Thought leadership is not a writing problem

The easiest way to produce thought leadership is to remove the thinking.

Choose a subject everyone is discussing. Announce that it is transforming the world. Add three recommendations that no reasonable person could oppose. Finish by reminding readers that the future belongs to organisations willing to adapt.

An article has appeared. Nothing has happened.

Generative AI makes this exercise remarkably convenient. But the formula predates the technology. Businesses have long published articles that mistake a topic for an argument and a senior job title for a reason to listen.

The harder work begins before anyone writes the opening sentence. What does this person know that is not already obvious? What supports their position? Where does their experience complicate the prevailing story? What could a reader understand differently afterwards?

Those questions shaped my editorial work at the Indian School of Business. They also mattered, in different forms, in communications at ELMS and internally at Microsoft (via partner).

These were not interchangeable assignments. Academic research, physical literacy and the digital workplace require different kinds of explanation. Their audiences have different concerns. Their evidence comes in different forms.

But they exposed the same underlying problem: organisations possess considerably more expertise than they know how to communicate.

The answer is not necessarily more content. It is better judgement about what deserves to become content in the first place.

Expertise does not arrive as an article

Expertise usually arrives untidily.

It comes in a research paper written for other researchers. A presentation whose most interesting point appears somewhere near the appendix. A specialist’s explanation that assumes everyone understands the terminology. A leadership priority expressed so broadly that nobody could object to it—or act on it.

This does not mean the expert has failed. The researcher is doing research. The programme specialist is developing a programme. The leader is making decisions. Writing for a particular audience is another task, with its own demands.

The mistake is assuming that task consists mainly of polishing sentences.

A piece can be grammatically immaculate and intellectually empty. It can sound authoritative while leaving the reader unable to explain its argument. It can contain a great deal of information without making any of it useful.

The editorial problem is therefore not simply: how do we make this easier to read?

It is: what is the idea, what can it legitimately claim, and why should this reader spend time with it?

That is where thought leadership becomes a service rather than a label.

At ISB: finding the business question inside the research

At ISBInsight, I led editorial strategy connecting management research with business leaders, alumni, policymakers and wider public audiences. I worked with faculty, deans and institutional leadership to interpret research and shape executive-facing narratives. My ISB Insight story describes that work in more detail.

The central challenge was preserving rigour while making relevance visible.

Those objectives are sometimes presented as opposites. Academic writing is assumed to be serious but inaccessible; business writing accessible but inclined to remove the qualifications. The editor is then invited to find a tolerable point between them.

I do not think that is the right bargain.

Accessibility should not require making a finding more certain than it is. Rigour should not require hiding its significance behind the conventions of a research paper.

The work was to identify the research question, understand the evidence and establish which implications mattered to a non-academic audience. Then came the narrative: the angle, structure, context and language through which the reader could encounter the argument.

A paper’s sequence is not automatically an article’s sequence. The order that establishes a contribution within a scholarly field may not be the order that helps an executive understand a business problem.

Changing the sequence is not the same as changing the substance.

The difficult judgement is deciding what must remain, what needs explanation and what belongs outside the article. An omission can improve clarity. It can also quietly distort the finding.

The editor needs to know the difference.

The qualification may be the most useful part

Business writing likes a clean conclusion. Research often supplies conditions.

A finding may apply to a particular population, market or period. An association may not establish causation. An intervention that works in one setting may depend on conditions absent in another.

These details can look inconvenient when an article is supposed to deliver a confident recommendation.

But they may be the information a reader most needs.

Consider the difference between “this approach improves performance” and “this approach improved performance when these conditions were present”. The second is less portable as a slogan. It is more useful as the basis for a decision.

At ISB, my work involved identifying the central argument and connecting findings to live business and policy conversations without oversimplifying the underlying research. Accuracy, institutional voice and audience relevance had to work together.

That experience shaped a principle I still use: clarity is not the removal of complexity. It is the organisation of complexity around a question the reader can follow.

This requires more than replacing technical terms with shorter words. Sometimes the term needs to stay and the explanation needs to improve. Sometimes a familiar word conceals a distinction that must be made explicit.

An article earns its confidence through those decisions.

The editor should contribute more than agreement

There is a deferential model of thought-leadership production in which the expert speaks, the writer transcribes and everyone hopes the result sounds important.

It produces recognisable problems.

The argument contains an unexplained leap because nobody questioned it. The recommendation assumes resources the audience does not have. The example illustrates success but reveals nothing about the decisions behind it.

A useful editorial partner has to do more than record expertise. They have to interrogate it.

That need not be adversarial. Often the best question is straightforward:

• What do you mean by that?
• What happened that made you change your mind?
• Where would this advice fail?
• What would someone have to stop doing to follow it?
• What does the evidence allow us to say—and what does it not?

These questions help move an article beyond a polished restatement of familiar advice.

They also protect the expert. A weak argument does not become safer because nobody challenges it before publication.

The writer’s contribution is not to manufacture a provocative opinion. It is to help the person distinguish a defensible position from an attractive sentence.

At ELMS: the audience was not one audience

At ELMS, the translation problem took a different form.

My communications and stakeholder-engagement work involved a national ecosystem that included educators, coaches, parents, students, government agencies and corporate partners. The subject matter included physical literacy and grassroots development.

