Patent Eligibility Checklist: Claim Drafting, Alice Screening, and the Prosecution Record

By Casey Scott McKay · 2025-04-27

Eligibility is decided at the drafting desk years before anyone argues about it, which makes this a checklist you run before filing rather than after a rejection. It works in nine phases: screen the disclosure, find and document the technical improvement, write a specification that supports a practical-application argument, build a claim set at three scopes, file with the record in order, then diagnose, respond to, and escalate a rejection if one arrives, and finally manage the family. Each box gives the reason it exists, the authority behind it, and the trap that catches people who skip it. A worked matter runs throughout, following one software application from disclosure to issued claim. The companion article supplies the doctrine and the companion guide the response drafting.

IP and Technology > Patent Counseling Transactions | Checklist | Published 27 April 2025 - Updated 4 July 2026 | Casey Scott McKay - marksy.us

Summary. Eligibility is decided at the drafting desk years before anyone argues about it, which makes this a checklist you run before filing rather than after a rejection. It works in nine phases: screen the disclosure, find and document the technical improvement, write a specification that supports a practical-application argument, build a claim set at three scopes, file with the record in order, then diagnose, respond to, and escalate a rejection if one arrives, and finally manage the family. Each box gives the reason it exists, the authority behind it, and the trap that catches people who skip it. A worked matter runs throughout, following one software application from disclosure to issued claim. The companion article supplies the doctrine and the companion guide the response drafting.

Keywords: patent eligibility checklist, Section 101 screening, Alice framework, claim drafting, technical improvement, specification support, practical application, conventionality evidence, office action response, examiner interview, continuation strategy, prosecution record, mental process, abstract idea, invention disclosure, eligibility risk, patent drafting, claim scope, prima facie case, subject matter eligibility

How to use this checklist

| Phase | What it covers | Who owns it | When | |---|---|---|---| | 1 | Screen the disclosure for eligibility risk | Patent counsel | At intake | | 2 | Find and document the technical improvement | Counsel + inventors | Before drafting | | 3 | Write the specification to support the argument | Drafter | Drafting | | 4 | Build the claim set at three scopes | Drafter | Drafting | | 5 | File with the record in order | Counsel | Filing | | 6 | Diagnose the rejection | Counsel | On office action | | 7 | Respond | Counsel | Within the period | | 8 | Escalate | Counsel + client | After final | | 9 | Manage the family | Counsel | Continuous |

Phases 1 through 5 are the whole game. A disclosure screened properly, a technical improvement documented, a specification written to support it, and a claim set at three scopes will survive most eligibility rejections. Phases 6 through 8 are recovery from work not done earlier. Boxes marked [Gate] should clear before the next phase begins.

The matter. Ferrisbrook Analytics builds software that flags anomalous shipments for a logistics customer. The founders describe the invention as "using machine learning to detect supply chain fraud." Nothing in that sentence is patentable. What the engineers actually built — a streaming partition scheme with a per-segment probabilistic filter that cut index traversals by an order of magnitude — is. The boxes below show how that difference was found and preserved.

Phase 1. Screen the disclosure

[ ] Ask whether a person could perform the claimed steps mentally or with pen and paper, given time.
- Why. If yes, the claim is exposed as a mental process and the drafting must move toward the mechanism that makes the approach possible at machine scale.
- Authority. 35 U.S.C. § 101 and the judicially created exceptions applied through the two-step framework.
- Trap. Accepting the inventor's framing. Inventors describe outcomes; the eligibility question is about mechanisms.
[ ] Ask whether the invention is a thing or a way of deciding something.
- Why. Machines, devices, compositions, and manufacturing processes carry essentially no eligibility risk. Decision procedures carry most of it.
- Trap. Assuming a software product is uniformly risky. Most contain a genuinely technical core alongside the decision logic.
[ ] Ask what breaks if the computer is removed.
- Why. "It takes longer" means there is no technical improvement. "It cannot work because the method depends on how the machine stores or transmits data" means there is one.
- Trap. Accepting speed alone as the answer. Doing a known thing faster is the paradigm ineligible claim.
[ ] Classify the subject matter by industry risk.
- Why. Mechanical, chemical, and materials inventions can skip most of this checklist. Fintech, diagnostics, and enterprise software cannot.
- Trap. Running a full eligibility workup on a valve geometry, which wastes budget.
[ ] [Gate] Give the client a written risk assessment before the drafting budget is committed.
- Why. A client who learns of eligibility risk after eighteen months and a full prosecution spend has been failed by counsel, not by the law.
- Trap. Softening the assessment because the client is enthusiastic.

