What Does an Organizational Development Consultant Actually Do?
“We need a leadership development program.” I hear versions of this request regularly. Sometimes it is management development. Sometimes leadership development. Sometimes it is coaching for a manager who has already been labeled as underperforming. The objective is usually clear: improve performance, and do it quickly. My response today is rarely to discuss the program. I start with a series of conversations with the executive who made the request. One of the first questions may be very simple: “Define a good manager for me.”
That question tends to lead somewhere else. What begins as a conversation about managers can expose an undefined business model, an absent or contested vision, outdated assumptions about management, incompatible expectations across functions and roles, blocked communication, or a growing disconnect between the organization and the environment in which it operates. The original request was not necessarily wrong. The managers may indeed need development. But what appeared to be an individual capability problem may be produced or reinforced by wider organizational conditions. That is where organizational development consulting begins. Not with the intervention, but with the explanation that made the intervention seem necessary.
An organizational development consultant may conduct interviews, analyze data, facilitate meetings, work with leadership teams, design interventions, and support implementation. None of those activities adequately explains the profession. Organizations can increasingly perform most of them themselves. Technology has made analysis cheaper, and generative AI can summarize open-text responses, compare frameworks, identify themes, and produce plausible recommendations in minutes. The more important question is what an external consultant makes possible that the organization cannot, or will not, do effectively from inside itself. My answer is this:
An OD consultant's product is not the intervention. It is a change in what the organization is able to notice about itself, say out loud, and act on.
The intervention is the instrument. What matters is what becomes possible afterward.
1. What the Organization Is Actually Buying When It Calls a Consultant
Organizations rarely call a consultant with an unframed problem. They call with a name for the problem: “Our managers need to become more strategic.” “People are resisting the change.” “We need to break down silos.” “We have an accountability problem.”
The name matters because it already limits the field of possible solutions. Call something an accountability problem and performance management becomes plausible. Call it a capability problem and training becomes plausible. Call it a structural problem and reorganization appears reasonable.
The first professional act is therefore to renegotiate the name. Not to prove the client wrong, but to test whether there is a better problem to solve. Thomas Wedell-Wedellsborg found something striking in surveys of 106 C-suite executives representing 91 organizations across 17 countries: 85% agreed that their organizations were poor at problem diagnosis, while 87% agreed that this weakness carried significant costs [1]. His more important point is that reframing is not about discovering the one “real” problem. It is about asking whether there is a more useful problem to solve (see: Are You Solving the Right Problems? - Harvard Business Review Harvard Business Review).
In my work, I use a simple formulation to begin that reframing:
We declare X. We do Y. Because we assume Z.
Z is where I start looking. Take the complaint: “Our managers do not take enough responsibility.” The organization declares that managers should take ownership. In practice, senior leaders continue to pull difficult decisions upward. Why? Perhaps the organization assumes that mistakes made lower in the hierarchy are more dangerous than delays created at the top. If that assumption is operating, another leadership program might improve managerial skills while leaving untouched the condition that prevents those skills from being exercised.
The request itself is therefore evidence. Organizations select interventions through the way they already understand themselves. An organization that sees expertise as the source of performance reaches readily for training. One that prizes efficiency reaches for restructuring. One whose identity is built around entrepreneurial autonomy may resist standardization even when coordination has become its central constraint. The solution being requested tells me something about how the organization explains itself.
There is an uncomfortable objection here. “Diagnose before you prescribe” also happens to serve the consultant's financial interest. Diagnosis can turn a small assignment into a larger one. That objection is valid. Yet, if the problem is already well defined, the evidence is strong, and the appropriate response is known, an additional diagnostic process can be wasteful. A specific knowledge gap may need training. A technical failure may need technical expertise. A management decision that everybody understands but nobody has made may simply need to be made. A consultant who cannot tell you when diagnosis is unnecessary is not defending rigor. They may be defending scope.
