Immunity to Change and the Need to Be Indispensable
Nour
Founder, Coachful
The need to be indispensable can make an ordinary leadership improvement feel unexpectedly difficult. A founder may want colleagues to answer their own questions, yet repeatedly speak before they can begin. Immunity to Change offers a way to explore whether that behavior protects another commitment, such as remaining visibly useful.
That is a possibility for the client to investigate, not a motive the coach should assign. Real authority, access, knowledge and staffing can also make a leader necessary. This guide explains the model and follows an original fictional experiment in which a founder changes one interaction while keeping legitimate decisions and support intact.
Understand the framework without diagnosing the person
Robert Kegan and Lisa Laskow Lahey developed Immunity to Change as an approach to examining why a desired improvement can conflict with another commitment. The Harvard Graduate School of Education overview attributes the work to them and describes attention to competing commitments and assumptions. Its immune-system language is a reflective metaphor, not a medical diagnosis.
In a Harvard interview, the authors distinguish technical changes in skills from adaptive work involving skills and mindset. That distinction is useful here: a leader may know how to delegate and still experience answering first as protection against becoming irrelevant.
It does not follow that every delegation difficulty is adaptive. A missing permission, unclear mandate or untrained colleague may need an operational change. The coach should examine both the work and the client's account rather than using a psychological explanation to dismiss a real dependency.
A client can want to help, value their expertise and still notice a costly intervention pattern. Avoid diagnosing a savior complex, inferring childhood causes or treating a wish to contribute as inherently unhealthy. The question is whether this particular way of contributing serves the client's chosen goal and the actual work.
Explain the four areas and how they connect
Kegan and Lahey's 2009 book, including the selected excerpts provided for HarvardX, organizes the inquiry around an improvement commitment, contrary behavior, competing commitments and big assumptions. The explanations and working examples here are original commentary, not a reproduction of the published form or its cases.
1. Choose an improvement the client actually wants
Make the desired behavior specific enough to recognize. “Become a better leader” does not show what would change in tomorrow's meeting. “Let the authorized operations lead answer routine status questions before I decide whether to add anything” identifies a usable interaction.
The client needs a reason to care beyond satisfying the coach. Ask what the change would make possible for them and for colleagues. Also identify what the goal does not change: approval limits, safety duties, client promises and access arrangements remain real constraints.
2. Describe the behavior that works against it
Record what the client does or fails to do, rather than a personality label. Answering a question directed to another person, rewriting a competent response before it is sent or insisting on being copied into every routine exchange are observable examples.
“I care too much” is harder to examine because it mixes an interpretation with a flattering explanation. Ask for the actual sequence. Who was asked, what did the client do, what opportunity did the colleague have and what happened next?
3. Explore what the behavior may protect
Invite the client to imagine not doing the familiar behavior and describe what worries them. The worry may suggest a competing commitment. A person might say, “I want to keep proving that I am needed,” alongside sincerely wanting a more independent team.
A competing commitment is more than another task on the calendar. It describes something the client may be trying to preserve. Do not insist on a hidden answer when none fits. Keep it tentative and allow the client to revise or reject the proposed account.
4. Make the assumption discussable
An assumption links the feared consequence to the familiar protection. For example: “If I do not answer routine questions first, colleagues will stop seeing any useful contribution from me.” The statement is broad enough to matter and specific enough to begin questioning.
Do not replace it immediately with a positive slogan. Ask what information would bear on it, where it seems valid and what it predicts in one suitable situation. A useful test gathers information about the assumption; it is not a public demonstration that the client can perform perfect delegation.
The model's connection matters. If the assumption makes invisibility seem inevitable, maintaining visible usefulness can feel necessary; answering first then protects that commitment while undermining the improvement. Describing the connection gives the client something to investigate, rather than another command to try harder.
Separate indispensable contribution from unnecessary intervention
Some decisions genuinely belong to the leader. A bank authorization may require an authorized signer. A priority conflict may exceed the operations lead's mandate. A colleague may need information that only the founder currently holds. Removing the founder from those situations without changing the arrangement can create risk or delay.
