You are sitting at the far end of a conference table in a two-story office building off the 101, inside one of those mid-Peninsula office parks where a hundred-person semiconductor company can sit next to a multi-billion-dollar venture fund and the only difference is the size of the sign out front. The room is professional but spare. No art. No mahogany. A whiteboard still carrying the ghost of last quarter's projections, half-erased.
The VP of Finance is presenting. She has chosen the seat next to the CEO, which you have noticed, and she is walking the board through a spreadsheet projected on the screen at the far wall. Three thousand patent files, organized by internal reference number, color-coded by status. Green means current. Yellow means thirty days out. Red means urgent. She fumbles early in the presentation, loses her place in a column, but recovers with confidence. She knows these numbers. She has spent four months building this tracker, and it shows.
The two board members dialing in from San Francisco and Vienna are small rectangles on a wall-mounted monitor. The executives in the room have laptops open. You do not. You have a legal pad, bad posture, and the quiet resignation of someone who warned them this would happen eighteen months ago.
The company is Verada Semiconductor. A hundred and forty people designing AI-accelerated inference chips for the two or three customers whose names you would recognize from earnings calls. In the semiconductor food chain, Verada is the pilot fish. Fast, specialized, essential to the larger animal. The company's entire enterprise value lives in its patent portfolio. And right now, a finance executive is presenting that portfolio to a board that does not know the difference between a filing tracker and a prosecution strategy.
The VP finishes. She looks at the CEO. He nods. "We have visibility," she says. "We are monitoring closely. We understand what needs to get done and when."
The board is reassured. You are not.
Six weeks earlier, the IP partner called. She had been patient for a year, maybe longer than she should have been. She had built this portfolio from the ground up over four years. Sixty active matters. Chip architecture, on-chip memory designs, a novel inference engine that several larger players had already noticed. That attention was not a compliment. It was a threat posture. Two of those larger players had filed Notices of Investigation. They had nine-figure litigation budgets and institutional patience. Verada had neither.
What Verada also did not have was the ability to pay its patent lawyers.
The Series B had been raised at the peak of AI enthusiasm. The burn rate was built around landing a major customer by Q3. Q3 came and went. By September the company was managing cash week to week, prioritizing R&D salaries and letting other invoices age. The patent firm's invoices aged with them. You had been telling management since the previous spring that this pattern would not hold. That the firm would eventually withdraw. That this was the kind of risk that belonged in front of the board, not buried in accounts payable.
Management disagreed. The CEO, an MD by training who leaned on that credential when business questions made him uncomfortable, said you were being alarmist. The relationship with the firm was strong, he said. They would wait. Raising the issue with the board would create unnecessary stress during an already difficult fundraise.
The firm did not wait. The IP partner did not call the CEO. She called you, because you were the person she trusted to understand what she was actually saying. In the careful language of one lawyer speaking to another, she said the files would be transferred within thirty days. The representation was over.
Now the gap was real. And the VP of Finance had stepped into it with a spreadsheet.
A spreadsheet tells you when something is due. It does not tell you what to do about it.
You wait until the board's initial questions are finished. The questions are soft. How many matters are active. Whether the foreign associates have been contacted. Whether there is a timeline. The VP handles them well. She is prepared for what she was prepared for.
Then you ask your question.
"What is our prosecution strategy for the European office action on the inference engine architecture?"
She checks the spreadsheet. "That one is flagged yellow. Sixty days out."
"I know when it is due. I am asking what we are going to do about it. Whether we contest the examiner's claim construction. Whether we file a divisional. What our theory is for protecting the broadest claims without creating estoppel problems in other jurisdictions."
The room goes quiet. Not the dramatic silence of a television courtroom. The administrative silence of people realizing that the conversation just changed categories. The VP looks at the CEO. The CEO looks at his laptop. The board member in San Francisco leans forward in his rectangle.
