Counting the Handoffs: How a Single Patent Filing Moves Between Systems

Ask an intellectual property operation where its time goes and the answer will name systems. The docketing platform is slow. The drafting tool does not integrate. The client portal requires duplicate entry. Each of these is usually true, and none of them is the constraint.

The constraint is the set of moments when work changes hands. At each of those moments the file changes owner, changes format, and waits. The waiting is the part that does not appear on any dashboard, because a queue that sits between two systems is owned by neither.

Counting handoffs requires no instrumentation and no project. Take one matter type, follow one real matter through it end to end, and write down every point at which the work moved from one person or system to another. Most operations find six. The exercise takes an afternoon, and it produces something no platform comparison can produce: a map of where the time actually sits. The reason to do it before buying anything is straightforward. A platform placed on an unexamined process automates the wait rather than removing it, and the automation makes the wait harder to see afterwards.

The first transition is the arrival of the instruction. It comes by email, through a portal, or by telephone. It carries a deadline, an invention description, and an assumption about scope. The assumption is rarely written down, and it is the thing most often renegotiated later. The instruction enters as prose and must leave as structured data. Someone performs that conversion, and no system records that they did.

The second is conflict and intake clearance. The matter is checked against the conflicts database and opened. This is the one handoff that is almost always instrumented, because a missed conflict is a known catastrophic risk. It is worth including precisely because it shows what a measured handoff looks like, and how different it is from the others.

The third is the drafting assignment. The matter moves to an attorney or an outside drafter, with the file, the disclosure, and whatever context exists. Context that lives in the assigning partner's head does not move. This is the most expensive silent handoff of the six. Work restarts here whenever the transfer was incomplete, and the restart is invisible because it looks like drafting.

The fourth is review and client approval. The draft goes to the client, comes back with comments, and is revised. This cycle repeats an unpredictable number of times and is the largest single source of calendar variance in the whole journey. It is rarely counted as a handoff at all, because it involves the client. It is nonetheless the transition where the file waits longest.

The fifth is filing preparation. The approved draft becomes a filing package: forms, fees, formalities, signatures, sequence listings where relevant. It moves from the attorney to a paralegal or a formalities team, and it moves as documents rather than as data. Format changes here. Whatever structure existed in the drafting system is flattened into what the office requires, and the mapping between the two exists in one person's practice.

The sixth is submission and docket confirmation. The package is filed, a receipt returns, and the docketing system is updated. The matter is now, from the firm's point of view, in a stable state. The loop closes only when someone confirms it. The interval between the receipt arriving and the docket reflecting it is unmeasured in most firms, and it is where the missed-deadline risk concentrates.

Leadership reads cycle time: instruction to filing, measured from two timestamps that sit at either end of the journey. That number is real, and it is almost useless for improvement, because it aggregates six intervals of very different character into one figure that can only move for reasons nobody can name.

The paralegal sees something else entirely. They see which attorneys return formalities complete and which return them requiring reconstruction. They see that a particular client's approvals arrive in a predictable rhythm and another's do not. They see the second week of every month absorb whatever slack existed. None of this is written anywhere, and all of it is a measurement.

The gap between those two views is the finding. A firm that instruments its transitions is not buying new information. It is writing down what one person already knows, in a form that survives that person taking a holiday.

Once each transition carries a timestamp and an owner, four questions become answerable, and all four are currently guesses. Which transition holds the longest queue, in median rather than in average. Which transitions vary most, since variance costs more than duration. Where rework enters, and whether it traces to an earlier incomplete transfer. And what each transition costs, once internal time is attached to it.

The fourth question is the one that changes a procurement conversation. A tool that improves drafting speed is addressing the third transition. If the queue is at the fourth, the tool will perform exactly as advertised and the matter will not move faster.

To run the exercise, pick the matter type that carries the most volume, not the one that causes the most complaint. Follow one real matter that has already completed, so the timestamps exist rather than being estimated. Record, at each transition, who held the work, what format it was in, and how long it sat. Then repeat with two more matters of the same type, because a single matter shows a path and three show a pattern.

Do this before the next platform demonstration rather than after it. The demonstration will be more useful, and considerably shorter.