Ferrisbrook, Phase 1. The screen killed the framing. "Detecting supply chain fraud" is a method of organizing human activity performed on a computer. The engineers were asked what the machine does that it could not do before, and the answer — suppressing index traversals for records that cannot match — was the actual invention.

Phase 2. Find and document the technical improvement

[ ] Interview the engineers, not the founders.
- Why. The commercially interesting feature and the patentably interesting feature are usually different, and only the engineers know the second one.
- Trap. Working from the pitch deck.
[ ] Ask the specific question: what does the machine do differently?
- Why. Open questions produce marketing answers. This question produces engineering answers.
- Trap. Accepting "it uses AI," which describes a tool rather than an improvement.
[ ] Look for the named improvement categories.
- Why. Reduced memory footprint, fewer processing operations, lower latency, reduced bandwidth, fewer round trips, eliminated synchronization, improved error rate, a new security property, or enabling a previously infeasible computation. These are the categories that map onto a practical-application argument.
- Trap. A benefit stated only in business terms — cost, revenue, user satisfaction — which supplies no eligibility argument at all.
[ ] Quantify it if possible.
- Why. A measured improvement is far more persuasive than an asserted one, in prosecution and in litigation.
- Trap. Numbers that cannot be supported later, which become a validity problem.
[ ] Record what was conventional at the priority date, and what was not.
- Why. Conventionality is a factual question, and the record is far easier to build contemporaneously than three years later.
- Authority. Evidence traversing a rejection is submitted under 37 C.F.R. § 1.132.
- Trap. Waiting for the examiner to assert conventionality before thinking about proof.

Phase 3. Write the specification to support the argument

[ ] [Gate] State the technical problem in the background, in engineering terms.
- Why. A court or examiner characterizing the claim will use the specification's own vocabulary. A background describing a commercial frustration produces a commercial characterization.
- Trap. A background written by marketing.
[ ] Describe the deficiencies of prior approaches specifically.
- Why. The contrast is what establishes that the advance is technical.
- Trap. Disparaging prior art so broadly that it reads as an admission about scope.
[ ] Explain how the invention improves system operation.
- Why. This is the passage you will cite by paragraph number in every eligibility response.
- Trap. Describing only what the user sees.
[ ] Include enough implementation detail to support the narrow claims.
- Why. Amendments cannot add new matter, so anything you may need later must be in the application as filed.
- Authority. 35 U.S.C. § 132(a); support requirements at 35 U.S.C. § 112(a).
- Trap. A thin specification that forecloses the amendment that would have saved the claim.
[ ] Provide structure for any functional claim term.
- Why. A functional limitation may be construed under 35 U.S.C. § 112(f), and without corresponding structure the claim is indefinite.
- Trap. Algorithmic functions described only as results.
[ ] Number paragraphs and keep a map of which paragraph supports which argument.
- Why. Every response will cite them; assembling the map at drafting time costs minutes and saves hours later.

Phase 4. Build the claim set at three scopes

[ ] Draft a broad claim you expect to fight for.
- Why. It is the appeal claim and the continuation claim, and it cannot be created later if the family closes.
- Trap. Filing only broad claims and having nothing allowable.
[ ] Draft a middle claim carrying the technical hook.
- Why. This is usually the allowance path — specific enough to recite a mechanism, broad enough to cover competitors.
- Trap. A middle claim that adds hardware rather than mechanism, which adds nothing.
[ ] Draft a narrow claim tied to the disclosed embodiment.
- Why. It secures an issued patent while the broader fight continues.
- Trap. A narrow claim so specific that no competitor would ever infringe it.
[ ] [Gate] Chart each claim against the two or three products the client actually worries about.
- Why. An allowable claim that reads on nothing is a certificate, not an asset.
- Authority. See Patent Infringement Contention Checklist.
- Trap. Discovering after allowance that the amended claim missed the market.
[ ] Include claims in more than one statutory class where the disclosure supports it.
- Why. Method, system, and computer-readable-medium claims can fare differently, and a system claim reciting specific structure sometimes survives where a method claim does not.
- Trap. Boilerplate parallel claims that merely restate the method, which fail together.
[ ] Avoid the collect-analyze-display architecture in the independent claim.
- Why. Receive data, apply a rule, output a result is the single most frequently invalidated claim structure in the field.
- Trap. Writing it because it is the clearest description of what the product does.

