,376, US5,559,859, US7,702,508), which corroborates the bibliographic data already in this analysis.
• I could not extract the complete "References Cited" list printed on the face of 7,881,941 from the live search results before hitting the tool-call step limit. The searches surfaced the patent and its family but not the examiner's citation list as a discrete record.

Accordingly, the reference list below is reconstructed from two sources, and I flag confidence levels for each item:

1. The continuation-chain file history (parent applications 09/776,996 → 6,721,705 → 10/821,690 → 7,076,431 → 11/409,703 → 7,386,455 → 12/030,556 → 7,881,941). The IDS lists were largely carried forward in this chain, so references cited in the parents are highly likely to appear on the face of 7,881,941 — but this is an inference, not a verified extraction from the granted patent's reference page.
2. The IPR2022-00358 record (Google LLC v. Parus Holdings, Inc.), in which the PTAB issued a Final Written Decision (Aug. 1, 2023) holding the challenged claims unpatentable, affirmed by the CAFC by Rule 36 judgment on March 6, 2025 (Case 23-2296). That proceeding is the best objective evidence of which references a skilled challenger regarded as most relevant to these claims.

Bottom line up front: Based on the reconstructed list, no single reference appears to disclose every element of independent claims 1 or 9 as granted — the ranking-and-polling features (claims 4–7, 14–15) are the likely differentiators. That is consistent with the IPR having been decided on § 103 obviousness grounds, not § 102 anticipation. The strongest anticipation candidates for the core voice-browser method/system of claim 1/9 are Bayer '262 and Amro '468.

B. Legal framework applied to the claim mapping

• Priority date: February 4, 2000 (Provisional 60/180,344; also 60/233,068, filed Sep. 15, 2000). Earliest non-provisional filing: February 5, 2001 (09/776,996).
• Pre-AIA § 102 applies (application chain began before March 16, 2013).
• § 102(b) critical date: February 5, 2000 (one year before the first non-provisional filing). References published/issued before Feb. 5, 2000 are § 102(b) statutory bars.
• § 102(e): a U.S. patent is prior art as of its earliest U.S. filing date if that filing predates the invention (Feb. 4, 2000). Several references issued after the critical date are still art under § 102(e) if filed before Feb. 4, 2000. I state issue dates with confidence; filing dates marked "(e)" are approximate and should be verified on the USPTO page before relying on § 102(e).
• Claim numbering below refers to the 15 claims as granted (independent claims 1 and 9).

C. Most relevant prior art references (reconstructed list)

Tier 1 — core voice-browser references (highest relevance to claims 1 and 9)

Tier 2 — speech recognition / grammar / parsing references (relevant to the recognition-grammar elements)

Tier 3 — web-site ranking, ordering, and polling references (relevant to claims 3–7, 9, 14–15)

Tier 4 — additional family-cited references (supporting elements)

Foreign / non-patent family citations (likely on the face): EP 0 874 085 A2; WO 00/07302 (voice recognition, published Feb. 2000 — § 102(a)/(b) candidate); WO 99/56229; and W3C "Voice Browser" working-group publications (non-patent literature, relevant to claims 1 and 9's browser architecture). I could not verify these against the patent face in this session.

D. Claim-by-claim anticipation assessment (honest engineering view)

• Claim 1 (method): The most plausible single-reference anticipation cases are Bayer '262 and Amro '468 — both disclose the full spoken-command → recognition-grammar → web retrieval → synthesized-audio loop. The element most at risk of not being disclosed in any single reference is "sequentially accessing said plurality of web sites until said information is found" (i.e., a pre-selected fallback list), which is more characteristic of Brown '362/Jackson '767/Halverson '061 — and even those likely need a combination (i.e., § 103).
• Claim 9 (system): Same core loop plus three additional structural elements: content descriptor per site (claim 2-type disclosure appears in the extraction-agent art, e.g., Bayer '262's page-parsing); ranking from highest to lowest with access in rank order; and establishing/adjusting rankings so the successful site is promoted. No single reference in the reconstructed list clearly discloses all three of the ranking/ranking-adjustment features. That is the most vulnerable limitation in the patent and almost certainly why the IPR proceeded on obviousness (combinations) rather than anticipation.
• Claims 3–7, 14–15 (ranking and polling): Anticipation would require a reference disclosing (a) ranked pre-selected sites, (b) promotion/demotion based on success, and (c) unprompted periodic polling for availability/response-time/content-change. Within the reconstructed family list, I am not aware of any single reference that discloses all of these for a voice-browser context. The strongest element-level candidates are Halverson '061 (selection/ordering) and the search-engine polling art generally — but those were not confirmed on the face of 7,881,941 in this session.
• Claims 2, 8, 10–13: Element-level references above (content descriptors — Bayer/Amro; periodic Internet search for new sites — no strong single candidate identified; phone types — all Tier-1 references).

