Every company has someone like this. Not necessarily the most senior lawyer, or the highest-ranking person on the legal team. It's the person everyone already knows to ask when the awkward case shows up: the contractual exception that isn't in any policy, the unusual clause that has already appeared three times in similar contracts, the right way to negotiate with that difficult supplier. When this person leaves the company, HR opens a position, interviews candidates, closes the hire in a few weeks and considers the problem solved. Except the role was filled. The legal knowledge that supported those decisions was not.
This gets treated as a recruitment problem because that is easier to measure than what actually happened: the legal department lost access to a set of opinions and decisions that never existed in any document, template contract or system, only in the head of the person who left. This kind of knowledge has no direct relationship with tenure or seniority. A mid-level lawyer who resolves a specific type of contractual exception every month builds up this asset just as much as a legal manager with fifteen years at the company. The difference is that nobody asks either of them to document how they do it. It applies to any management area, but in legal, the risk of a poorly grounded decision, or of a negotiation redone from scratch, costs more than in almost any other team.
The problem shows up first as a queue, not a resignation
An article from SHRM Executive Network, published on September 21, 2026, describes exactly this symptom before anyone even resigns. According to Surojit Chatterjee, CEO of the agentic AI platform Ema, internal questions, including contractual and legal ones, used to take up to five days to be resolved, not for lack of an answer, but because the answer was concentrated in a few people, and everyone had to wait their turn. Five days is not how long it takes to think of an answer. It is how long it takes for someone with the right information to have a free slot in their calendar, and in legal, five days can be the entire window available to answer a client or file a response in a lawsuit.
The article also cites two companies that have already stopped treating this as an HR issue and started treating it as an infrastructure issue: Hitachi, with a center dedicated to capturing institutional knowledge through generative AI, and ABB, which built what it calls an industrial knowledge vault on top of Azure OpenAI Service, including legal and compliance areas. Neither solved the problem by hiring faster. Both solved it by no longer depending on specific people being available at the right moment.
The wrong question is "who will replace them"
The question most companies ask when a lawyer or legal manager leaves is who will replace that person. The question that should be asked is what that person knew that isn't written down anywhere, which internal precedent they applied without recording it, which clause they always asked to adjust before closing a contract. They are different questions because they have different answers: the first is solved with a hiring process; the second can only be solved if the company has already created, before the departure, some way to record the legal reasoning behind decisions, not just the final opinion, but the why behind it.
This is where workload management for the legal and advisory team blurs into risk management. An outdated internal policy is still a document, even if a bad one. Tacit knowledge that was never recorded isn't even that. It is a gap nobody knows exists until they need it, usually at the worst possible moment: in the middle of a negotiation or an urgent response.
The cost isn't the open position, it's the time spent rediscovering
Picture a mid-sized legal department with ten professionals, each handling on average three or four contractual exceptions a week that aren't documented in any policy, only in the experience of whoever solved them before. When one of these people leaves, a realistic estimate of how long it takes the team to rediscover, on its own, the equivalent of what they knew sits between two and four months, depending on the complexity of the area. In that interval, the work doesn't stop: someone takes over, gets wrong what had already been gotten right in a negotiation, repeats a question that already had an answer in a previous opinion, treats as a new case a clause that was already routine. The real cost doesn't show up in the payroll of the hiring process. It shows up in the time the whole team takes to know again what it already knew.
The most common miscalculation is measuring the cost of turnover by time to hire. The true cost is the time spent rediscovering, and most companies don't even measure it, because there is no way to measure what was never recorded in the first place. This applies to any management area, but it weighs more in legal, where rediscovering on your own can mean repeating a mistake that was already expensive the first time.
At AskLisa, every legal answer validated by a specialist becomes knowledge that whoever comes next can consult, not because someone remembered to write it down before leaving, but because the record is created at the moment the answer is given. The person may leave. The answer doesn't.