Other interventions are optional. Answering a routine status question before its owner can respond may deprive the colleague of an opportunity to do their actual job. The same leader can be legitimately needed for one question and unnecessarily first for another.
| Situation | First question | Appropriate direction |
|---|---|---|
| Only the founder has required access | What permission and continuity arrangement does the work need? | Resolve the real dependency through authorized processes. |
| A decision exceeds the colleague's mandate | Who has the actual authority? | Use the proper decision route; do not treat escalation as failure. |
| The colleague has scope and information but is interrupted | What happens if the leader allows an answer? | Explore the leader's chosen behavior and possible assumption. |
| The colleague already has too much work | What capacity or priority must change? | Agree workload rather than assigning more through a coaching exercise. |
The 2025 ICF Core Competencies emphasize client autonomy in designing goals, actions and accountability, including attention to support and barriers. That is a professional practice reference, not proof that this model will work for every client or a substitute for appropriate management.
Work through a complete fictional example
All people, dates, dialogue and outcomes in this case are fictional. Theo founded a service firm. Rina is its operations lead and is authorized to explain the current delivery plan and routine status. Theo retains decisions about accepting new work that would displace existing commitments.
In a weekly internal meeting, a colleague asks Rina for a status update. Theo answers before she begins, then later says the team relies on him too much. In coaching, he chooses to examine that sequence rather than claim that Rina should simply become more confident.
Build the client-owned account
| Working area | Theo's own entry |
|---|---|
| Desired improvement | Let Rina answer routine questions within her role, while I contribute when a decision actually belongs to me. |
| Current pattern | I answer first, add unrequested explanations and then complain that I carry the conversation. |
| Worry if I change | I might have nothing useful to say. People might wonder why I am there. |
| Possible competing commitment | Keep demonstrating that my presence is necessary in each discussion. |
| Tentative assumption | If I am not the first source of routine answers, colleagues will stop seeking my judgment and see no useful leadership contribution from me. |
The worry is included to explain the inquiry, not as a fifth formal column or a diagnosis. Theo says the competing commitment fits his experience as a possibility. He does not endorse a claim that every founder has the same concern.
Coach: What feels at risk if Rina answers the routine question well?
Theo: I may have nothing to add. I have treated having the answer as evidence that I still matter.
Coach: What could we learn in one ordinary interaction without deciding your worth as a person?
Theo: Whether colleagues still seek the decisions that actually belong to me when I stop taking the routine answers.
Coach: What does Rina need for that interaction?
Theo: Her actual role, the current plan and a choice about participating. I should ask rather than assume.
Obtain a real agreement with the colleague
Theo tells Rina he is working on his tendency to answer first. He proposes using their next ordinary internal meeting to let question owners respond, with the usual escalation route intact. He explains that the experiment concerns his conduct, not a secret performance test of her.
Rina agrees in principle but checks the preparation. She has delivery-plan version two, while Theo has recently changed priorities in version three. She cannot accurately explain the current position from the earlier document. She also asks for preparation time and a clear agenda.
They do not run the planned 12 May test. Theo records it as postponed for a real information gap, rather than evidence that Rina needs him to answer everything. He shares the authorized current plan, explains the changed priorities and confirms her existing mandate. No new approval rights are transferred by that conversation.
Rina accepts the rescheduled 19 May interaction after reviewing version three. Her manager protects thirty minutes of preparation and fifteen minutes of feedback within her paid allocation, moving an optional internal report revision to a later agreed slot. The ordinary meeting itself is already part of her role. The experiment does not create an unlimited new advice duty.
