Experimental results
Especially results that were unexpected, more efficient, more stable, or achieved with fewer steps.
For researchers and research groups
Your research may contain more than a publication.
Findings, methods, systems, compositions, datasets, processes, devices, and results that were not supposed to happen can all carry invention potential. Not all of them do — and the difference is worth establishing before your work becomes public.
Non-confidential summary first · Your dates drive the sequence
Fig. 2 — What we look for
- A repeatable technical effect, not just an observation
- A mechanism that explains why it works
- A difference from what the literature already does
- Variants that widen what could be claimed
- Evidence that supports the effect
Timing
Publication is a deadline, not just an achievement.
In most countries, a public disclosure before filing can be used against your own application. That includes a journal article, a preprint, a conference talk, a poster, a public thesis repository, a demo, a pitch without confidentiality, a product sale, and often a detailed social media post.
Some jurisdictions provide a limited grace period for certain disclosures by the inventor. Others do not. Relying on a grace period narrows where you can protect the work later.
The full pre-publication checklistBring these dates to the first conversation
- Journal or preprint submission date
- Conference abstract, poster, and presentation dates
- Thesis submission and repository release dates
- Funding, demo day, and industry showcase dates
- Any sale, quotation, or public trial already made
Important
If disclosure has already happened, options may still exist in some jurisdictions but they are narrower and time-bound. Tell us the exact date — it changes the advice.
Starting material
What is usually enough to begin.
You do not need a finished invention, a prototype, or patent vocabulary. You need something repeatable and a clear description of what it does better.
A method or protocol
A sequence of steps that produces a technical outcome others have not achieved the same way.
A system or device
An arrangement of components — including software components — that solves a defined problem.
A composition or material
A formulation, blend, coating, or structure with a measurable property advantage.
A model or pipeline
Where the technical contribution is in how it is built, trained, deployed, or constrained.
A failed experiment
Understanding why something fails is often the seed of the mechanism that fixes it.
Method
How a research review runs.
Two to three working sessions for most projects, sequenced around your publication dates rather than ours.
Start a review01
Non-confidential intake
You describe the field, the problem, and the shape of the result — without enabling detail. We confirm scope and set up confidentiality terms if you need them before going further.
02
Technical reading
We work through the results, methods, and anomalies rather than the paper's argument, and list every candidate technical effect.
03
Separation
Each candidate is sorted: knowledge, know-how, or a repeatable effect with a mechanism. Only the third category is a realistic invention candidate.
04
Differentiation
We search the closest literature and patent record for the leading candidate and state honestly how close the nearest references sit.
05
Invention architecture
The strongest candidate is written up as problem, mechanism, embodiment, variants, and evidence — the document a drafter can actually work from.
06
Direction
A recommended sequence with dates: file first, publish first, gather more data, or protect a different way. Sometimes the answer is do not file yet.
Redacted example
From a research statement to an invention architecture.
An illustrative transformation, not a client matter. It shows the shift in language that the work performs.
As written for a paper
"We observed that treating the substrate with a low-temperature plasma step before coating improved adhesion in 9 of 12 samples, with an unexplained drop in variance at longer exposure times."
As an invention architecture
- Problem: coating adhesion fails unpredictably on this substrate class
- Mechanism: pre-treatment alters surface energy in a specific window
- Effect: adhesion improves and variance falls — a measurable outcome
- Parameters: exposure time and temperature ranges worth claiming
- Variants: other substrates, other coatings, inline vs batch
- Open question: is the variance drop reproducible? That decides scope.
Note
The example is illustrative and does not describe any client's work. Whether an equivalent concept could be protected would depend on prior art, claim scope, jurisdiction, and disclosure history.
Paths
Different starting positions, different first step.
More detail on who we help.
Students
Your institution may own rights in the work. We help you find out before anything is disclosed or filed, and we work with your supervisor and IP cell rather than around them.
Faculty and research groups
Reviews are sequenced around submission and conference dates, and the disclosure is written so that a drafter needs fewer clarification rounds later.
Institutions and incubators
Repeatable triage: researcher education, disclosure intake, a consistent way to sort what is worth searching, and filing support for what survives triage.
Questions
What researchers ask first.
More in the full FAQ.
Not always, but it narrows them. Some countries allow a limited grace period after certain disclosures by the inventor; many do not. The safest sequence is to talk before the abstract goes out, not after the paper is online.
Usually the institution or employer, depending on your contract, funding terms, and local law. We will help you identify who needs to be involved before anything is filed, and we will not advise you to bypass your institution.
A method often can, depending on its technical character and the jurisdiction. A dataset by itself usually cannot, although the system that produces or uses it sometimes can. This is one of the questions a review answers for your specific case.
You get that answer in writing with the reasoning. That is a useful result: it saves the drafting and filing spend and often points to a different protection route, such as know-how, copyright, or publication strategy.
Send a non-confidential overview, not your data.
Enough to understand the field, the problem, and your dates. We will tell you what we need next and how to send it safely.