Our Offerings · Invalidity Analysis

A granted claim is a set of limitations. Test every one.

Give the platform a granted patent and the agents hunt the art that reads on its claims: element by element, before the priority date, charted with evidence and combination arguments, in a claim-chart report you approved.

Claim-chart report X · Y · A grading §112(f) aware Pre-priority filtering

Why it matters

Why invalidity work is won and lost at the level of the element

An invalidity position is not an opinion about a patent. It is a demonstration, limitation by limitation, that every element of a claim was already disclosed: by one reference if you can get it, by a defensible combination if you cannot. The standard is unforgiving: miss one limitation and the argument fails, no matter how close the art feels. Which is why real invalidity work has always meant claim charts, and why claim charts have always been brutally expensive to build.

The mechanics are the problem. The claim must be decomposed into its limitations precisely, including the means-plus-function elements where the claim’s real scope lives in the specification’s corresponding structure, not in the claim language itself. Every candidate reference must predate the priority date, which means the date of each reference must actually be verified, not assumed from a database field. And the search has to keep going after the obvious hits, because the reference that completes the combination is usually the one indexed under vocabulary nobody was searching.

Esgenix runs this as a directed workflow. Capture the target by publication number and the claims and priority date are fetched; the claims are parsed into a claim ribbon: each limitation a discrete unit with its own date logic. Means-plus-function elements are flagged under §112(f), and the workflow will not proceed past them until the corresponding structure is confirmed, because a chart built on the wrong claim scope is worthless no matter how good its art is.

Any limitation still marked NOT DISCLOSED is the exact gap between your position and a complete one.

The search runs semantically across 220 million patents and the research literature, in rounds you direct, and everything it returns is filtered to art that predates the priority date. References whose dates cannot be verified are not quietly included or quietly dropped: they are parked in a date-verification tray until a person confirms them. Hits come back graded X (anticipates on its own), Y (works in combination, with the missing limitations named, not implied), or A (background), and you triage every round.

The chart builds as the rounds converge: limitations against references, with the disclosing passage quoted in every cell. A subtractive mode searches for art covering the claim without a chosen limitation: the fastest way to find which element is genuinely doing the work. An element-to-reference flow diagram keeps one thing impossible to overlook: any limitation still marked NOT DISCLOSED, the exact gap between your position and a complete one.

It closes as a claim-chart report with combination arguments assembled, reviewed, edited, and approved by you. Whether it is headed to a negotiation, a proceeding, or a decision about whether to fight at all, the position arrives the way it will eventually be judged: element by element, with the evidence attached.

The workflow

How a position gets built

Parsed precisely, searched before the priority date, graded in examiner vocabulary, and charted with evidence.

  1. 01

    Capture the patent

    A publication number or pasted claims. The claims and priority date are fetched and confirmed.

  2. 02

    Confirm the claim ribbon

    The claims parsed into discrete limitations with per-limitation date logic, plus a hard stop on §112(f) elements until the corresponding structure is confirmed.

  3. 03

    Pre-priority rounds

    Semantic search in rounds you direct, filtered to art before the priority date. Undatable references park in a verification tray until a person confirms them.

  4. 04

    Graded triage

    X anticipates alone; Y works in combination, with its missing limitations named; A is background. You triage every round on the board.

  5. 05

    Claim chart

    Limitations against references with the disclosing passages quoted, plus subtractive search to find which element is really doing the work.

  6. 06

    Claim-chart report

    The charted position with combination arguments assembled, reviewed and approved by you before it exports.

Capabilities

Built for the standard the chart will be judged by

The claim ribbon

Claims parsed into discrete limitations, each with its own date chips: the decomposition the whole analysis stands on, confirmed by you before anything searches.

§112(f) gate

Means-plus-function elements are flagged and the workflow stops until the specification’s corresponding structure is confirmed. Claim scope is settled before art is graded against it.

Pre-priority filtering

Only art before the target’s priority date survives, and the cutoff is applied per limitation where the date logic differs.

Date-verification tray

Same-year and undatable references are neither trusted nor discarded: they wait in a tray until a person verifies them. The chart never rests on an assumed date.

X / Y / A with named gaps

Y references carry their missing limitations explicitly: the exact elements a combination partner must supply, not a vague "relevant in combination".

The claim chart

Limitations against key references, every cell carrying the disclosing quote. Exportable, reviewable, and built to the format the argument will be judged in.

Subtractive search

Search for the claim without a chosen limitation and watch the art flood in. There is no faster way to prove which element is genuinely novel.

Element-to-reference flow

A flow diagram from claim elements to disclosing references, with a NOT DISCLOSED sink that makes any gap in the position impossible to miss.

Position analytics

Claim-element coverage per reference, grades by round, the priority timeline against the target date, and chart review progress, so you can see when the position is complete.

Deliverables

What lands in your inbox

Documents you can hand over. Every one is reviewed and approved by a person before it leaves the platform.

DOCX · PDF

Claim-chart report

The charted position with combination arguments, reviewed, edited, and approved by you before export.

XLSX

Claim chart workbook

Limitations against references with evidence quotes in every cell: the working record of the analysis.

PPTX

Position deck

The invalidity position as slides, for counsel review or the negotiation table.

Questions

What practitioners ask

The claim ribbon flags §112(f) elements and the workflow stops until the corresponding structure from the specification is confirmed. The real scope of a means-plus-function limitation lives in that structure. Grading art against the claim language alone would chart the wrong claim.

Everything is filtered to art before the target’s priority date. References whose dates cannot be verified automatically, such as same-year publications and undated disclosures, are parked in a verification tray until a person confirms them. No grade rests on an assumed date.

That is what the Y grade and combination groups are for: Y references carry their missing limitations by name, and suggested combinations show which references together cover the claim. The obviousness argument arrives assembled with its evidence, for you to accept, edit, or reject.

You do, at every stage: the claim ribbon, every triage round, the chart, and the report. The agents assemble; the judgment on what the position claims, and whether it is strong enough to use, is yours.

Chart a patent you actually face

Bring a publication number to the demo. Watch the claims parse into a ribbon, the pre-priority rounds land, and the chart build, with the gaps named.

© 2026 Syncinvent Labs Private Limited. All rights reserved.