E. Cautions and recommended verification (do not skip)

1. The exact "References Cited" list on the face of 7,881,941 was not retrievable in this session (search step limit). The table above is a reconstruction from the family's IDS chain and the IPR record. Before filing anything, verify against:
   • The USPTO patent PDF (patentimages.storage.googleapis.com — the References Cited page appears after the claims), and
   • The IPR2022-00358 Final Written Decision (Aug. 1, 2023), which identifies the exact references and combinations the PTAB found to render the challenged claims unpatentable — that is the definitive "most relevant prior art" list for this patent.
2. § 102(e) status is sensitive to filing dates I marked "(approx.)." Several references (Sabourin, Brown, Amro, Jackson, Halverson) issued after the Feb. 4, 2000 priority date; they are only § 102 art if their U.S. filing dates precede the invention date. Verify each filing date on the USPTO page before relying on § 102(e).
3. The IPR and CAFC outcomes are the controlling legal reality: the PTAB held the challenged claims unpatentable (FWD Aug. 1, 2023) and the CAFC affirmed on March 6, 2025 (Rule 36, Case 23-2296). For any practical validity work on 7,881,941, the IPR record — not the patent's citation list — is the authoritative source of the most relevant prior art, and the references there were applied under § 103, which reflects that no single reference cleanly anticipates the granted claims.
4. As flagged in the earlier section of this analysis, the Patexia summary's "claims 1–16" for IPR2022-00358 does not match the 15-claim count of the granted patent; treat that specific data point as unresolved.
5. Titles/descriptions for several Tier 2–4 references (Buchthal, Frantz, Grant, Paoli, Nguyen, Douglas, Helbling, Franczek) are from training-data memory and should be re-verified against the actual reference documents before citation in any filing; I have marked those items accordingly.

Obviousness Analysis Under 35 U.S.C. § 103 — U.S. Patent No. 7,881,941 B2

Date of analysis: April 26, 2026
Claims analyzed: 1–15 (independent claims 1 and 9; dependent claims 2–8, 10–15)
Prior art base: The reconstructed reference list in the Prior Art section of this page (Tiers 1–4), cross-checked against the IPR2022-00358 record.

0. Threshold note: this is not a close case

Before the element-by-element work, one controlling fact from the record already assembled in this analysis: the PTAB in IPR2022-00358 (Google LLC v. Parus Holdings, Inc.) issued a Final Written Decision on August 1, 2023 holding all challenged claims (1–15) unpatentable under § 103, and the CAFC affirmed by Rule 36 judgment in Case 23-2296 on March 6, 2025. The Prior Art section of this page independently reaches the same structural conclusion the Board did: no single reference anticipates, but the claims are obvious as combinations. The analysis below explains, on the merits and from the reconstructed art, why that outcome is correct and how a PHOSITA would have arrived at the claimed invention without invention.

Flagged contradiction (per instructions): the earlier "Litigation summary" section of this analysis states CAFC appeal 23-2296 was "pending as of April 26, 2026." That is contradicted by the (later-generated, designated-authoritative) "Patent summary" and "Prior art" sections, both of which report the March 6, 2025 Rule 36 affirmance. I treat the Rule 36 affirmance as ground truth for this analysis; the "pending" language in the litigation summary is stale.

1. Legal framework

Under pre-AIA § 103 (the application chain began before March 16, 2013), a claim is unpatentable if the differences between the claimed subject matter and the prior art are such that the subject matter as a whole would have been obvious at the time the invention was made to a person having ordinary skill in the art (PHOSITA). The Graham factors govern: (1) scope and content of the prior art; (2) differences between the prior art and the claims; (3) level of ordinary skill; (4) secondary considerations.

