My executive assistant left. The role was not refilled.
…because he learned the tools before closing the recruitment.
A CFO peer, running a service company, told me this story recently. His executive assistant had resigned to take a manager position at another firm. Clean departure, well planned, no crisis. Recruitment started immediately; tight budget, role judged indispensable, job description reused almost unchanged.
Between candidate interviews, a friend introduced him to Claude and a few common productivity tools: writing assistants, calendar automation, meeting transcription and summarization, AI-assisted document management. A few hours one weekend. Then a week. Then a month.
By the end of the month, he closed the recruitment and informed HR the role would not be refilled.
What the decision reveals
This is not a story about cost-cutting. It is a story about recomposition.
The question most leaders ask when someone leaves is: how do I find an equivalent hire, fast, at the best price? The question this CFO answered is structurally different: is this role, as it existed, still the right component for what I need to produce today?
The first question reproduces the existing organization; the second reinterrogates it at every window a departure opens. Across ten years running a service SME, I can count on one hand the times I asked the second question before the first. The replacement reflex dominates, and it dominates even among leaders who consider themselves adaptive.
Why identical replacement almost always wins
Identical replacement is easy to justify: fast, apparently risk-free, no change to the org chart, no conversation to hold with the team. The role existed; therefore the role must continue.
Teece (1997) showed nearly thirty years ago that durable competitive advantage does not come from what a firm owns but from its capacity to reconfigure resources when context shifts. Most service SMEs still operate as if competitive capital lived in the list of open roles, not in the capacity to recompose them.
Eisenhardt and Martin (2000) extended this thesis by demonstrating that recomposition is a teachable competency, not a talent reserved for so-called visionary leaders. It is built through explicit routines: a question asked at every departure, a decision format that frames the arbitrage within a tight deadline. Subsequent work on resource allocation routines (Bidault et al., 2023) links their formalization to sustained higher growth in European service SMEs.
What made the decision possible
The departure opened a window of arbitrage that would not have existed otherwise. Most roles are never reinterrogated between two vacancies; they are validated by usage.
At the same time, a new component had become available in the environment. Generative AI tools now make feasible, at usable quality, tasks that previously required qualified human work: correspondence writing, meeting summaries, document organization, reformulation, structured administrative follow-up. Without this component, the non-renewal question would not have had a workable answer.
The budget constraint acted as a trigger. In times of comfort, the potential saving from a non-refilled role stays invisible; pressure makes it legible.
The question most organizations fail to ask
In the majority of service SMEs I observe, no routine formalizes the inquiry at the moment of departure. The HR form asks: “what profile are you looking for to replace X?” It does not ask: “should X be replaced, and if so by which component?”
This micro-shift in phrasing produces massive effects. As long as the question is about profile, the answer will always be a CV. When the question becomes about component, the answer can be a CV, a freelancer, a tool, an internal redistribution, or nothing.
The decision carried at the right level
What this CFO did not do deserves naming: he did not ask his HR director or his team to automate the assistant’s tasks. He invested his own time to build competence on the tools. That is the difference between delegating recomposition, which fails most of the time, and carrying it at the leadership level, which produces decisions.
Eisenhardt and Martin were explicit on this point: dynamic capabilities anchor in leadership team routines, not in HR procedures. Delegating the review of a role to the function that manages roles amounts to asking that function to decide on its own suppression. The structural bias is guaranteed; the status quo wins every time.
What you can install
A simple routine is enough to change the reflex. At every announced departure, before a role is opened, the leader or a small committee takes a few hours to answer in writing two questions: are the deliverables expected from this role still relevant as they stand, and among those that are, which component (human, freelancer, tool, AI, redistribution, suppression) is now best suited to deliver them?
The exercise must happen before the job ad is drafted, not after sourcing has started. Once recruitment is live, organizational momentum makes it nearly irreversible; the candidates received, the expectations already communicated internally, the time already invested lock in the decision well before it is formally made.
The reflex test
One indicator is enough to place your organization. Over the past twelve months, what share of departures led to a strictly identical rehire, with no other option formally examined? Above 80%, the organization reproduces. Below 50%, an arbitrage routine is already active, consciously or not. Most service SMEs I observe sit above 90%.
The question to ask this week
To check where your organization stands, one question at your next leadership meeting is enough: over the past twelve months, how many departures were followed by a strictly identical rehire, with no other option formally examined? The answer, obtained in five minutes, is worth more than a long diagnostic. It will tell you whether you run an organization that reproduces or an organization that recomposes.
Lionel Jaquet — CVO @tebicom · Responsive Management · DBA Candidate @GEM