The same applies to the choice of discipline. If the desired change is already sufficiently defined and the central challenge is implementation and adoption, organizational development may not be the best starting point. I address that distinction separately in Organizational Development vs. Change Management: Which Does Your Organization Need?. The first value of an OD consultant, then, is not a solution. It is helping the organization notice that the problem it named may itself be part of the diagnosis.
2. What a Systemic Diagnosis Refuses to Examine
“We take a systemic view” is one of the least useful sentences in consulting. It sounds rigorous while committing the consultant to nothing. A serious systemic diagnosis is defined partly by what it refuses to examine. Every organization offers an almost unlimited field of possible inquiry: leadership, strategy, structure, culture, roles, incentives, processes, capabilities, relationships, workload, technology, decision rights, professional identities, and external pressures. Examining all of them is not systemic diagnosis. It is absence of method. The useful question is narrower:
Where does the system repeatedly produce, reinforce, or conceal the pattern we are trying to understand?
I look especially closely at interfaces and incongruities. A strategy says collaboration matters, while targets reward unit performance. A manager is accountable for an outcome but lacks authority over the resources required to produce it. The organization says initiative is valued, while consequential mistakes are escalated and remembered for years. Each component can look reasonable in isolation. The problem becomes visible in the relationship between them.
Diagnostic models help because they discipline attention. Weisbord, Nadler-Tushman, Burke-Litwin and other frameworks offer different ways of deciding where to look. They should guide inquiry, not predetermine the answer. I examine those choices in Organizational Diagnostic Models: A Comprehensive Review.
There is a risk built into selective inquiry: the lens can manufacture the finding. A consultant who specializes in culture will find cultural problems. A leadership specialist will discover leadership deficits. A consultant whose business depends on a proprietary intervention faces an even harder test: the answer may exist before the diagnosis begins. Every explanation therefore needs a way to fail. What evidence would make this interpretation wrong? Who sees the situation differently? What becomes visible if I use another lens?
In my own work, I also classify findings into three different destinations. Some are organizational problems that management can address. Others reveal a deeper dynamic that belongs in a separate professional track with someone qualified to hold it. And some are normal for the organization's stage of development and require patience rather than intervention. That last category matters because consultants are paid to intervene. Growth produces ambiguity. New leadership creates uncertainty. Redistribution of authority produces tension. Integration after restructuring creates temporary inefficiency. Discomfort is not evidence of dysfunction.
Sometimes the correct professional recommendation is to leave something alone. The same principle applies to professional boundaries. If an organizational diagnosis uncovers a clinical, legal, financial, technological, or other issue outside my competence, the answer is not to expand my scope until it includes the finding. The answer is to open another track with someone who has the relevant expertise.
A consultant's value is partly defined by the work they refuse to sell.
The Request-to-Diagnosis Test
Before approving the intervention that has been requested, I use a short sequence of questions to test whether the request and the organizational problem are actually the same thing. A leader can use the same test before calling a consultant.
Step | Question | What it tests |
1. Name the request | What are we asking for? | Separates the requested intervention from the problem it is supposed to solve. |
2. Define success behaviorally | What would “good” actually look like? | Converts labels such as “better leadership” or “more accountability” into something observable. |
3. Look outside the person | What currently makes that behavior difficult here? | Tests whether structures, incentives, authority, assumptions, interfaces, or routines reinforce the problem. |
4. Run the reproduction test | If the current people were replaced tomorrow, what conditions could reproduce the same problem a year from now? | Shifts attention from individual blame toward persistent organizational conditions. |
5. Set the boundary | What should we deliberately not intervene in? | Tests whether the diagnosis has enough discipline to distinguish an OD problem from something requiring another response - or no intervention at all. |
Consider the original request: “We need management development.” Step 2 changes the conversation immediately: “Define a good manager for me.” If the answer is “someone who makes decisions without constantly escalating them,” Step 3 is no longer primarily about management development. It becomes: What happens here when a manager actually makes an important decision without escalating it? That is a very different inquiry.