Complete an original experiment record
The authors' selected book excerpts describe small, safe tests designed to learn about an assumption. The record below is an original planning aid. It does not reproduce their guide sheet, acronym or published case.
| Field | Agreed 19 May experiment |
|---|---|
| Question | When Rina has the current information and answers routine questions, what evidence appears about Theo's prediction of losing a useful contribution? |
| Changed action | Theo lets the addressed person finish and adds only a relevant clarification or decision within his actual role. |
| Prediction | Theo expects colleagues to stop seeking his judgment and his presence to appear unnecessary. |
| Required support | Version three, named agenda topics, Rina's accepted role and the protected preparation/feedback time. |
| Limits | Routine internal discussion only. Required interventions, client commitments and priority decisions remain intact. No manufactured mistake or disappearance. |
| Evidence | Who answers, what is asked next, Theo's actual contribution and specific feedback about the interaction. |
| Review | 20 May: separate observations, interpretation, remaining support needs and a revised or retained assumption. |
They do not ask colleagues to reassure Theo that he is valuable. Feedback questions concern what helped the work and what remained unclear. A favorable response may inform the narrow prediction; it cannot measure a person's overall worth or guarantee future influence.
Show the actual meeting, including appropriate intervention
On 19 May, a colleague asks Rina which activities are complete. Theo notices the urge to answer but lets her explain from version three. She identifies the current position and a missing status entry, checks the underlying record and corrects the entry. Theo does not take over simply because a clarification is needed.
A second routine question concerns an absence already reflected in the plan. Rina explains the existing cover arrangement. Theo listens rather than repeating her answer in his own words. These are two ordinary questions within her role, not two situations in which the founder has abandoned required oversight.
The third question is different: can the firm accept a new request while preserving all existing dates? Rina explains that the current capacity does not support both and asks Theo to decide the priority. She has not been given authority to break an existing promise.
Theo considers the options and decides to retain the existing commitments and offer the new requester a later start. He asks the account owner to communicate that offer, rather than promising an immediate place. The requester subsequently accepts the later start, and the agreed delivery record is updated. The capacity decision belongs to Theo's role; making it does not invalidate the experiment.
After the meeting, Rina says: “I had room to answer the status questions. Your decision about the new request was needed. I still want the agenda to identify which version is current before each meeting.” Another colleague confirms that the distinction between reporting status and changing priorities made the route clearer.
Those are specific accounts of one interaction. They are not evidence that every colleague now understands the arrangement or that Rina needs no future support. The version-control request remains an operational action.
Review the assumption without forcing a positive story
At the 20 May review, Theo separates the following evidence:
- Rina answered two routine questions and corrected one status entry without Theo answering for her.
- The team still asked Theo for the priority decision within his mandate.
- He made that decision, and the later-start offer was actually accepted and recorded.
- Rina and a colleague reported useful aspects of the interaction, while Rina requested a clearer current-version agenda.
- Theo felt less visible during the routine answers and still found that uncomfortable.
The prediction that nobody would seek his judgment was not borne out in that meeting. The evidence does not establish that he is indispensable to every decision, that people would never question his role or that every intervention has become unnecessary.
Theo revises the assumption in his own words: “Colleagues can answer routine questions and still need my priority judgment. I may feel less useful when I am not answering first, but that feeling does not describe everything I contributed.” He treats this as a more workable interpretation of the observed situation, not a permanent cure.
He also changes a concrete process. Before the next weekly meeting, he and Rina confirm the current delivery-plan version in the agenda. On 26 May the agenda identifies version four, Rina confirms access and the meeting uses it. The next review asks whether that information arrangement works rather than turning every new problem into evidence of hidden resistance.
Theo chooses to repeat the ordinary interaction once more with the same boundaries. He does not escalate automatically to going offline, withholding expertise or testing a high-stakes client meeting. He can remain available and contribute while learning more about when answering first is unnecessary.
Contrast a genuine key-person dependency
In a separate fictional case, leader Nessa is the only authorized approver for a particular payment. The team asks for approval because the current arrangement requires it. A coach does not treat that request as proof that Nessa secretly wants dependency.
Nessa contacts the responsible finance owner and asks which continuity options are permitted. The owner confirms that a proposed backup requires a separate authorization process and appropriate checks. Until that is completed, Nessa retains the approval responsibility. A note saying “delegate more” does not authorize another person.
