An office action is a letter from a USPTO examiner explaining why your application, as filed, is not allowable. The correct response is a written reply filed before the deadline that addresses every rejection and objection individually — usually a combination of narrowly targeted claim amendments and a technical argument explaining why the amended claims are still not obvious. Receiving one is not a setback; the overwhelming majority of applications get at least one, and many get two or three. What determines the outcome is the quality of the reply, not the existence of the rejection.
This article is general and educational. Projects House is an engineering firm, not a law firm — responses should be prepared or reviewed by a registered patent attorney or agent.
What the Document Actually Contains
An office action has two main parts. The first is the list of cited references: earlier patents, published applications, and other prior art the examiner found and considers relevant. The second is a rejection-by-rejection analysis, keyed to specific claims, citing the statutory basis and pointing to the exact passage in each reference the examiner is relying on.
Read the reasoning, not just the conclusion, because the reasoning shows how the examiner interpreted your claims. Quite often the whole disagreement stems from ambiguous claim language rather than from genuinely overlapping subject matter — and that is a much easier problem to fix. If working through the cited documents feels daunting, our guide to how to read a patent makes the structure far less intimidating, and what to do when you find a similar patent covers how to assess how close a reference really is.
The Common Types of Rejection
- Anticipation (lack of novelty). The examiner says one single reference discloses every element of the claim. Serious, but relatively easy to defeat if you can show one claimed element is genuinely absent from that reference.
- Obviousness. Two or more references combined, which the examiner says a person of ordinary skill would have combined. This is the most frequent rejection and also the most open to argument — often because there was no motivation to combine them, or because the combination would not actually work.
- Indefiniteness and lack of support. The claim is vague, uses a relative term, or recites something the specification never described. Usually resolved by rewording.
- Subject-matter eligibility. An assertion that the claimed subject matter is not patentable at all — an abstract idea, for instance. Common in software-adjacent filings; see can you patent software.
- Formal objections. Drawings, reference numerals, claim format, abstract length. Technical, but they still have to be cured.
Deadlines and Extensions
Every office action states a response period, typically a few months from the mailing date, and extensions are generally available for an additional fee. Missing the deadline can abandon the application, so the first action on receipt is to put the date in a calendar with reminders well ahead of it.
Extensions cost money and push out any eventual grant, so starting early is cheaper on both counts. For the wider timeline and budget picture, see how much a patent costs and what patent pending means for what your status is while all this is happening.
Amend the Claims or Argue?
This is the central decision in every response.
Argue without amending when you are convinced the examiner is wrong: an element of the claim was overlooked, unrelated references were combined without a technical rationale, or a reference was read far more broadly than it supports. The advantage is that you keep your original scope intact.
Amend to narrow when the reference really is close — typically by pulling a distinguishing feature from the specification up into the independent claim. The cost is a narrower monopoly that is easier for a competitor to design around.
In practice a good response does both: a measured amendment paired with a technical explanation of why even the amended combination would not have been obvious. A useful rule is not to surrender broad scope in the first round unless the cited art genuinely overlaps the invention, because scope given away early is very hard to recover later. This is also where the drafting quality of the original claim set pays off — well-layered dependent claims give you ready-made fallback positions, as explained in patent claims explained.
An examiner interview — a scheduled conversation about the rejection — is frequently the fastest route to agreement, because it lets you find out in twenty minutes what the examiner would accept instead of guessing across two more written rounds.
Where Engineering Input Matters
The attorney writes the legal argument, but the substance of the argument is usually engineering. The difference between "our device also has a spring" and "the cited mechanism cannot achieve the claimed preload without the second linkage, which the reference does not disclose and could not be added without defeating its stated purpose" is technical, not legal. Projects House often works on that side, helping the attorney identify the real engineering distinction between the invention and the art. The division of labor between the two roles is explained in patent attorney vs patent agent.
If the Second Answer Is Also No
A final rejection is not the end of the road. The usual options are a request for continued examination, a continuation or divisional application that lets you pursue different claim language, an appeal, or an after-final amendment. Alongside the procedural choice, weigh the business one: it is sometimes smarter to secure protection in one strong market than to fight in every jurisdiction, a tradeoff laid out in PCT application cost.
Preparation earlier in the process pays off here. A specification rich in technical detail and variations gives you raw material to amend from; a thin one leaves you with nothing to add. That is one of the strongest arguments for a thorough prior art search before drafting, and for filing in step with engineering progress. More on the whole subject is collected in our patents and intellectual property hub.
If you have an office action in hand and need engineering support to articulate the real technical difference between your product and the cited art, send us the details through our contact form.