The test does not tell you which diagnostic model to use, nor does it prove that you need OD consulting. Its purpose is more basic: to prevent the requested solution from silently becoming the diagnosis.
Systemic diagnosis is not the ability to notice everything. It is the discipline to notice what reproduces the pattern - and leave the rest alone.
3. How a Consultant Finds Out What Is Already Known but Unsaid
Organizations measure more than they used to. That changes the traditional justification for external consulting. The argument cannot simply be that leaders do not know what is happening inside their organizations. They have engagement surveys, operational dashboards, HR analytics, customer data, pulse surveys, performance metrics, open-text responses, and increasingly sophisticated analytical tools. Yet measurement does not guarantee understanding.
Perceptyx's 2025 State of Employee Listening data show that 75% of organizations were listening to employees at least quarterly, compared with 18% surveying more than once per year a decade earlier [2]. Its 2026 research identifies the next problem: 71% of employees report that survey results are shared, but only 51% report seeing actual improvements from their feedback [3]. These are findings from successive reports by the same research program, not two independent studies. (see: The State of Employee Listening 2025 The State of Employee Listening 2026).
The bottleneck is moving. Collection is easier. Analysis is cheaper. The difficult part is what happens when the information becomes consequential. Morrison and Milliken's work on organizational silence helps explain why. They describe organizational conditions that can make withholding information about potential problems appear sensible to employees [4]. Silence, in other words, can be produced by the system rather than by an absence of information. (see: Organizational Silence: A Barrier to Change and Development in a Pluralistic World).
An external consultant changes some of those conditions. That does not mean employees suddenly reveal an objective truth. People use consultants politically too. An interview can become a channel for sending a message upward. What someone tells a consultant may be accurate, strategic, incomplete, defensive, or several of these at once. That is why I do not treat an interview as a measurement device. The interview changes the conditions under which something can be said.
In one focus-group design I use, participants first respond anonymously to questions in real time. The aggregate responses appear immediately on a screen. Only then does the discussion begin. The shift is subtle but important. Nobody has to begin by declaring, “I don't trust management.” They can say, “Look at what is on the screen.” The data become an object that the group can examine together. Anonymity makes some information easier to express. Collective visibility makes it possible to discuss. I often close with a different question: “What was the most important thing we said here, and what did we go silent about?” The silence is not a finding. It is a lead. Nor do I assume that employees understand the organization better than leadership. That would be another convenient consulting story. Senior leaders possess information about external constraints, resources, trade-offs, and strategic choices that other organizational members do not have.
Contradiction is more useful than deciding prematurely who is right. If a senior executive and a frontline employee describe the same decision process in incompatible ways, their disagreement may itself tell us something important about the organization. Before an interpretation becomes a formal finding in my work, I look for it across independent sources. My working rule is three sufficiently independent sources. Until then, it remains a hypothesis.
This is also where the impact of AI becomes clearer. AI can cluster comments, summarize interviews, compare frameworks, and generate credible interpretations. The analytical component of consulting is becoming less scarce. But analysis and access are different things. The harder problem is who can be told something consequential, under what conditions, and whether that information can travel toward the place where it becomes expensive to hear.
The consultant has not necessarily discovered information that nobody possessed. The work has made fragmented, contested, or costly information possible to say - and examine - in the same room.
4. What Happens When the Organization Sees the Diagnosis
The most consequential meeting in an OD engagement can be the one in which the findings come back. It is commonly treated as the end of diagnosis. I regard it as the first intervention. Until that moment, different parts of the organization may each hold fragments of a pattern. Returning the findings places those fragments into a shared account and asks the organization to look at itself. That changes the system before any formal intervention begins.
The next temptation is to search for “the root cause.” Root-cause language promises that organizational complexity can be reduced to one thing beneath everything else. That is a poor assumption in complex systems. Sidney Dekker's work on complex-system failure provides a useful alternative: outcomes can emerge through interactions among individually understandable decisions and conditions rather than from one defective component [5]. I therefore use a different question: What conditions, left unchanged, would reproduce this pattern a year from now even if every current person were replaced?