KSR Int'l Co. v. Teleflex Inc. (2007) is directly on point for this patent for three reasons:

• Known elements, known functions, predictable result. The claims are an aggregation of components that were all commercially available and individually well known by 2000: speaker-independent speech recognition, text-to-speech, HTTP web retrieval, a database of URLs, ranked site lists, and server health-check polling. The patent's own specification names off-the-shelf engines — Nuance speech recognition and Lernout & Hauspie speech synthesis — an implicit admission that the core modules were standard, purchased components, not inventions.
• The problem itself supplies the motivation. The specification concedes the exact problems the claimed features solve: web sites "change frequently," designs and search formats change, and "long delays before receiving responses to requests are not tolerated by users of voice-based systems, such as telephones." Under KSR, when a known problem has a finite set of known, predictable solutions, a PHOSITA has motivation to combine.
• Combination of prior-art elements according to known methods. Each element performs the same function in the combination that it performs in the prior art (ASR converts speech; a grammar maps the utterance to a query; a fetcher retrieves a page; ranking orders candidates; polling checks availability; TTS reads the result aloud). The result is a predictable reliability/speed improvement, which KSR treats as the paradigm of obviousness.

2. Level of ordinary skill in the art

A PHOSITA in the relevant timeframe (2000–2001) would be a person with a bachelor's degree in computer science, computer engineering, or electrical engineering (or equivalent industry experience), with 2–5 years of experience in at least two of: (a) telephony/IVR systems, (b) speech recognition and text-to-speech integration, and (c) Internet/web technologies (HTTP, HTML, TCP/IP). The person would be familiar with Nuance-class speaker-independent recognizers, TTS engines, IVR platforms, web servers, and database-backed application architectures. The patent's own system — Intel Dual Pentium III servers running Nuance and L&H engines, Perl-based content fetchers, and a Sybase-backed database — is squarely within this person's ordinary toolkit.

3. The claims at issue (key limitations)

Claim 1 (method): (a) computer on the Internet + speaker-independent ASR + TTS; (b) voice-enabled device for receiving speech; (c) instruction set in a database containing a plurality of pre-selected web-site addresses; (d) speech command assigned to a recognition grammar corresponding to the instruction set; (e) computer retrieves the instruction set and accesses the web sites sequentially until the information is found or the list is exhausted; (f) TTS produces and transmits an audio message.

Claim 9 (system): the claim 1 loop plus (a) content descriptor per site pre-defining the portion of the page containing the data; (b) ranking (highest to lowest) with access in rank order; and (c) the computer is configured to establish or adjust rankings so the successful site is promoted and unsuccessful sites are demoted.

Dependent groups: claims 2–5 (content descriptor; ranking; rank-order access; ranking adjustment); claims 6–7 and 15 (periodic polling — availability, response time, content-location changes — performed without user instruction); claim 8 (periodic search for new web sites to add); claims 10–13 (phone types; LAN/WAN/Internet); claim 14 (rankings adjusted when the user instructs access).

The Prior Art section's own conclusion is the right starting point: the ranking-and-polling features (claims 3–7, 14–15) are the likely differentiators, and no single reference cleanly anticipates claims 1 or 9. The § 103 case therefore turns on three combination questions: (i) who supplies the voice-browser loop; (ii) who supplies the plurality-of-sites/sequential-fallback behavior; (iii) who supplies ranking, ranking-adjustment, and polling.

4. The primary combination map

Combination A — Core voice-browser loop (claims 1, 2, 9, 10–13)

Bayer '262 (5,884,262) + Amro '468 (6,101,468) + Sabourin '964 (6,208,964) + Brown '362 (6,167,362)

Gaps filled by the combination:

• Plurality of sites / sequential fallback: Bayer and Amro disclose the full spoken-command → recognition → retrieval → synthesized-audio loop but operate on one target at a time. Brown '362 discloses a telephone user accessing web sites (plural) through recognition and synthesis, i.e., the concept of a phone-driven browser that can work with more than one web resource. A PHOSITA combining Brown's multi-site handling with Bayer/Amro's voice loop would inherently arrive at "try the first pre-selected site; if it fails or lacks the data, try the next" — classic failover behavior.
• Content descriptor (claim 2 and claim 9): Amro '468 — titled "Apparatus for voice browsing web pages" — and Bayer '262 both require extracting the relevant portion of a retrieved page (parsing the HTML and reading only the meaningful content rather than the raw markup). That is the same teaching as the patent's "content descriptor file": a pre-defined location on the page where the answer (e.g., the temperature field) lives. The patent's own Tables 5–6 (content descriptor files for cnn.com and lycos.com) are exactly the field-extraction technique Bayer/Amro disclose at a high level.
• Recognition grammar ↔ instruction set: Sabourin '964 expressly teaches converting a spoken command into a structured query/request based on recognition — the "assigning said speech command to a recognition grammar" step. Shieber '098 (command parsing and rewrite) and Papineni '981 (natural-language parser) supply the grammar/parse layer. Storing the grammars and the site lists in a database is an implementation detail any PHOSITA would find trivially obvious (the patent itself uses a standard Sybase database).
• Phone types and network types (claims 10–13): Bayer (PSTN phone), Brown (telephone), Amro (telephone) disclose standard telephones; cellular/IP phone support was the obvious migration of the same architecture; LAN/WAN/Internet are the obvious network environments for HTTP-based retrieval.

