- When skilled attorneys keep falling behind, the cause is design, not effort: the reactive model was built for a rate of invention that no longer exists.
- At $15,000 to $25,000 per traditional filing, the budget is consumed by inventions that already arrived, leaving nothing for the strategic work.
- The most expensive loss is the invention that never became a disclosure; it leaves no record and cannot be recovered.
- Run three measurements: disclosure growth against R&D growth, days from arrival to decision, and whether each pass and continuation was deliberate.
- Adding people to the current model will not close the gap; the model itself has to get cheaper.
- Run the diagnostic now, not six weeks before a funding round; capture losses become permanent every quarter you wait.
A disclosure that landed three weeks ago is still sitting in your inbox, buried under something more urgent that arrived the next day, and the one after that. The queue behind it is longer than it was last quarter, and longer than the quarter before. You are doing the same job, applying the same judgment you always have, and the backlog grows anyway.
Your engineers are inventing faster than your IP function can capture, and the gap grows wider every quarter. The reason has nothing to do with your team and everything to do with the cost of traditional filing, which keeps the function locked in reaction. The result is a weaker portfolio at every stage: slower to exclude competitors, thinner on freedom to operate, and less able to support the company’s strategy. A fundraise or an acquisition only puts a price on what was already true. This piece gives you a three-part diagnostic to find where your pipeline is leaking, and the move that closes each leak.
Why sound teams still fall behind
When skilled attorneys running a sound process keep falling behind, the cause isn’t effort or discipline. It’s design. The IP function most companies run today was built for a rate of invention that no longer exists, and over the last two or three years your engineers started producing patentable work faster than the model you inherited was ever meant to handle.
The flaw is foundational. Your function can only react. Every dollar of the IP budget, and nearly every hour of the team’s time, goes to drafting and filing applications on the inventions already in front of you, the evidently valuable ones that arrived on their own. That work is expensive. At the fifteen to twenty-five thousand dollars a traditional outside-counsel filing runs, a portfolio of any size consumes the whole budget just keeping pace with what has already been disclosed.
The cost isn’t what stops you from filing a good invention. Show most IP leaders a clearly valuable, commercially important disclosure and they will find the money. The damage is subtler. The cost leaves no budget or hours for the work that nobody ever schedules. That is the work of going into engineering to surface inventions that were never disclosed, deciding where the portfolio needs depth before a competitor moves in, and pursuing the variants and continuations that turn a handful of filings into a defensible position.
That strategic work is what builds lasting value, and a reactive function never reaches it. The team knows it matters, but the capacity it would take is already spent on filing. The portfolio ends up assembled from whatever happened to arrive, instead of built toward where the business is going.
What AI changed
None of this is new. A reactive IP function has always left inventions uncaptured and the portfolio thin in places, but at the old pace of invention the loss was slow enough to ignore. You lost a little ground each year, and each year you mostly made it back.
AI ended that. Engineering now produces patentable work at a rate the same team wouldn’t have recognized three years ago: code written and rewritten with AI assistance, designs iterated in a fraction of the time, experiments running in parallel instead of in sequence. The reactive function on the other side didn’t change. Its budget didn’t double and its hours didn’t grow. The gap that used to be a rounding error now widens every quarter and won’t close on its own. AI now pushes far more invention through the same leak, and the losses have grown to match.
The loss you never see
The most expensive loss is the one you never see. A deprioritized filing at least leaves a record: you evaluated it, made a call, and can defend that call later. The invention that never became a disclosure leaves nothing behind. It never entered a queue, never got a docket number, never registered as a decision at all. You can’t recover it once a competitor files first or once the window your own disclosure opened has closed, and you can’t even count what you gave up. It keeps accumulating, and by the time it matters, two or three years of it have piled up: a whole category of work your engineers explored and your company never claimed.
The loss isn’t only the inventions you never captured. A patent portfolio is supposed to do substantial work: keep competitors out of the space you built, defend your freedom to operate, and back the moves the business wants to make. A reactive one, assembled from whatever happened to arrive, does all three worse, and it does them worse every day, not only when someone is underwriting you.
