Patent due diligence is often reduced to a simple question: what patents does the company have?
That’s rarely enough. For an early-stage technology company — particularly one built around a platform rather than a single product — the value of an IP portfolio depends on far more than the number of applications on a spreadsheet. The real exercise is understanding what the patents actually protect, how that protection has developed through prosecution, whether it maps onto the technology being commercialized, who owns it, and what sits entirely outside the portfolio. Having run this kind of review on early-stage portfolios, here’s the method broken into the steps that actually do the work — each one capable, on its own, of changing a deal’s outcome.
1. Define the scope before opening the portfolio
Start by establishing what the exercise is actually meant to answer, not by searching the patent database.
A preliminary review based on public records is fundamentally different from a full legal opinion. It may cover published applications, granted patents, prosecution records, and public company disclosures. On its own, it does not establish validity, enforceability, infringement, or freedom to operate — and a review of public ownership records does not establish a complete chain of title. State these limitations explicitly, near the top of the report and again near the conclusions, so the reader never mistakes a preliminary finding for legal certainty.
2. Reconstruct what the portfolio actually contains
Before assessing strength, the portfolio has to be reconstructed accurately: identify the relevant patent families, understand how they relate to one another, check applicants and inventors against every corporate name the company has used, and determine where protection has actually been pursued versus merely claimed.
This usually reveals whether a handful of published families are genuinely independent inventions or, more often, technically complementary pieces of a single underlying platform. That distinction is far more informative than a bare count of families, and everything else in the review builds on it.
3. Read the claims — then read the prosecution history
A specification tells you what the applicant says the invention is. The claims tell you what protection is actually being sought. The prosecution history tells you how much of that protection has survived scrutiny.
Search and examination records typically show prior art disclosing the individual elements of an invention without disclosing the specific combination actually claimed — a meaningfully different signal from prior art that anticipates the claim outright. Just as important is separating the character of any objections raised: clarity, support, or claim-breadth objections may be capable of being addressed through amendment, but can result in meaningful narrowing of the claims; novelty or inventive-step objections can present a more fundamental challenge to the underlying concept. A report that doesn’t draw this line reduces prosecution to a false, uniform “pending” or “granted” status.
4. Track how the claims changed
Compare claim language across successive prosecution stages, not just the current version. It’s common to see broad functional language narrowed into specific structural or compositional limitations tied to the working examples in the specification, often alongside a sharp drop in claim count as overlapping embodiments are consolidated. That produces a more focused, generally more defensible claim set.
But there’s a question a rushed review skips: what commercial territory was lost in the process? A narrower claim may be more focused while covering fewer real-world implementations. The right conclusion isn’t that prosecution simply “improved” the patent — it’s whether the resulting scope remains commercially meaningful for what the company actually intends to do.
5. Map the claims against the commercial story
This is where the analysis has to move beyond the patent database entirely.
A company will typically describe its business in terms of specific products or applications. Its patents may instead protect an underlying platform capable of supporting many different products — which is not a weakness, but it means product-specific protection can never be assumed from the existence of platform patents. Take every commercial application the company describes and check, claim by claim, whether any language in the portfolio actually reaches it, rather than inferring coverage from proximity.
6. Map the IP across the operational chain
Sketch the value chain the company actually operates:
Input → core process or platform step → intermediate output → further processing → finished product or service.
Then plot the claims against each stage.
It’s common to find dense coverage clustered around the stage the company considers its core innovation, with no independent published protection for adjacent stages that remain essential to delivering the commercial output. That gap is significant: a company can hold strong patents around one link in the chain and still need third-party rights or additional filings to cover the rest. Owning the patent for one stage and being free to operate across the whole chain are two different things.
7. Separate ownership from what appears on the register
A register showing a single, uncomplicated applicant name, with no apparent co-ownership or irregularity, is reassuring — but it is not a completed chain-of-title review.
That requires documents a database won’t show executed assignments, employment agreements with invention-assignment provisions, and any collaboration or joint-development agreements that could create obligations invisible from the public record. A preliminary review can flag the absence of a visible red flag; it cannot confirm ownership is legally complete. Those are different claims, and conflating them is a common and consequential drafting error.
8. Look beyond the patents entirely
Some of the most commercially important elements of a platform technology may never appear in a patent at all — process protocols, screening methods, operating parameters, analytical techniques, and technology-transfer materials retained as trade secrets rather than disclosed in a filing.
This creates a distinction worth stating outright: the patented technology versus the complete technical package actually required to reproduce and commercialize it. Where commercial value depends substantially on know-how, diligence needs a separate line of inquiry — what’s proprietary, how it’s protected, who has access, and whether it will actually transfer with the business.
9. Do not confuse ownership with freedom to operate
A patent portfolio tells you what a company may be able to exclude others from doing. It tells you nothing about whether the company is free to practice its own technology.
Patent due diligence asks what IP does this company have? Freedom to operate asks what third party IP could stop this company from doing what it wants to do? Same source material, entirely different analysis. A preliminary review that doesn’t flag FTO as a distinct, unaddressed question — specifying roughly which areas it should cover — leaves the reader with a false sense of completeness.
10. The real output is the list of unresolved issues
A good report doesn’t conclude that a portfolio is “strong” or “well positioned.” It’s useful function is to state what’s established, what’s uncertain, and what specifically needs investigating before a commercial decision is made — chain-of-title verification, unpublished or subsequent filings, pending prosecution, product-specific protection elsewhere, and a dedicated FTO assessment. Each item should tie back to a specific gap the preceding steps actually surfaced, not read as a generic checklist appended at the end.
The five questions underneath all ten steps
Strip away the mechanics and every one of these passes is really answering one of five questions:
- What technology is actually protected, in plain terms?
- How broad and defensible are the claims likely to be once prosecution is complete?
- Does that protection map onto the product, process, or application that’s commercially at stake?
- Does the company demonstrably own — not just appear to own — the rights needed to exploit the technology?
- What third-party rights, unpublished filings, or unpatented know-how still need to be investigated before anyone relies on the portfolio?
The goal was never a patent list. It’s a clear-eyed account of where protection is genuinely strong, where it’s still being shaped by prosecution, and where the real gaps sit — because that determines whether an IP position can carry the weight of the deal it’s being asked to support.