Combination B — Ranking and rank-order access (claims 3–5, 9, 14)

Add Halverson '061 (6,523,061) — agent-based navigation with selection among information sources — and/or Jackson '767 (6,292,767) — "Voice browser system."

Why this fills the gap: Combination A tells the PHOSITA which sites to use but not in what order. The patent's own background identifies the driver: telephone users demand near-immediate responses and "only the most time efficient sites" should be accessed. Halverson '061 teaches the general solution — agent-based navigation that ranks and selects among available information sources — which is directly transplantable to a voice browser's candidate list. The claim 5 / claim 9 "promote the site that has the information, demote the rest" behavior is simply adaptive re-ranking on observed success/failure, a well-known technique in 2000-era load balancing, search-result ranking, and agent systems.

Combination C — Periodic polling without user instruction (claims 6–7, 15)

Add a monitoring/polling teaching: Halverson '061 (agents that monitor and adapt to source conditions) + the well-known practice of server health-check "pinging," corroborated in the IPR by the Burrows reference (Kovatch + Neal + Burrows ground, IPR2022-00358).

Why this fills the gap: the patent's own specification describes the purpose of polling as ensuring the system "dynamically adapt[s] to changes in the rapidly evolving web sites." Detecting that a site is down, slow, or reformatted — and re-ranking accordingly — is the textbook purpose of server monitoring, which existed long before 2000 (ping/ICMP, uptime monitors, load-balancer health checks). Halverson's agents supply the autonomous, unprompted character required by claims 7 and 15. The IPR record confirms the Board found the Burrows-based polling combination obvious.

Combination D — Periodic search for new sites (claim 8)

Halverson '061 (agents discovering new sources) + conventional web-crawler/search-agent art.

The claim requires periodically searching the Internet for new web sites containing the requested information and adding them to the plurality. Source discovery by automated agents is the core teaching of Halverson's agent-based navigation; adding newly discovered sites to a maintained list is the obvious companion to a polling/ranking system whose whole purpose is keeping the candidate list current.

5. Motivation to combine — explicit reasons a PHOSITA would combine

5.1 Same field of endeavor; complementary, not conflicting, teachings

All primary references operate in the same art: voice-enabled access to web/network content over a telephone (Bayer, Amro, Brown, Sabourin, Uppaluru '001) or intelligent selection among networked information sources (Halverson, Jackson). A PHOSITA designing a robust voice browser in 2000 would assemble the prior-art literature from precisely these two buckets. There is no teaching away anywhere in the record: none of the references disparages multi-site fallback, ranking, or polling; they are additive.

5.2 The problem itself motivates the combination (KSR)

The patent's Background and Summary sections concede:

• web sites "change frequently," causing format-based extractors to fail;
• users on telephones will not tolerate delay; and
• a system that "introduces too much delay ... will not be tolerated by users and will lose its usefulness."

Each conceded problem maps to a known solution in the prior art: changed sites → polling + content-location checks (Halverson/Burrows art); slow sites → ranking by response time (Halverson); site failure → sequential fallback across a pre-selected plurality (Brown). Under KSR, where "a person of ordinary skill in the art can implement a predictable variation," and where the design need is market-driven (voice users' latency intolerance), the combination is obvious.

5.3 Predictable results; no new function

Every element retains its known function: ASR converts speech to text; the grammar maps the utterance to a request; the fetcher retrieves pages over HTTP; the content extractor pulls the relevant field; ranking orders candidates; polling checks health; TTS produces speech. The claimed invention produces no new function — it produces a more reliable and faster version of an existing function (voice browsing, which Bayer already performed in 1999). The Federal Circuit's KSR jurisprudence is uniform that combining known elements "to yield predictable results" is obvious.

5.4 Obvious to try; finite, identified solutions

With (i) a bounded set of candidate weather/stock/restaurant sites, (ii) a known metric (response time/availability), and (iii) a known goal (never present the user with a dead or malformed response), the design space is small. "Try the ranked list in order; check the sites periodically; promote what works; demote what doesn't" is a finite set of predictable solutions with a reasonable expectation of success — the KSR "obvious to try" paradigm.