Ferrisbrook, Phase 4. Three independent claims. The broad one recited segment-wise filtering generally. The middle one recited the partition rule and the per-segment filter. The narrow one recited the specific filter width and eviction policy from the embodiment. All three were charted against two competitor products; the narrow one missed both and was kept anyway as the allowance path.

Phase 5. File with the record in order

[ ] Confirm inventorship before filing.
- Why. Inventorship follows conception, and errors are correctable but awkward.
- Authority. 35 U.S.C. § 115; 35 U.S.C. § 116; correction under 35 U.S.C. § 256.
- Trap. Listing the founders rather than the conceivers.
[ ] Confirm assignments are executed and recorded.
- Authority. 35 U.S.C. § 261.
- Trap. An "agrees to assign" clause where a present assignment was needed.
[ ] File the information disclosure statement, and keep the duty running.
- Why. The duty of candor continues while the application is pending, and late-learned material art must be disclosed.
- Authority. 37 C.F.R. § 1.56; 37 C.F.R. § 1.97.
- Trap. Treating the IDS as a one-time filing task rather than a standing obligation. See Prior Art and Patentability Checklist.
[ ] Decide the provisional question deliberately.
- Authority. 35 U.S.C. § 111(b); benefit claimed under 35 U.S.C. § 119(e).
- Trap. A thin provisional that does not support the claims it is later used to antedate. See The Priority Chain.
[ ] Docket every deadline the day the receipt arrives.
- Trap. A docket living in one person's calendar.

Phase 6. Diagnose the rejection

[ ] Identify the alleged exception and the grouping.
- Why. Abstract idea, law of nature, or natural phenomenon — and within abstract idea, mathematical concept, method of organizing human activity, or mental process. The response differs for each.
- Trap. Responding generically to a rejection you have not classified.
[ ] Identify the specific claim language mapped to the exception.
- Why. A rejection paraphrasing the claim as a whole into one sentence has arguably not established a prima facie case.
- Authority. 37 C.F.R. § 1.104.
- Trap. Accepting the examiner's characterization by arguing only step two.
[ ] Identify what the examiner said about the additional elements.
- Trap. Missing that the examiner conceded additional elements exist, which frames the whole response.
[ ] Check whether any conventionality finding is supported.
- Why. An unsupported assertion can be traversed for free, before spending on a declaration.
- Trap. Commissioning the declaration first.
[ ] Note whether art was also cited.
- Why. A claim rejected under 35 U.S.C. § 101 with no rejection under 35 U.S.C. § 102 or 35 U.S.C. § 103 is in a different posture — a fact to use carefully and never as a novelty argument.
- Trap. Arguing novelty at an eligibility rejection, which is the most common failure in the field.

Phase 7. Respond

[ ] Choose the arguments and label them in order.
- Why. No exception recited, in the alternative integration into a practical application, in the further alternative significantly more. Clear labeling lets the examiner adopt the narrowest ground for allowance.
- Trap. Blending the arguments into a paragraph.
[ ] Amend to add mechanism, not hardware.
- Trap. Adding a processor and a memory, which examiners disregard.
[ ] Cite specification paragraphs by number for the technical improvement.
- Trap. A practical-application argument with no specification citation, which is an assertion.
[ ] [Gate] Request an interview, with proposed claim language in hand.
- Why. Free, fast, and the highest-yield step available. The diagnostic question is what language would place the claim in a practical application in the examiner's view.
- Authority. 37 C.F.R. § 1.133.
- Trap. Interviewing after two written rounds instead of before the first.
[ ] Respond substantively to every ground.
- Authority. 37 C.F.R. § 1.111.
- Trap. A response that leaves a dependent-claim rejection unaddressed.
[ ] Keep claims at different scopes across the set.
- Trap. Amending every claim to the narrowest formulation, which forecloses the appeal and the continuation.