Suppose decisions are consistently slow. Replacing a cautious manager may accelerate them temporarily. But if decision rights remain ambiguous, information still travels through four layers, and the personal cost of an incorrect decision remains much greater than the organizational cost of delay, the pattern is likely to return.
Those conditions are more interesting than the person currently carrying the symptom. The diagnosis should therefore change the intervention. But even then, organizational problems rarely offer one value that should simply defeat another. Organizations need autonomy and consistency, speed and due process, innovation and reliability, local responsiveness and system-wide coordination. A technically correct intervention can fail by optimizing one side. Centralization can reduce inconsistency while destroying initiative. Radical decentralization can improve speed while creating fragmentation.
“Hold both” is not an answer. It can become elegant language for avoiding a decision. The intervention needs a mechanism. That might mean differentiated decision rights, explicit escalation thresholds, changed interfaces, boundary roles, or review routines that make both legitimate requirements more possible in practice. The point is not compromise for its own sake. It is organizational design capable of carrying a tension that cannot simply be removed. And the organization may reject the proposed intervention. That rejection deserves examination before it is labeled resistance. It can reveal who would bear the cost of the change, which legitimate value the intervention ignored, or where the diagnosis itself was wrong. Sometimes resistance is correct. Diagnosis becomes organizational development only when what the organization can now see changes what it is prepared to do.
5. Who Is the Client When the Finding Becomes Uncomfortable?
The sponsor pays the consultant. That does not settle who the work is for. Employees provide evidence and live with the consequences of change. That does not make them the client either. Organizations also carry obligations toward people who may never enter the consulting room: customers, patients, citizens, shareholders, partners, regulated communities, future employees, or others who depend on what the organization does. That creates the hardest structural problem in consulting.
The sponsor may be part of the finding. In my first conversation with a sponsor, I often ask: “What matters to you that happens here, and what matters to you that does not happen here?” The second half matters more. It begins to reveal the boundary around what the engagement is allowed to discover.
A sponsor can genuinely want an independent diagnosis and simultaneously hope that one particular explanation does not emerge. That is not unusual. It is diagnostically important. This is why the gravest ethical issue in consulting is not only confidentiality. It is whether the consultant has been hired to produce a finding or to produce cover.
Consultants can provide expertise. They can also legitimize a decision that has effectively already been made. Mazzucato and Collington's critique of the consulting industry is useful here because it forces attention onto the role the consultant is actually performing [6]. There is nothing inherently illegitimate about implementation work. If leadership has already decided to restructure and wants external help carrying it out, say so. Yet, do not call it independent diagnosis. Independence is therefore not a personal virtue. It is a structural role that has to be protected through contracting, access, confidentiality, supervision, and the consultant's willingness to return findings that make the relationship uncomfortable.
I work under professional supervision. Buyers should ask consultants who supervises their work. Supervision does not guarantee good consulting. It creates somewhere outside the client relationship where the consultant's own behavior can be examined: Why did I avoid that subject? Why did I intervene there? Whose explanation have I started adopting? What role might the organization be recruiting me to play?
The same discipline applies when values enter the diagnosis. In research with colleagues, I co-developed and validated the Publicness Perceptions Scale (PPS) to examine how people experience dimensions of organizational publicness across different types of organizations [7]. The two-study research identified five dimensions: perceived ownership, funding and political control; market control; legality and equity; accountability; and transparency. Importantly, the research was designed to examine perceptions of publicness across organizational types rather than equating publicness simply with public-sector ownership. (see: Unveiling the Spirit of Publicness: Conceptualization and Validation of a Publicness Perceptions Scale).
