The difference
Prosecution optimizes for allowance. Litigation optimizes for survival.
Those are not the same goal, and a patent drafted for the first can fail badly at the second. This page is the concrete version of our pitch — what actually changes in the document when the person writing it has spent years attacking patents like it.
The incentive problem nobody mentions
A prosecution-only practice is measured on one number: how many applications get allowed. That's the feedback loop. A drafter files, argues with an examiner, gets a notice of allowance, and the file closes. Everyone is happy. The client frames the certificate.
What that loop never surfaces is what happens five years later, when a competitor's counsel reads the same patent looking for a way through it. By then the matter is closed, the client has moved on, and the drafter never learns which of their habits produced a patent that held and which produced one that folded.
Litigation closes that loop. When you've sat on the other side — building invalidity contentions, taking apart a specification in claim construction briefing, watching a client's best claim get read out of existence over a single ambiguous word — you stop writing certain things. Not because a treatise told you to, but because you watched them fail.
To be clear about what this isn't
No drafter can make a patent invalidation-proof. Prior art you never saw can surface; the law shifts; examiners and judges vary. What litigation experience buys you is better odds and fewer self-inflicted wounds — the failures that trace back to a drafting choice rather than to the invention itself.
In practice
What changes in the document
1. Claim architecture built for a validity fight
Most claim sets are a broad independent claim plus dependent claims that add whatever detail was convenient. If the independent claim survives, the dependents were decoration. If it falls, the dependents often fall with it, because nobody positioned them to be useful on their own.
We build them as a ladder instead:
- A broad independent claim for negotiating leverage and to cover design-arounds — knowing it carries the most validity risk.
- Intermediate claims at deliberately chosen widths, each one still commercially meaningful on its own. If the broad claim dies, you're not left holding a claim so narrow that only your exact prototype infringes.
- A defensible core claim reciting the feature we believe is genuinely novel over the closest art we found — the one we expect to survive almost anything.
- Multiple statutory categories where the invention supports it — apparatus, method, and system claims fail for different reasons, so covering more than one category is cheap insurance.
Before we file, we run the set against the standard invalidity theories: anticipation under §102, obviousness under §103 including the combinations an examiner or an IPR petitioner would reach for first, eligibility under §101 where the subject matter invites it, and written description and enablement under §112. Where a claim looks weak, we'd rather learn it now than in a petition.
2. A specification written for the Markman hearing
In litigation, claim construction usually decides the case. The judge takes your claim terms and construes them, and the single most important piece of evidence is your own specification. Whatever you wrote years ago is now being used to define what you own — by someone actively looking for a narrow reading.
Writing with that in mind changes specific, checkable things:
- Deliberate lexicography. Where a term carries real weight, we define it explicitly and consistently rather than leaving it to a dictionary fight later.
- No accidental disavowal. Phrases like "the present invention requires…" or "the invention must…" read as a promise to the court that your claims are limited to exactly that. We write in embodiments, not in absolutes.
- Genuine alternative embodiments. Describing several ways to implement each element supports a broad construction and provides fallback support. A spec describing one embodiment invites a court to limit the claims to it.
- Clean antecedent basis and consistent terminology. Calling the same part three different names across a specification is a gift to opposing counsel.
- Structure behind functional language. After Williamson, terms like "module" or "unit" can be treated as means-plus-function and limited to disclosed structure — or held indefinite if none was disclosed. If we use functional language, the structure is in the spec.
3. Prosecution with an estoppel budget
Here's a cost most clients never hear about. When you narrow a claim to get around a rejection, you may permanently surrender the territory you gave up — you can't later recapture it under the doctrine of equivalents. That's Festo. The amendment that made an examiner happy in year two can be the reason you lose in year seven.
So before we amend, we price it: what scope does this give up, does it cover a design-around a competitor would plausibly attempt, and is there an argument-based route instead? Sometimes the right move is to argue rather than amend, or to take the rejection and pursue the broader claim in a continuation. Those choices need someone who has seen the bill come due.
4. Keeping the family alive
A single patent is a snapshot of what you thought mattered on your filing date. Markets move. We routinely keep a continuation pending so that when a competitor ships something adjacent, there's still a live application to pursue claims aimed at it — with your original priority date. Letting a family go abandoned at the first allowance is one of the most common and least recoverable mistakes we see.
Side by side
Same invention, two approaches
Both produce an issued patent. They behave very differently the first time somebody pushes on them.
| Typical prosecution-only drafting | Litigation-tested drafting | |
|---|---|---|
| Success measured by | Getting the application allowed | Whether the claims still stand when challenged |
| Claim set | One broad claim plus convenient dependents | A ladder of independently meaningful fallback positions |
| Prior art review | Is it novel enough to allow? | How would an IPR petitioner combine these references? |
| Specification | Describes the embodiment that was built | Written for claim construction, with alternatives and defined terms |
| Amendments | Whatever clears the rejection fastest | Priced against surrendered equivalents before filing |
| Freedom to operate | A separate engagement, if raised at all | Flagged during drafting, while design changes are cheap |
| After allowance | File closes | Continuation kept pending to reach future competitor products |
Straight talk
When you don't need this
We'd rather turn away work than sell you something that doesn't earn its cost. The Litigation-Ready tier is genuinely not worth it if:
- You're in an empty field with no visible competitors and no plans to raise capital.
- The patent is purely defensive — you want to publish and avoid being blocked, not enforce.
- Your product's commercial life is shorter than the time it takes a patent to issue.
- The invention is a minor improvement whose value doesn't justify the analysis.
- Budget is genuinely tight and the alternative is filing nothing at all — a solid Standard filing beats no filing.
In any of those cases we'll quote the Standard tier and say so plainly. The differentiator is only worth paying for when your patent is actually likely to be tested. Both tiers are laid out here.
Find out which you need
Tell us what you've built
Describe the invention and your situation — competitors, funding plans, timeline — and we'll tell you which tier actually fits, with a flat-fee quote. If Standard is the right answer, that's what we'll recommend.