You are not trying to embarrass anyone. You are doing what you have done for over twenty-five years. Lawyers learn a framework in their first year of law school called IRAC. Issue, Rule, Application, Conclusion. The entire discipline lives in the third step: Application. That is where you ask why. Why does this fact matter under this rule? Why is this answer incomplete? That single word, repeated patiently, is the difference between describing a problem and solving one.
You ask the next question. The VP has told the board that some filings are being allowed to lapse deliberately. "Focusing on critical markets," she said. "Letting the non-critical ones go." You ask which filings have been abandoned and on what basis.
The answer is where the quiet resignation settles into your chest. Several abandoned filings had years of remaining patent life. The VP looked at the titles and made judgments about which markets mattered. But a patent title tells you the subject of the application. The claims tell you what you own. Those are different things. Letting a filing lapse in one jurisdiction because the market seems unimportant today can destroy your ability to enforce related claims elsewhere. The word for that is estoppel. The VP has never heard it. She has been making irreversible decisions about the company's most valuable assets using criteria that had nothing to do with the assets' actual scope.
A patent title tells you the subject of the application. The claims tell you what you own. Those are different things.
Each lost claim is a building block pulled from the wall that protects Verada's enterprise value. And nobody on the board knows it is happening, because the spreadsheet shows green.
You request a closed session. General counsel and directors only. This is your prerogative as corporate secretary when fiduciary obligations are at issue, and the executives in the room know it. The CEO glances at the VP. She gathers her laptop. The other executives follow. The door closes.
You present three findings to the board. You do not use the spreadsheet.
The patent law firm withdrew because management did not pay the invoices for over a year. You flagged this risk eighteen months ago. Management characterized your concern as alarmist. A finance executive filled the gap with a tracking spreadsheet and has been making prosecution decisions without patent counsel. Filings with years of remaining life have been abandoned based on market assumptions rather than claim analysis. Some of those decisions cannot be reversed. The board was told the situation was under control. The spreadsheet supported that impression. The underlying reality does not.
You are careful. You do not accuse anyone of dishonesty. There is no fraud here. There is something more common and, in its way, more dangerous: a competent person doing real work on the wrong problem, performing for an audience that cannot tell the difference. The VP's spreadsheet is genuinely useful as an organizational tool. But an organizational tool is not a legal strategy. Tracking deadlines is not managing prosecution. And the board's fiduciary duty runs to the outcome, not to the effort.
The board member in Vienna unmutes. "What do you recommend?"
Three things. Retain specialized patent counsel immediately, even on a limited engagement, to triage the portfolio. Conduct a claim-by-claim review of the active matters to determine actual scope of protection, not just filing status. And establish a protocol so that any changes to the patent portfolio, intentional or otherwise, are reported to both management and the board as material events.
The board adopts all three. Within two weeks, specialized counsel is retained. The review identifies four filings abandoned in error, two of which are reinstated within the grace period. The claim analysis reveals that Verada's inference engine patents are broader than anyone had realized, which becomes a significant factor in the customer negotiation that closes three months later.
After the meeting, the VP stops you in the hallway. She is not happy. She spent four months on that spreadsheet. She tracked every deadline. She contacted every foreign associate. She organized every file. She worked hard. She wants you to know that.
You tell her you know. You mean it. Then you tell her something she does not want to hear. In corporate governance, the legal standard is not "did you try?" It is "did you act with the care that a reasonably prudent person would exercise under similar circumstances?" A reasonably prudent person would not have made prosecution decisions without a patent attorney. That question requires no specialized expertise. It requires the willingness to ask it.
She walks away. You stand in the hallway for a moment.
The most dangerous problems in a company are not the ones that are hidden. They are the ones that are visible, organized, color-coded, and maintained with great discipline.
Every significant failure I have observed in over twenty-five years of corporate practice shares one feature. Somewhere in the story, there is a person who worked very hard at something that did not need to be done. And nobody asked why.