For consulting, the implication is practical. A broad statement such as “we are not living our commitment to accountability or transparency” need not remain a vague cultural complaint. Perceptions can be examined across organizational groups and contexts to locate where a declared commitment and experienced organizational practice diverge. Yet, that does not authorize the consultant to decide what the organization should value. No consultant appointed themselves to represent customers, citizens, employees, or “the public.” The professional role is visibility, not representation. If an organization has declared that fairness, transparency, safety, service, or another obligation matters, the consultant can examine whether organizational practice is experienced as consistent with that commitment.
The value must belong to the organization. The consultant makes the gap visible. I learned the importance of this boundary from both sides of consulting. Before working independently, I held executive responsibility for systemic organizational development in a system of roughly 6,000 people. At that scale, under sustained operational pressure, the difference between an intellectually attractive consulting idea and something an organization can actually absorb becomes impossible to ignore. The consultant can illuminate, challenge, and help design. Management still has to manage. Independence matters because noticing and saying are useless if the finding becomes unsayable precisely where action has to be authorized.
6. What Should Still Be Running a Year After the Consultant Leaves?
“Building capability” is one of the most comfortable promises in organizational development. It is also too vague to be useful unless you can point to where that capability now lives.
I look for routines. Feldman and Pentland distinguish between the ostensive aspect of an organizational routine - the abstract idea or structure of the routine - and its performative aspect: what specific people actually do at specific times and places [8]. That distinction explains why putting a new process on paper proves very little. (see:Reconceptualizing Organizational Routines as a Source of Flexibility and Change).
A weekly learning meeting on the calendar is not evidence of learning. A decision protocol is not evidence that decisions changed. A leadership framework in a presentation is not evidence of different leadership. A capability that never enters organizational routines may be little more than the memory of a workshop. Implementation therefore belongs in the places where work already happens: the leadership meeting where priorities are negotiated, the decision process where authority becomes visible, the handoff where two units repeatedly lose coordination. A workshop can create protected space for work that cannot initially happen inside those routines. It should not become a substitute for changing the routine that reproduces the problem every week.
This also creates a danger. Embedded implementation is an accurate description both of useful OD work and of a consultant who never leaves. A consultant who becomes indispensable to recurring decisions has not necessarily built organizational capability. They may have become shadow management without managerial accountability.
The role should decline. At first, I may lead a process the organization cannot yet hold. Then I design it with the people who will own it. Later, my role becomes observation and reflection. Eventually, I should not be in the room.
My six-year longitudinal research reinforced this point. The study followed an OD process through 67 interviews, field observations, and organizational documentation collected across two phases, and examined how change became connected to ongoing organizational functioning rather than remaining a separate developmental activity [9]. It is one longitudinal case, not a universal recipe. What it made especially visible was that sustainable change required developmental work to connect back into the system in which everyday work was being performed. (see: OD Consulting as a Catalyst for Change - The Journal of Applied Behavioral Science).
Three signs matter to me when thinking about exit. A question that once required the consultant is now asked without them. People describe a finding as their understanding of the organization rather than “what the consultant told us.” And people who were never part of the original intervention enter the organization and participate in the changed practice because it has become part of how work is done. Even that is not enough. Routines can become hollow. People can perform the meeting without performing the inquiry. A routine introduced by a consultant can eventually ossify into the next organizational constraint. The durable residue of consulting therefore cannot be a fixed routine alone. It has to include a routine of noticing - a recurring way of testing whether the assumptions, conditions, and relationships on which current practice rests still hold. This is why the strongest signal that an organization needs OD is not a symptom. It is repetition.
A leadership team discusses slow decisions every quarter. Each time, it agrees that managers need more authority. Each time, the next high-risk decision is pulled back upward for senior approval. Six months later, “empowerment” is back on the agenda. The repetition is the diagnostic signal. The organization does not lack a decision. It lacks an explanation for why its own decisions do not survive contact with its routines.
