The June 18, 2026 pub drop is the weekly batch of newly published U.S. patent applications, and it is worth being precise about what "published" means: these documents are filings that have cleared the publication window, not issued grants. The claims in them are what an applicant has asked the office to examine. So when a record lands titled Method and apparatus for decentralized privacy preserving audit based on zero knowledge proof protocol, the right question is not whether the idea is novel but what the pending claims actually reach. This one, US20260172254A1, is assigned to PwC Product Sales LLC, the entity through which the accounting firm files much of its software estate, and it is directed at a problem auditors have always had: how to check two parties' records against each other without either party handing over the underlying data.

Read at the level of the independent method, the application is directed to a sequence rather than a single act. A first local system receives, from an orchestrator system, instructions to perform a local audit. It then executes what the application calls a local audit model epoch: it analyzes information from one or more data sources in its own local data set, requests and receives input information from a second local system through a privacy-preserving communication framework, and determines whether its local data is reconciled based on that input. If it is not reconciled, the system executes a second epoch. The iteration is the mechanism. The audit is not a one-shot comparison but a loop that keeps exchanging privacy-preserved inputs and re-running until the local data either reconciles or the process is exhausted.

Provided herein are systems and methods for decentralized privacy preserving audits comprising receiving by a first local system, from an orchestrator system, instructions to perform a local audit; executing, by the first local system, in response to receiving the instructions to perform the local audit, a first local audit model epoch, wherein executing the first local audit model epoch comprises: analyzing information from one or more data sources from a local data set; receiving first input information from a second local system through a privacy preserving communication framework in response to requesting the first input information; determining whether information from the one or more data sources from the local data set is reconciled based on the first input information; and in accordance with a determination that the information from the one or more data sources from the local data set is not reconciled, executing a second local audit model epoch.— Method and apparatus for decentralized privacy preserving audit based on zero knowledge proof protocol, US20260172254A1

Two features of that recited flow are worth isolating because they shape the scope. First, the architecture is described as decentralized: an orchestrator coordinates, but the audit work happens at the local systems, each operating on its own data set, rather than at a central node that ingests everyone's records. Second, the cross-system exchange runs through a "privacy preserving communication framework" — the abstract and title tie that framework to a zero-knowledge proof protocol, the cryptographic primitive that lets one party convince another a statement is true without revealing the data behind it. In an audit context, that is the disclosed answer to the data-sharing problem: a counterparty can demonstrate that its figures reconcile without exposing the figures themselves.

Where the application sits in the classification landscape

The record carries three CPC classifications, and they read as a fairly tight description of what the filing is about. H04L 9/3221 is the subgroup for zero-knowledge proofs and related interactive proof protocols within the cryptographic-mechanisms class — the same H04L 9 family that anchors much of the blockchain and distributed-ledger patent estate, including the H04L 9/50 ledger-specific subgroup. H04L 9/083 covers key distribution involving a central authority or key-distribution center, which fits the orchestrator-plus-local-systems shape of the disclosure. H04L 63/0428 sits in the network-security class and covers transmitting data with the payload itself encrypted. Notably, the classification places this filing in the cryptography-and-network-security neighborhood rather than the G06Q 20 payments cluster where a large share of the week's other blockchain-adjacent applications landed; the June 18 drop returned a substantial pool on the keyword, and its CPC facet was led by H04L 9/50 and several G06Q 20 payment subgroups. This application is on the cryptographic-protocol side of that map, not the payments side.

It is also worth noting what the title's "zero knowledge proof protocol" language does and does not settle. The independent flow quoted above recites a "privacy preserving communication framework"; the zero-knowledge characterization comes through the title, abstract, and the H04L 9/3221 classification rather than being spelled out as a limitation in the lead recitation. For a reader tracking scope, that is the kind of gap between the marketing-level framing of a filing and the literal recitations that is worth watching as the application moves through examination — the framework is the recited element, and the proof protocol is the disclosed way of building it.

The application against PwC's recent filings

This publication does not arrive in isolation. PwC Product Sales LLC has been filing steadily in audit and accounting automation, and the same named inventors recur. The reconciliation problem at the center of this application also drives US20250209092A1, directed to composable microservices for record linkage and reconciliation, and the broader workflow appears in US20250182217A1, an AI-augmented auditing platform with automated document processing. Several of the inventors on the zero-knowledge audit filing — including the team of Li, Cheng, Flavell, Hamer, and Davies — are listed on those records too, which makes the cluster read as connected work rather than scattered filings.

Around that core sit applications directed at adjacent pieces of an audit-and-advisory toolchain: US20250165537A1 on knowledge representation and reasoning in accounting, US20250231971A1 on an AI-assisted virtual consultant, US20250322170A1 on data merging using machine learning, and US20260148007A1 on automated revision of document language by readability and sentiment. Most of that estate is classified in the G06F, G06N, and G06Q families — software, machine learning, and business-method classes. The zero-knowledge audit filing is the one that reaches into the H04L 9 cryptography classes, which is what makes it the blockchain-adjacent entry in an otherwise AI-and-accounting portfolio.

For a sector that talks constantly about "privacy-preserving" and "zero-knowledge" as adjectives, the value of reading a record like this one is in pinning the terms to recited steps: an orchestrator issuing instructions, local systems analyzing their own data, an encrypted-payload exchange framework between them, and an epoch loop that repeats until reconciliation. Whether the claims survive examination in their current form, and how the zero-knowledge characterization is ultimately reflected in the allowed claim language, is a question for the file history. What the published application establishes today is the scope an applicant in the audit business has asked the office to consider — and the corner of the cryptography map, H04L 9/3221, where it has staked the filing.