Invention Disclosure Programs: Capturing IP Before It Walks Out the Door

Your engineers solve hard, potentially patentable problems every sprint — and tell no one outside their standup. Six months later, one of them joins a competitor with the whole solution in their head, or your team demos the breakthrough at a conference and quietly torches your foreign patent rights. An invention disclosure program is the fix: a deliberately lightweight system for capturing inventions when they happen and deciding, on purpose, what to do with each one. This guide covers how to design the intake form, staff the review committee, set incentives that don’t backfire, and run the whole thing at startup scale. It’s one pillar of the broader IP strategy and portfolio management guide.

The problem: inventions leak by default

Companies systematically underestimate how much patentable work they produce. Patentability doesn’t require a eureka moment — it requires something new, useful, and non-obvious, and “non-obvious” includes the clever data-pipeline fix, the sensor calibration trick, and the manufacturing tweak that cut scrap 30%. Engineers rarely self-identify these as inventions; they’re just Tuesday.

Left uncaptured, that work leaks through three drains:

1. Departures. When the engineer who invented your core technique leaves, the undocumented invention leaves too — along with your evidence of who invented what and when, which matters enormously in later disputes over ownership and inventorship.
2. Disclosure clocks. Under 35 U.S.C. § 102, an inventor’s own publication, public use, sale, or offer for sale starts a one-year grace period for U.S. filing — miss it and the invention is unpatentable. Most foreign countries have no grace period at all: a conference demo before filing kills European rights that same day. And the Supreme Court held in Helsinn Healthcare v. Teva (2019) that even a confidential sale can trigger the on-sale bar. Sales and marketing teams start these clocks constantly without knowing they exist.
3. Accidental trade-secret destruction. An invention you might have protected as a secret gets published in a blog post or open-sourced because nobody flagged it as valuable.

A disclosure program doesn’t make you file more patents. It makes every protect-or-don’t decision deliberate.

What is an invention disclosure program?

Three components, kept as light as your organization allows:

• An intake form any employee can submit in under 30 minutes.
• A review committee that evaluates disclosures on a set cadence.
• A decision pipeline with four exits for every disclosure:
  1. Patent — file a provisional or nonprovisional application.
  2. Trade secret — formally designate it, restrict access, and log it in a trade-secret register.
  3. Defensive publication — publish it to create prior art that blocks competitors from patenting it.
  4. Drop — decide, in writing, that it isn’t worth protecting.

The written drop matters more than it looks. A documented “we considered this and passed” is a strategy; an undocumented non-decision is how companies discover, mid-acquisition, that their crown jewel was never protected and nobody can say why. Buyers and investors probe exactly this — see how it surfaces in IP diligence for fundraising and M&A.

What should the disclosure form capture?

The form is a triage document, not a patent application. One to two pages:

• The problem. What couldn’t be done before, and why it mattered.
• The solution. How it works, at whiteboard depth. Attach diagrams, code snippets, or lab notes rather than demanding polished prose.
• Alternatives. What else was tried or exists, and why this beats it — this is early non-obviousness evidence, and it helps the committee judge whether competitors would want it.
• Dates. Conception, first working prototype, first internal demo. These anchor the record.
• Contributors. Everyone who contributed to the inventive concept, with roles. Inventorship is a legal question with real consequences — misnaming inventors can jeopardize a patent, as the disputes in our patent case-law archive show — so capture the raw facts now and let counsel sort the legal inventorship later.
• Bar-date risks. The section that should be printed in red: any past or planned publications, conference talks, demos, sales, offers for sale, beta programs, or investor decks containing the invention. This is what converts the form from paperwork into an alarm system, because of the § 102 clocks described above. A disclosure that reveals a launch in three weeks jumps the queue.

Keep the barrier low. Every required field you add costs you disclosures; engineers route around friction.

Who sits on the review committee, and how do they decide?

A workable committee is three to six people: a senior technical leader who can judge the invention, a product or business owner who can judge its commercial weight, and IP counsel (in-house or outside) who can judge protectability and bar dates. If you don’t yet have a regular patent attorney relationship, this is one of the classic triggers — see when to hire IP counsel.

Cadence: monthly or quarterly, matched to your disclosure volume, with an out-of-band fast lane for anything facing a bar date.

Scoring criteria should be tied to business value, not novelty-pride. A useful rubric asks:

1. Detectability — could we tell if a competitor used this? (Undetectable inventions often make better trade secrets than patents.)
2. Design-around difficulty — would a patent actually block anyone, or is there an easy alternative path?
3. Alignment — does it cover where the product line is going, or where it’s been?
4. Competitor appetite — would the companies you worry about want this?
5. Revenue exposure — how much of your (or their) business does it touch?