A transaction is just the moment that weakness gets a price. When you raise, go public, or get acquired, the gaps stop being an internal concern and become diligence, examined by people with leverage over the outcome. This isn’t a recent development; IP has always been a diligence item in markets where it matters. Having a coherent IP story removes a reason for investors or acquirers to say no. As everyone starts inventing and filing faster, those reasons multiply: more prior art, more competitors holding positions in your space, more gaps a careful diligence team can surface. A portfolio that maps without friction to your product clears those objections before they come up. One with holes invites more diligence questions.
Find your leak
The path from invention to filed patent runs through three stages: an idea becomes a disclosure, a disclosure becomes a decision, and a decision becomes a deliberate position in the portfolio. A reactive IP function springs a leak at one of the three, and each one has its own measurement. Run all three; the stage that reads worst is where your gap is widest.
Capture: is the invention even reaching you?
Compare the growth in disclosures over the last two years against the growth in R&D, by headcount or by spend. If engineering grew forty percent and your disclosure count is flat, the inventions aren’t arriving; they’re being lost before anyone writes them down. This is the leak that shows up in no queue, because nothing ever entered one, and that is what makes it the most expensive kind.
Throughput: can you act on what reaches you?
Measure how long a disclosure sits between arriving and getting a decision. Days is healthy. Weeks means disclosures are aging in an inbox while the team is so buried in active filings it can’t process its own pipeline. The inventions reached you; they’re just sitting long enough that your engineers notice.
Protection: are you building deliberately, or just keeping up?
The obvious number, conversion rate, can mislead. A low filing rate isn’t automatically a problem, since holding an invention as a trade secret can be the right call. The test is whether your decisions are deliberate: can you give a strategic reason for each disclosure you passed on, and when did you last file a continuation or pursue a variant because the portfolio called for it, rather than in reaction to whatever arrived? If the passes are the budget running dry, and the continuations aren’t getting filed at all, then cost is making your strategy for you.
None of this should require new tooling. You can answer all three this week from data you already have, and what you find tells you which leak to close first. If you truly can’t answer one of these questions, that might be your answer.
Closing each leak
The reflexive solution most attorneys jump to once they’ve identified an issue is to hire another attorney or send more work to outside counsel. It would help; more capacity does close the gap. But the budget isn’t there, and more reactive spending only buys more of the same. You file the obvious applications a little faster and still have nothing left for strategy. Adding people to the current model won’t fix it. The model itself must get cheaper.
If the leak is capture, the fix is the cheapest of the three, because it isn’t a budget problem. Stop asking engineers to halt what they’re doing and write up an invention they’ve mentally moved past. Put the prompt where the work already happens: a standing question in design reviews, or a named person on each team who knows when to flag something. It adds no process; it just shortens the trip from where an invention happens to where it gets written down.
If the leak is throughput, the fix is to stop personally touching every disclosure. First-pass triage (prior art, routing, the clear no’s and the obvious yes’s) doesn’t need your judgment; it needs to happen fast and consistently, which is what AI-assisted triage now does well. You’re left with the smaller set of decisions that need a lawyer’s judgment. The queue stops aging because the work clogging it never needed you in the first place.
If the leak is protection, the fix is the one that requires changing the economics. Continuations go unfiled and counsel never weighs in on strategy because every available dollar is committed to getting applications on file at full freight. Lower the cost per filing and two things free up at once: budget to protect more deliberately, and your attorneys’ time to spend on where the portfolio should go, not just on what arrived.
The economics are what must change, and that shift is beginning. AI-assisted drafting paired with attorney review and filing is pulling the cost per application well below hourly outside counsel, without giving up the legal accountability that every filing still needs. The budget math that has held the function in reaction for years is finally moving, and a filing volume that keeps pace with how fast your engineers invent is becoming possible.
Run the diagnostic now
Speed is already your advantage. Your engineers reach the market ahead of competitors still on the old cadence, and an IP function that keeps pace turns that head start into protection a competitor must overcome twice, in the market and in the portfolio.
Run the diagnostic now, not when a funding round is six weeks out. Every fix is cheap today and impossible in hindsight. A throughput leak you can close in a quarter; the invention that went unrecorded two years ago is beyond recovery. That asset isn’t behind. It’s gone. Every quarter you wait, more of it becomes permanent.
Related: IP Diligence Readiness
For IP leaders who’ve run the diagnostic and found the gaps: IP Diligence Readiness: Pass the Audit, Win the Narrative walks through what a diligence team will surface in a portfolio assembled reactively, and how to close those gaps before a funding round, acquisition, or IPO turns them into leverage against you.