Even then, OD may not be the answer. A defined skill gap may need training. A problem concentrated in one executive's behavior may call for coaching. A defined change that needs adoption may call for change management. The HR team may already have the capability and standing to do the work. A competent internal manager may be enough. And sometimes everyone knows exactly what needs to happen. Management has simply deferred the decision.
If you are considering an external OD consultant, I would ask him/her six questions:
What would you need to see before recommending an intervention?
What would make you conclude that we do not need OD consulting?
What happens if your diagnosis points back at me as the sponsor?
Who supervises your work?
What should still be happening here after you leave?
Tell me about a finding that made an engagement more difficult rather than easier.
There is another evidence test worth adding: ask where the consultant's numbers come from. The familiar claim that 70% of organizational change initiatives fail is a useful example. Mark Hughes traced five published instances of that claim and found no valid and reliable empirical evidence supporting the asserted failure rate [10]. (see: Do 70 Per Cent of All Organizational Change Initiatives Really Fail?). Evidence discipline is part of consulting discipline.
AI now adds one more buyer question: What part of your work can a model do, and what exactly am I paying you for beyond that? A credible answer cannot simply be “insight.” Analysis is becoming commoditized. The harder value lies elsewhere: access to information that does not travel easily through hierarchy, independence from the explanation that commissioned the work, responsibility for carrying uncomfortable findings back into the system, and the ability to help the organization act on them without becoming dependent on the consultant.
That brings us back to the request in the opening. “We need leadership development.” The organization may indeed need leadership development. But before discussing a program, I return to the question that changes the conversation: “Define a good manager for me.” The answer tells me where to start looking. What follows determines whether the organization needs a leadership program, a different intervention, or no OD engagement at all. The original request may not have been wrong. It may simply have been incomplete.
Good OD consulting should leave the organization better able to do that work itself - to notice what it was missing, say what was difficult to hold in the room, and act on what it now understands.
A good first conversation should work the same way. Its purpose is to determine whether organizational development consulting is actually needed, not to assume that it is.
At PublicWise, one legitimate outcome of that conversation is that I tell you that you do not need me.
References
[1] Wedell-Wedellsborg, T. (2017). Are you solving the right problems? Harvard Business Review, 95(1), 76-83.Harvard Business Review article
[2] Perceptyx. (2025). The State of Employee Listening in 2025.State of Employee Listening 2025 report
[3] Perceptyx. (2026). The State of Employee Listening 2026.State of Employee Listening 2026 report
[4] Morrison, E. W., & Milliken, F. J. (2000). Organizational silence: A barrier to change and development in a pluralistic world. Academy of Management Review, 25(4), 706-725.Academy of Management Review article
[5] Dekker, S. (2011). Drift into Failure: From Hunting Broken Components to Understanding Complex Systems. Farnham, UK: Ashgate.
[6] Mazzucato, M., & Collington, R. (2023). The Big Con: How the Consulting Industry Weakens Our Businesses, Infantilizes Our Governments and Warps Our Economies. New York: Penguin Press.
[7] Dryzin-Amit, Y., Vashdi, D. R., & Vigoda-Gadot, E. (2024). Unveiling the spirit of publicness: Conceptualization and validation of a publicness perceptions scale. Journal of General Management, OnlineFirst. DOI: 10.1177/03063070241258806.
[8] Feldman, M. S., & Pentland, B. T. (2003). Reconceptualizing organizational routines as a source of flexibility and change. Administrative Science Quarterly, 48(1), 94-118. DOI: 10.2307/3556620.Administrative Science Quarterly article
[9] Dryzin-Amit, Y. (2026). OD consulting as a catalyst for change. The Journal of Applied Behavioral Science, 62(3), 486-517. DOI: 10.1177/00218863251376151. First published online September 22, 2025; issue publication September 2026. Sage JournalsOpen-access SAGE article
[10] Hughes, M. (2011). Do 70 per cent of all organizational change initiatives really fail? Journal of Change Management, 11(4), 451-464. DOI: 10.1080/14697017.2011.630506.Journal of Change Management article
Comments