Phase 8. Escalate

[ ] Consider a pre-appeal brief request before a full appeal.
- Why. A panel review at a fraction of the cost, effective against clearly deficient rejections.
- Trap. Skipping straight to appeal and spending the budget.
[ ] Evaluate the request for continued examination against appeal.
- Authority. 37 C.F.R. § 1.114.
- Trap. Repeated requests that buy rounds with an examiner who is not moving.
[ ] If appealing, organize the brief by claim grouping.
- Authority. 35 U.S.C. § 134; further review at 35 U.S.C. § 141 or 35 U.S.C. § 145.
- Trap. Arguing all claims together, which lets a single weak claim sink the group.
[ ] [Gate] File a continuation before the appeal concludes.
- Authority. 35 U.S.C. § 120.
- Trap. Losing the family while winning or losing the appeal.
[ ] Reassess whether the fight is worth it against the technology's commercial life.
- Trap. A three-year appeal on a product with a two-year market.

Phase 9. Manage the family

[ ] Keep something pending, always.
- Why. Continuations must be co-pending, and once the parent issues the opportunity is gone permanently.
- Authority. 35 U.S.C. § 120.
- Trap. Paying the issue fee without filing the continuation first. This is the most consequential single deadline in the whole prosecution.
[ ] Vary claim scope and framing across the family.
- Why. The same disclosure supports claims at the data-structure layer, the system layer, and the method layer, and they present different eligibility postures.
- Trap. Filing continuations with the same claims that were just rejected.
[ ] Review the notice of allowance before paying.
- Why. Examiner's amendments change scope, and a reasons-for-allowance statement will be quoted in litigation.
- Trap. Paying the fee without reading either.
[ ] Verify the patent term adjustment calculation.
- Authority. 35 U.S.C. § 154(b).
- Trap. A miscalculation left uncorrected past the window.
[ ] Docket the maintenance fees.
- Authority. 35 U.S.C. § 41(b).
- Trap. A lapsed patent, which is the cheapest way to lose an asset that cost six figures.
[ ] Record what this examiner and art unit responded to.
- Why. It is the most useful intelligence you have for the next application in the family.

Ferrisbrook, outcome. The first office action rejected all claims under § 101 as a mental process, mapping only the receive-and-analyze limitations and omitting the partition and filter steps. The response identified the omission, argued no exception was recited, argued integration in the alternative citing paragraphs 0042 through 0047, and requested an interview. The examiner proposed adding the eviction policy to the independent claim. The middle claim issued nine months later; a continuation with the broad claim remains pending.

Phase 10. Technology-specific screens

The base checklist assumes software. Six categories need extra boxes.

Diagnostics and personalized medicine

[ ] Determine whether the claim recites a natural correlation.
- Why. A method of detecting a relationship that exists in nature, using conventional detection steps, is the paradigm law-of-nature failure.
- Trap. Assuming the discovery's difficulty confers eligibility. It does not.
[ ] Consider claiming the treatment step instead.
- Why. A method of treating a condition by administering a specific compound at a specific dose is eligible even though it exploits a natural relationship.
- Trap. A treatment claim so broad that it reads on administering anything to anyone.
[ ] Consider claiming the detection technology.
- Why. A novel assay, probe, reagent, or instrument is a composition or a machine, and eligibility follows.
- Trap. Claiming a conventional assay applied to a new correlation.
[ ] Preserve the trade secret alternative for the underlying model.
- Authority. 18 U.S.C. § 1839(3).
- Trap. Publishing the correlation in a specification that never yields an eligible claim, destroying both options.

Biotechnology

[ ] Confirm the claimed material has markedly different characteristics from what occurs in nature.
- Trap. Claiming an isolated natural sequence.
[ ] Claim the engineered construct, the vector, the cell line, or the method of making.
- Trap. Relying on isolation as the point of novelty.

Financial technology

[ ] [Gate] Identify the infrastructure claim, or advise that there may be none.
- Why. The financial product maps directly onto the fundamental-economic-practice category. Settlement mechanics, cryptographic protocols, and data structures do not.
- Trap. Filing on the product and hoping.
[ ] Check the regulated-naming and disclosure overlay separately.
- Authority. See Branding Money.
- Trap. Treating patent risk as the only legal risk in a fintech launch.