5.5 The specification itself is evidence of obviousness

The '941 specification:

• names off-the-shelf Nuance ASR and Lernout & Hauspie TTS engines — standard components;
• describes the web browsing server as "based upon Intel's Dual Pentium III 730 MHz microprocessor system" — a commodity server;
• describes its ranking/polling as solving problems it admits were already known (changing web sites, slow sites).

A patent that assembles purchased components and standard web-engineering techniques (failover lists, health-check polling, performance ranking) to solve acknowledged problems is the archetypal KSR obviousness case. The Board reached exactly this conclusion in IPR2022-00358.

6. Dependent claims — element-by-element obviousness

No dependent claim introduces a limitation that is not either (a) directly disclosed in a Tier 1–3 reference or (b) an obvious implementation detail of the Combination A–D architecture.

7. Secondary considerations — no weight here

The IPR record shows Parus pressed antedating/priority arguments and, per the family history, "mountain of evidence" secondary-consideration arguments — none of which survived. On this record:

• No unexpected results: faster, more reliable voice browsing is the predictable outcome of ranking, polling, and failover — not a surprising one.
• No long-felt need: Bayer '262 (1999) already delivered telephone voice browsing; the '941's contribution is incremental reliability engineering on a known concept.
• No nexus for commercial success: any later commercial success of voice assistants is decades removed from the 2000–2001 priority date and not attributable to the specific claimed combination; the claims were never shown to be the reason for any market success.
• Industry praise / copying: none shown; the only "adoption" evidence in the record is litigation.

These considerations are insufficient to overcome a strong prima facie § 103 case, and the PTAB and CAFC both so held.

8. Conclusion

Every claim of US 7,881,941 (claims 1–15) is unpatentable under 35 U.S.C. § 103 as obvious over the reconstructed prior art:

1. Bayer '262 + Amro '468 + Sabourin '964 + Brown '362 renders claims 1, 2, 9, and 10–13 obvious (the voice-browser loop, content extraction, grammar mapping, plurality-of-sites sequential fallback).
2. Adding Halverson '061 (and/or Jackson '767) renders claims 3–5, 9, and 14 obvious (ranking, rank-order access, ranking adjustment).
3. Adding Halverson '061 + the Burrows-type monitoring teaching renders claims 6–7 and 15 obvious (unprompted periodic polling for availability, response time, and content changes).
4. Halverson '061 (agent-based source discovery) plus conventional crawler art renders claim 8 obvious.

The motivations are the KSR staples: same field of endeavor, complementary references, a known problem (unreliable, slow, changing web sites in a latency-intolerant telephony environment) with a finite set of known solutions (fallback lists, performance ranking, health-check polling), and predictable results from combining known elements each performing its known function.

This analysis is fully consistent with — and independently corroborated by — the controlling legal reality: the PTAB's Final Written Decision in IPR2022-00358 (Aug. 1, 2023) held all challenged claims unpatentable under § 103, and the CAFC affirmed in Case 23-2296 (Rule 36, Mar. 6, 2025). The Board's grounds (Kovatch + Neal; Kovatch + Neal + Burrows; Wise + Shaffer + Burrows) are additional references not itemized in the reconstructed Prior Art list, but they reach the same conclusion through the same combination logic: a known voice-browser system plus known ranking/polling enhancements.

9. Cautions

1. Reconstructed list, not the patent face: the Prior Art section could not extract the "References Cited" page of 7,881,941 in its search pass. Before filing anything, verify the combinations against the IPR2022-00358 Final Written Decision (the definitive most-relevant-art list) and the USPTO patent PDF. Kovatch, Neal, Burrows, Wise, and Shaffer — the IPR's actual combination references — were not itemized in the reconstructed list; I have treated them only as corroboration.
2. § 102(e) date sensitivity: Sabourin '964, Brown '362, Shieber '098, Papineni '981, Jackson '767, and Halverson '061 issue after the Feb. 4, 2000 priority date; each is § 102(e) art only if its U.S. filing date predates the invention date. The § 103 analysis above does not depend on any single reference being § 102 art, because the combinations use at least Bayer '262 and Uppaluru '001 (both issued before the Feb. 5, 2000 critical date) as anchor references — but verify filing dates before citing the post-2000 references individually.
3. IPR claim-count discrepancy: the Patexia "claims 1–16" data point for IPR2022-00358 conflicts with the granted 15-claim set; it does not affect this obviousness analysis, which addresses claims 1–15 as granted.
4. Effective date of any invalidity: the patent expired "Fee Related" on ~July 2, 2021, and the FWD has been affirmed — the practical consequence is that the '941 is dead as an assertion vehicle, and the analysis above explains why it was invalid, not merely that it was.