The most common committee failure mode is filing on what’s cleverest rather than what’s most valuable. An elegant algorithm nobody can detect infringing scores below a mundane mechanical feature every competitor will copy. Landscape data on where competitors are actually filing sharpens these calls considerably.

Incentives: what works and what backfires

What works:

• Filing awards. A payment when the company files on a disclosure — commonly a few hundred dollars to around $2,000 per inventor, sometimes with a second award at issuance. Modest, but it makes the invisible visible.
• Recognition. Patent plaques, an inventors’ wall, shout-outs at all-hands, an annual inventor dinner. In most engineering cultures, status beats cash — being a named inventor is a durable résumé line, and companies that celebrate it get more disclosures.
• Manager accountability. Disclosure counts reviewed at the R&D-org level, so managers prompt their teams after big technical wins.

What backfires:

• Volume-only bonuses. Pay per raw disclosure, regardless of quality, and you’ll get exactly what you paid for: a flood of trivial submissions that buries the committee and teaches everyone the program is a game. Pay at filing (a quality-filtered event), not at submission.
• Only rewarding patents. If trade-secret designations and defensive publications earn nothing, inventors lobby for filings the committee should reject. Reward the disclosure that leads to any protection decision.
• Winner-take-all contests. They discourage the collaborative disclosures that reflect how invention actually happens.

The fork in the road: patent, secret, or publish?

The committee’s core judgment for each keeper is which protection fits:

• Patent when the invention is detectable in competitors’ products, hard to design around, and worth 20 years of protection at filing-and-maintenance cost. Before spending on a filing, a search is worth the modest cost — see patent search before filing.
• Trade secret when the invention is invisible from outside (internal processes, manufacturing parameters, algorithms behind an API) and you can realistically keep it secret. The full trade-off analysis is in patent vs. trade secret; the key insight for the committee is that this choice is irreversible in one direction — a published patent application can never become a secret again.
• Defensive publication is the cheap third door: publish a technical disclosure so the invention becomes prior art, preventing anyone (including competitors) from patenting it, while you keep freedom to use it. Outlets include defensive-publication services and prior-art databases (IP.com’s Prior Art Database and the Research Disclosure journal are the traditional venues), or simply a well-dated technical blog post. Cost: roughly a hundred to a few hundred dollars, versus five figures for a patent. It’s the right answer for inventions you’ll practice openly but wouldn’t pay to enforce.
• Drop, in writing, for the rest.

The startup-scale version

If you’re pre-Series B, do not build enterprise machinery. Build a habit:

• A Notion or Google Form with the six intake sections above, linked in your engineering onboarding docs.
• A quarterly 30-minute review — founder/CTO plus outside IP counsel — with a standing rule that anyone can flag a bar-date emergency between meetings.
• A pre-launch checklist item: “anything in this release we should disclose or file on first?” This single question catches most § 102 clock-starters.
• Provisionals as the pressure valve. A provisional application is a comparatively cheap way to lock a filing date on a promising disclosure while the business case matures.

This slots into the broader early-stage sequence covered in the startup IP playbook. A Notion form and a quarterly review genuinely beat nothing by a wide margin — most startup IP disasters are capture failures, not strategy failures.

How do you measure whether the program works?

Three metrics tell you most of the story:

Also watch distribution: if all disclosures come from two teams, the other teams aren’t inventing less — they’re capturing less.

Tie it to assignments and exits

A disclosure program only pays off if the company actually owns what’s disclosed. Every employee and contractor should have signed a present-tense IP assignment (“hereby assigns,” not “agrees to assign” — the difference decided Stanford v. Roche (2011)) before contributing. Then close the loop at both ends of employment: onboarding surfaces prior inventions the new hire claims as their own, and the exit interview includes a final disclosure sweep — “anything you’ve invented here that isn’t written down?” — plus a reminder of surviving confidentiality obligations. That exit conversation, documented, is cheap insurance against the most expensive version of this problem: watching your invention ship from a competitor’s building.

The bottom line

An invention disclosure program is a net for value your company is already creating and currently losing. The mechanics are deliberately boring — a short form, a small committee on a regular cadence, four possible exits (patent, trade secret, defensive publication, documented drop) — because boring is what people actually use. The urgency is real: § 102 disclosure clocks, foreign forfeiture, and employee departures don’t wait for you to get organized. Start at whatever scale you are — even a shared form and a quarterly half hour — and let the machinery grow with the portfolio.

This article is general legal information for educational purposes only. It is not legal advice, does not create an attorney-client relationship, and may not reflect the most current law in your area. Patentability, inventorship, and ownership questions turn on specific facts. For advice about your situation, consult an attorney licensed in your jurisdiction.