When the CEO leaves, what must they truly leave behind?

Stepping down as CEO does not necessarily mean leaving the field.

Accumulated knowledge, relationships, internal reputation, Board access and the ability to influence people do not disappear the moment a title changes.

And it is precisely here that a succession ceases to be a simple handover and becomes a matter of the architecture of authority.

Over the last few months, a number of large organisations have made this tension particularly visible.

Adobe has announced that from 1 December 2026, Anil Chakravarthy will become President and CEO, while Shantanu Narayen will transition to the role of Executive Chair and will continue to work closely with the new CEO during the transition and transformation of the company.

AIG has opted for a different configuration. Peter Zaffino, after moving from CEO to Executive Chair, will also step down from the executive chairmanship and the Board on 15 September to become Senior Advisor to the CEO. The chairmanship of the Board will instead pass to John Rice, previously Lead Independent Director.

At first glance, both solutions seem to answer the same need: not to squander experience and continuity during a major transition.

But structurally they are very different.

Executive Chair and Senior Advisor are not the same thing

An Executive Chair remains within the formal architecture of governance.

A Senior Advisor, on the other hand, can remain close to the CEO without necessarily holding a position on the Board and without wielding the same formal authority.

The distinction matters because the real question is not just how much of the predecessor's experience is useful to retain.

The question is:

How much authority must remain alongside that experience?

Research into CEO succession has long warned us that the predecessor's continued presence is not neutral.

Timothy Quigley and Donald Hambrick, analysing 181 successions in high-tech companies, found that the former CEO remaining as Board Chair can limit the successor's discretion and reduce their ability to drive strategic changes.

This does not mean that keeping the predecessor is necessarily a bad choice.

It means something more interesting:

Continuity produces value only when the boundary of authority remains legible.

The problem is not only what is written down

Decision rights can be defined formally.

One can establish who signs, who approves, who reports to the Board, who leads the leadership team and who owns a given decision.

But in organisations there also exists another form of authority.

It is the one that does not necessarily appear on the organisational chart.

A former CEO who has led a company for many years can continue to possess an enormous capital of trust, relationships, information and legitimacy.

And this is where what I would call shadow authority can emerge.

A formal veto is not required.

It is enough that, when faced with a difficult decision, someone continues to ask themselves:

Before we proceed, what does the old CEO think?

At that point, the successor formally holds the role, but may not yet completely own the field.

Continuity and legacy authority are two different things

Protecting continuity means preventing organisational memory, knowledge of the system, critical relationships and the capacity to read signals that take years to understand from disappearing alongside a person.

Protecting continuity, however, does not necessarily require preserving the previous authority too.

This distinction becomes even more important in high-exposure transformations: successions, turnarounds, extraordinary transactions, strategic or technological changes.

The value of the predecessor can be enormous.

But the value of experience should not make the place where the decision now resides ambiguous.

For this reason, faced with a transition of this kind, I would look at at least five elements:

• Decision rights — which decisions now belong exclusively to the new CEO?
• Governance — does the predecessor remain on the Board, chair it, or leave it?
• Access — who can they interact with directly, and under what circumstances?
• Escalation — what happens when the predecessor and the successor read a situation differently?
• Time boundary — does the transition role have a duration, an exit condition, or does it risk becoming permanent?

But there is a sixth question that I consider even more important, and which rarely appears in corporate documents:

Who is perceived by the organisation as the person who, in the end, truly has the final say?

It is also a question of role identity

A succession changes an organisational chart in a single day.

Professional identity often requires much more time.

The predecessor must leave behind a part of the identity through which they have inhabited the organisation.

The successor must occupy a role that, until the day before, belonged to someone else.

The Board must stop implicitly turning to the old authority.

The leadership team must learn to recognise the new decision-making centre.

And the system must tolerate that the new CEO might see different things, decide differently and even change what the predecessor had built.

Here, succession becomes much more than a governance issue.

It becomes a problem of role, boundary and legitimacy.

It is possible to transfer a title without completely transferring authority.

And it is possible to retain experience without retaining command.

The quality of the transition lies precisely in distinguishing the two.

The paradox of the good advisor

An advisor can be very close to the decision-maker.

They can know the history, the people, the tensions and the alternatives.

They can see something that the decision-maker, immersed in the system, struggles to see.

They can challenge a reading or point out consequences that risk being overlooked.

But precisely because they are close, what does not belong to them must be absolutely clear.

A good advisor does not become a parallel authority.

They do not create a second decision-making centre.

They do not protect their own influence by turning it into dependence.

They bring seniority and judgment, leaving intact the ownership, accountability and decision rights of the role they are supporting.

In this sense, AIG and Adobe do not offer a definitive answer on which configuration is better.

They offer something more useful: two different architectures through which to observe the same problem.

And they bring us back to the question we started with.

When the CEO leaves, what must they truly leave behind?

Not necessarily their experience.

Not necessarily their availability.

Not necessarily their ability to contribute.

But it must become legible which authority no longer belongs to them.

Because a transition works when continuity protects the system without freezing it.

When experience remains available without turning into a veto.

When the predecessor can continue to be listened to without having to continue to be obeyed.

And when the new CEO can finally assume not only the title, but the full right to occupy the role.