Her practical action is to pursue the actual continuity process, with a review date and the existing controls preserved. The backup remains proposed until formally established. This case needs an operational change before it can support a meaningful experiment about being less central.
A person can have both a real dependency and a personal concern about losing importance. The presence of one does not disprove the other. Work with the part for which there is evidence, and avoid forcing every issue into the same model.
Use Coachful for reflection, action and review
Coachful is our #1 recommendation for organizing this coaching workflow. A client can work through original prompts, agree a specific preparation action and review the difference between their prediction and the event. It is a workflow recommendation, not a guarantee of mindset change or a test of leadership value.
Use a short worksheet with the client's goal, conduct, worry, tentative commitment, assumption and chosen observation. Explain the authorized audience and leave room for “not sure” or a revised account. Ask for permitted summaries rather than confidential pricing, payment approvals or employee records.
The supported response view lets the coach review a submission. Coach review is private to the authorized coach team and does not itself deliver feedback to the client. Discuss the review in session or send the agreed response through the client's authorized private conversation. Private journal entries remain separate from a shared experiment record.
Record the actual role conversation, information check and test as distinct agreed actions in the ordinary client plan. Completing a task to ask Rina does not mean she agreed. A planned test is not a completed test, and a completed meeting does not prove the assumption has been permanently changed.
Review recap approval and session-type overrides before relying on review before sharing workplace material; automatic recap sending is the default. Do not share a client's tentative worry with a sponsor as though it were an established finding about their personality. See Coachful features for these preparation and follow-through tools.
Use this ordered practice sequence
- Choose one interaction. Establish the improvement the client actually wants and the work it would support.
- Check real dependencies. Review access, competence, workload and authority before interpreting intervention as self-protection.
- Develop the client's account. Separate conduct, worry, possible competing commitment and assumption. Keep each revisable.
- Design a bounded learning opportunity. Specify the prediction, changed action, relevant observations, limits and review.
- Obtain actual agreement. Give colleagues the information, time and choice the arrangement requires.
- Run and record honestly. Distinguish postponement, help, required intervention and the selected behavior.
- Review and choose again. Compare evidence with the prediction, preserve discomfort and operational needs, then revise the assumption or next action.
Troubleshoot without blaming the team
The colleague needs help during the test. Check the type of help and the actual mandate. An appropriate escalation does not establish that every routine answer belongs to the founder. Supply required support and record what the event can still teach.
The client changes behavior but feels worse. Hear the experience without forcing a success narrative. Review scope and support. One uncomfortable interaction need not settle the value of the approach or the person's broader professional contribution.
The colleague never agreed to more work. Stop treating the proposal as a valid arrangement. Discuss actual scope, information and capacity, and accept that the colleague may decline. Their refusal is not evidence of a defect the client should overcome.
The coach already knows the hidden motive. Return to the client's words and concrete examples. A tentative model is not a privileged reading of another person's mind. Use the model only if it helps the client's chosen inquiry.
Frequently asked questions
Is the need to be indispensable a diagnosis?
No. This guide explores a possible client-owned concern about a particular behavior. It does not diagnose a syndrome or establish the origin of the pattern.
Should the leader stop helping the team?
No. Separate useful expertise and legitimate decisions from unnecessary intervention. The worked experiment retains the founder's priority decision and the colleague's support.
What if the first experiment cannot happen?
Record why. Missing information, unavailable people or inappropriate scope may require a changed arrangement. A postponed test does not supply evidence about the assumption's predicted consequence.
Does another person succeeding make the leader unnecessary?
It can reveal a more useful division of work. Examine the actual contribution and responsibilities rather than treating a routine answer as a verdict on a person's value.
How is this different from fear of losing control?
This case concerns visibility and perceived usefulness even when the client accepts the colleague's competence. The control guide considers a different concern about oversight, while the perfectionism guide examines repeated work against an agreed standard.
