Two Bad Extremes When FTO Feels Too Expensive

A startup in Shenzhen is preparing to launch a new product and asks U.S. counsel for an FTO quote. After seeing the number, the team wonders whether it is too early to spend that budget before the product is even on the market.

Two positions often emerge. One side says: do nothing now and deal with patent risk after the product becomes successful. The other says: let R&D search for a few days, find several similar patents, and then conclude internally that the product is safe.

Both approaches are risky. Doing nothing can miss the stage when redesign is cheapest. Treating a quick search as a non-infringement conclusion turns a preliminary screen into something it was never designed to be.

What an In-House Screen Can Actually Accomplish

A disciplined first-pass review can still be useful. A DIY FTO guide published by Patent AI Lab, for example, breaks the process into product decomposition, country selection, keyword and classification searching, legal-status checks, claim review, and a simple claim comparison table. That is a sensible way to think about early-stage screening.

At this stage, the company is trying to answer practical questions: Are there patents close to the product's core functions? Which competitors are filing in this area? Which product features repeatedly appear in active claims? Are there obvious obstacles that should be addressed before code is frozen, tooling begins, or the supply chain is locked?

The earlier those questions surface, the more room the team usually has to change direction.

1. Break the Product Down Before You Search

The first search term should usually not be the product name. Marketing language—“smart control platform,” “AI optimization system,” “automatic cleaning device”—rarely maps cleanly onto patent claim language.

A better starting point is a technical feature tree. For a software product, identify where data enters, how it is processed, what module performs the processing, and what output results. For a mechanical product, describe not only the components but also their positions, connections, movement, operating states, and control relationships.

This exercise often exposes product facts that the team itself has not yet documented clearly. That matters because FTO analysis is only as good as the product facts being compared to the claims.

Patent documents may describe the same component using very different terminology. Once the team finds a reasonably close document, the search can expand through CPC classifications, assignees, inventors, cited references, and family members.

The Cooperative Patent Classification (CPC) is jointly managed by the USPTO and EPO and can be a powerful supplement to keyword searching. Useful free tools include Google Patents, the USPTO's Patent Public Search, and the EPO's Espacenet.

For an internal screen, the goal is not to prove that the search is exhaustive. It is to build a more disciplined candidate set than “these five patents look similar.”

3. Read the Claims, Not Just the Title, Abstract, or Drawings

This is where many internal searches go wrong. Titles, abstracts, and drawings help explain the technical theme, but they do not define the infringement question. The USPTO itself notes that infringement analysis primarily involves comparing patent claim language to the accused product or process; the USPTO does not decide infringement, and courts make the final determination.

Start with the independent claims and then review dependent claims that are relevant to the product. A dependent claim includes all limitations of the claim from which it depends and adds further limitations. If the product clearly does not satisfy a required limitation of the base claim, that will ordinarily also defeat literal infringement of a dependent claim that incorporates that limitation.

4. Build a Simple Evidence-Based Claim Map

For each limitation, an internal team can use a restrained status such as:

  • Clearly present
  • Possibly present
  • Not identified in the current product
  • Insufficient product information to determine

Next to each entry, record the evidence source: product specification, engineering drawing, code logic, photo, video, test result, supplier document, or other material.

This kind of table is far more useful than a patent list with a similarity score. Even when it cannot support a legal conclusion, it shows where the potential overlap may be and where the product record itself is incomplete.

5. A Claim-Centered Review Is Not the Same as Reading Claim Words Literally

Claim scope can depend on the specification, prosecution history, and applicable law. A term that looks broad on its face may have been narrowed by the intrinsic record. A product may also use a different label for a structure that performs substantially the same role.

If a product, under a reasonable construction, clearly lacks a required claim limitation, that may substantially reduce literal-infringement risk for that claim. But an internal memo should be cautious about converting that observation into an absolute “no infringement” conclusion. Product facts may be incomplete, claim construction may be disputed, and U.S. law can require a separate doctrine-of-equivalents analysis on a limitation-by-limitation basis.

6. Legal Status Is More Than a Green Checkmark

For important targets, verify official records and look beyond a database status label. Relevant questions can include application status, patent term, maintenance fees, related family members, continuations or divisionals, reexamination or other proceedings, and ownership information.

Google Patents is excellent for rapid screening, but it should not replace authoritative records for important decisions. A U.S. utility patent can expire for failure to pay a maintenance fee, yet the USPTO provides a petition process to reinstate an expired patent when statutory and regulatory requirements are met.

Likewise, one abandoned application does not mean the entire family is irrelevant. The same technology may continue in a continuation, divisional, or other family member. One expired patent does not make every improvement patent expire on the same date.

7. Pending Applications Should Be Tracked Separately

A pending application cannot be enforced as an issued patent, but its claims may continue to change and it may mature into a patent after the product reaches the market. Early screening can therefore place relevant pending applications on a watch list rather than mixing them with currently enforceable patents.

There is also a narrower U.S. issue worth remembering. Under 35 U.S.C. § 154(d), an issued patent may in defined circumstances support a reasonable royalty for certain activities occurring between publication and issuance, including requirements concerning actual notice and substantial identity between the published and issued claims.

8. Know What the Internal Team Should Escalate

The best role for in-house screening is to make the next professional review more focused. It can organize product facts, identify target patents, flag uncertain claim limitations, and reduce duplicate data collection.

But it does not replace a professional search, formal claim construction, equivalents analysis, prosecution-history analysis, legal-status verification, or a legal opinion. Whether a project should be escalated depends on factors such as product value, concentration of risk, target market, development stage, financing or M&A requirements, and the consequences of being wrong.

A useful division of labor

Let the internal team reduce noise, clarify the product, and identify the patent questions. Let qualified professionals spend their time on the issues that require legal interpretation and accountable judgment.

9. A Formal FTO Opinion Is Not a Mandatory Certificate for Every Product

None of this means that every product must obtain a formal written FTO opinion before launch. In the United States, 35 U.S.C. § 298 provides that failure to obtain advice of counsel, or failure to present that advice to the court or jury, may not be used to prove willful infringement or intent to induce infringement.

The value of a formal opinion is different: qualified counsel applies legal analysis to a sufficiently developed factual record and assumes professional responsibility for that analysis. It is not a guarantee that a court will never find infringement.

The Bottom Line: Save Professional Time, Not Professional Judgment

The real value of a company doing the first round itself is not to produce a home-made clearance opinion. It is to sort the problem earlier: which results can be deprioritized, which product facts are missing, which patents justify deeper analysis, and which questions the internal team should not answer on its own.

When budget is limited, disciplined screening can reduce low-relevance results and repeated fact gathering. What it cannot do is turn an initial search into a substitute for professional judgment.

In-house FTO screening can reduce the cost of getting to the right questions. It should not be used to manufacture certainty where the legal analysis has not yet been done.

Sources & Further Reading