Machine learning and artificial intelligence

[ ] Distinguish applying a model from improving how models are built or run.
- Why. Applying a model to data is the collect-analyze-display architecture. A training technique, an architecture, a quantization scheme, or a hardware acceleration approach is a technical mechanism.
- Trap. Claiming "a trained model configured to output a prediction," which recites a tool and an outcome.
[ ] Address inventorship for machine-assisted conception.
- Authority. 35 U.S.C. § 115; 35 U.S.C. § 116.
- Trap. Assuming the question does not arise because a human filed the application.
[ ] Coordinate with the training-data and output-ownership analysis.
- Authority. See Who Owns What the Machine Made; Generative AI IP Compliance Checklist.
- Trap. A patent strategy that ignores the provenance of the data the invention depends on.

User interface and presentation

[ ] Determine whether the advance is in what is displayed or in how the system produces it.
- Why. Arranging information for a user is the archetypal abstract idea; a specific technique that makes a display possible or more efficient is not.
- Trap. Claiming a layout.
[ ] Consider design patent protection for the visual appearance instead.
- Authority. 35 U.S.C. § 171. See Three Ways to Own a Shape; Layered Design Protection Toolkit.
- Trap. Missing the design patent deadline while fighting a utility eligibility rejection.

Mechanical, electrical, chemical, and materials

[ ] Confirm the claim is to a machine, manufacture, or composition, and stop.
- Why. Eligibility is not a live issue in these fields, and running a full workup wastes budget.
- Trap. Method-of-analysis claims appended to a device application, which import the risk the device claims avoid.

Phase 11. The intake questionnaire

Most of Phase 1 can be delegated to a form the inventors complete before the first meeting. This is the working version.

[ ] What problem does the system solve, stated without reference to any business benefit?
- Why. The answer separates a technical problem from a commercial one in a single question.
- Trap. Accepting an answer containing the words cost, revenue, efficiency, or customer.
[ ] What did engineers try first, and why did it not work well enough?
- Why. This is the background section of the specification, and it is the contrast that establishes a technical advance.
- Trap. No answer, which usually means nobody tried anything else and the advance may be the application rather than the mechanism.
[ ] What measurable property improved?
- Why. Memory, latency, operations, bandwidth, round trips, error rate, throughput. A number here is worth more than a paragraph of description.
- Trap. "It's more accurate," which may be a model-quality claim rather than a system improvement.
[ ] Could this run on ordinary hardware without the claimed technique, just more slowly?
- Why. If yes, the practical-application argument is weak and the drafting must find a different hook.
- Trap. Confusing infeasible-in-practice with impossible-in-principle; the former is still a real argument if quantified.
[ ] Which parts of the system were standard at the time, and which were not?
- **Why. **This is the conventionality record, captured while people remember.
- Authority. 37 C.F.R. § 1.132.
- Trap. Asking three years later, when nobody can date anything.
[ ] What have you published, demonstrated, offered for sale, or shown to a customer, and when?
- Why. It is a prior-art and grace-period question that must be answered before filing, and it frequently changes the timeline.
- Authority. 35 U.S.C. § 102(a); 35 U.S.C. § 102(b)(1).
- Trap. A conference talk nobody mentioned. See Prior Art in a First-Inventor-to-File World.
[ ] Who conceived each element, and is each of them under an assignment obligation?
- Authority. 35 U.S.C. § 116; 35 U.S.C. § 261.
- Trap. A contractor who contributed a claimed element and signed nothing. See Employee, Founder, and Mobility IP Toolkit.
[ ] What would a competitor have to build to compete, and would the draft claims read on it?
- Why. It is the commercial test, and it belongs at intake rather than at allowance.
- Trap. A portfolio of claims covering the client's implementation and nothing else.

Use the answers directly. The first four become the background and summary sections. The fifth becomes the conventionality record. The sixth sets the filing deadline. The seventh clears title. The eighth sets claim scope. A completed questionnaire turns a two-hour intake meeting into a forty-minute one and produces a materially better application.

Phase 12. The annual portfolio eligibility review

Issued patents can be attacked on eligibility for their entire life, so a portfolio needs a periodic look rather than a one-time clearance.

[ ] Flag every issued patent whose independent claim follows the collect-analyze-display architecture.
- Why. These are the assets most likely to fail if asserted, and knowing which they are changes what you assert.
- Trap. Discovering it in a motion to dismiss.
[ ] Identify which patents rest on a specification that describes a technical problem, and which do not.
- Why. The specification is the resource for a step-one argument, and a patent without one has a much weaker defense.
- Trap. Assuming a strong claim compensates for a commercial specification.
[ ] Check whether a continuation remains pending in each significant family.
- Authority. 35 U.S.C. § 120.
- Trap. Families closed years ago that can never be reopened.
[ ] Consider reissue for a patent with a correctable defect.
- Why. Reissue can broaden within two years of grant or narrow at any time, and a narrowing reissue can cure an eligibility exposure before it is tested.
- Authority. 35 U.S.C. § 251.
- Trap. The recapture rule, which bars regaining subject matter surrendered in prosecution.
[ ] Track the maintenance fee decision against eligibility risk.
- Why. A patent unlikely to survive an eligibility challenge and unlikely to be licensed is a recurring expense.
- Authority. 35 U.S.C. § 41(b).
- Trap. Paying maintenance on everything because nobody wants to make the call. See Patent Portfolio Management Toolkit.
[ ] Re-check the law before relying on any prior opinion.
- Why. The case law accumulates on both sides continuously, and an eligibility memo from three years ago describes a different landscape.
- Trap. Reusing a dated opinion for a new assertion decision.

A closing note on who runs this checklist. The screening phases belong to whoever takes the invention disclosure, which in most organizations is not a patent lawyer — it is an engineering manager or an IP coordinator working from a form. Phases 1, 2, and 11 are written to be usable by that person, and pushing them upstream is what makes the rest affordable. The drafting and prosecution phases belong to counsel. The portfolio review at Phase 12 belongs to whoever owns the maintenance budget, because that is the person who can act on what it finds.

And the intake questionnaire at Phase 11 should be a standing form in the disclosure system rather than something counsel sends after a disclosure arrives, because the answers are far better when the engineers write them while the work is fresh than when they reconstruct them for a lawyer months later.

A form completed at the bench is worth three interviews conducted from a conference room.

That single change in where the questions get asked improves more applications than any drafting technique in this document.

Organizations that make the change usually discover a second benefit within a quarter: engineers who have answered the questions once start recognizing patentable subject matter on their own, and the quality of what reaches counsel rises without anyone being trained.

Key Authorities at a Glance

| Authority | Proposition | Phase | |---|---|---| | 35 U.S.C. § 101 | Eligible subject matter; the gateway | 1, 6 | | 35 U.S.C. § 102 | Novelty; the argument to avoid at § 101 | 6 | | 35 U.S.C. § 103 | Nonobviousness; also distinct | 6 | | 35 U.S.C. § 111(b) | Provisional applications | 5 | | 35 U.S.C. § 112(a) | Written description and enablement | 3 | | 35 U.S.C. § 112(f) | Means-plus-function construction | 3 | | 35 U.S.C. § 115 | Inventor's oath or declaration | 5 | | 35 U.S.C. § 116 | Joint inventors | 5 | | 35 U.S.C. § 119(e) | Benefit of a provisional | 5 | | 35 U.S.C. § 120 | Continuations; co-pendency | 8, 9 | | 35 U.S.C. § 132(a) | No new matter | 3, 7 | | 35 U.S.C. § 134 | Appeal to the Board | 8 | | 35 U.S.C. § 154(b) | Patent term adjustment | 9 | | 35 U.S.C. § 261 | Assignment in writing | 5 | | 35 U.S.C. § 41(b) | Maintenance fees | 9 | | 37 C.F.R. § 1.56 | Duty of disclosure | 5 | | 37 C.F.R. § 1.104 | Nature of examination | 6 | | 37 C.F.R. § 1.132 | Declarations traversing rejections | 2, 7 | | 37 C.F.R. § 1.133 | Interviews | 7 |

The five things people get wrong

One: they screen after drafting instead of before. Eligibility is a drafting problem, and by the time a rejection arrives most of the useful choices have already been made. The twenty-minute screen at intake is worth more than any response.

Two: they take the inventor's framing. "We use AI to detect fraud" is a description of a product, not of an invention. The engineers know what the machine does differently, and nobody asks them.

Three: they argue novelty. The absence of prior art is not an eligibility argument, examiners are trained to disregard it, and including it signals that the real argument is thin.

Four: they add hardware to the claim. A processor, a memory, a display. These are the amendments that cost scope and buy nothing, and they appear in an enormous share of responses.

Five: they let the family close. A narrow claim allowed, the issue fee paid, no continuation pending — and the broad claims can never be pursued again. It is the single most expensive mistake in this checklist and it is entirely a calendaring failure. See Overcoming a Section 101 Rejection.