The concept could remain consistent. The explanation could not assume the same starting point for everyone.

An educator might need to understand the relationship to teaching practice. A parent might need to understand what the concept meant for a child’s experience. A partner might need to understand the programme’s purpose and the role they could play.

These are examples of the questions such an audience mix raises, not quotations from a particular project. The important distinction is that changing the audience changes the work.

An organisation often begins with what it wants to say. The reader begins elsewhere: what is this, does it matter to me, and what am I expected to do?

Communication has to connect those positions.

It is tempting to call every audience a stakeholder and leave the analysis there. The word is administratively convenient. It tells you almost nothing about what a person needs to understand.

The useful work begins when the category becomes specific.

Translation is not the same as promotion

This matters because promotional writing and explanatory writing have different instincts.

Promotion wants the concept to appear compelling. Explanation wants the reader to understand it well enough to make a judgement.

The two can coexist. They are not identical.

A mission-led organisation may feel strongly about its work. That conviction does not automatically give an unfamiliar reader the context to appreciate it. Describing something as transformative, inclusive or essential cannot substitute for explaining what changes, for whom and through what mechanism.

At ELMS, translating specialist concepts for non-specialist audiences required attention to that gap. During the transition to virtual engagement in the pandemic, the communication task also involved helping people stay connected as the form of participation changed.

I would not relabel all of that work as thought leadership. Programme communication, community engagement and educational explanation have purposes of their own.

But they develop a capability that thought leadership depends on: recognising that relevance is not a quality the organisation can simply declare.

The reader has to be able to find it.

Inside Microsoft: the question was “what does this mean for my work?”

Internal communication offers a useful correction to the glamour around thought leadership.

Outside an organisation, a leader may be rewarded for sounding visionary. Inside it, people still need to know what happens next.

My internal communications work at Microsoft (via partner) included leadership communication, newsletters, digital-workplace resources, onboarding materials and adoption-related content. The setting differed from public editorial work, but it sharpened the same question: what does this information mean to the person receiving it?

A statement can be strategically correct and practically incomplete.

“We are changing how we work” may express the direction. It does not explain which work changes, where support is available, what remains uncertain or how an employee should respond.

A message that answers none of those questions is not necessarily false. It has simply transferred the explanatory burden to the reader.

That burden then appears elsewhere: another meeting, an unanswered question, a local interpretation, a manager trying to connect the announcement to actual work.

This is why I am wary of treating internal communication as a distribution problem. Sending the message is only one part of the task. The more demanding part is establishing what people need to understand.

Not every leadership message is thought leadership

This distinction deserves to be kept intact.

A policy update is not thought leadership because a leader signs it. An onboarding guide is not an opinion essay. A newsletter does not become influential through a more ambitious title.

These formats can be valuable without borrowing a grander label.

Internal thought leadership, where the term is useful, should offer a considered explanation of a problem or direction: why the organisation sees an issue this way, which choices follow and what the reasoning leaves open.

It should help people think, not merely tell them what has been decided.

That does not mean every message needs a debate attached. Some communications should be short, direct and procedural. The skill is knowing which kind of communication the situation requires.

My experience across these settings made me less interested in the prestige of the format and more interested in its job.

A strong research article, a clear explanation of a specialist concept and a useful internal leadership message do not need to be the same thing. They do need to respect the person reading them.

AI makes the distinction harder to avoid

Generative AI can assist with structures, summaries and draft language. That makes the question of what a person contributes more pressing.

If the entire service consists of turning a broad topic into plausible paragraphs, it is vulnerable for a good reason: plausible paragraphs are now easy to obtain.

The better service begins with material that is not interchangeable—experience, evidence, observations, decisions, disagreements—and helps develop it into an argument.

There is research that makes this tension worth considering carefully.

In a 2024 experiment, Anil Doshi and Oliver Hauser studied the effect of AI-generated ideas on short-story writing. Access to those ideas improved evaluations of individual stories, particularly for less creative writers, while the AI-assisted stories became more similar to one another. This was a study of short fiction, not executive articles; it does not prove that all AI-assisted thought leadership becomes homogeneous. It does show how individual improvements and collective similarity can occur together.

That is a more useful finding than either “AI destroys creativity” or “AI makes everyone creative”.

The practical question is where assistance becomes anchoring. If a suggested structure arrives before the writer has established their own argument, it can quietly become the argument.

The draft may sound better while becoming less particular.

Use AI after you have something worth developing

My rule would be simple: do not ask the model to supply the expertise you are claiming to communicate.

Begin with the source material.

That might mean research, an interview, project notes, documented decisions or a detailed account of what an expert has observed. Identify the central question and the position before allowing a fluent draft to make the choices for you.

AI can then help test structures, identify repetition or generate questions for further investigation. Its output still needs checking against the sources and the person’s intended meaning.

I explore the operational side in “Designing an AI-assisted content system that scales with the practice”: the separation of expert contribution, drafting, review and approval. The relevant principle is that assistance should sit inside an accountable editorial process, rather than replace one.

The test is not whether AI touched the article.

It is whether anyone can explain where its claims came from, why the argument takes this shape and who stands behind it.

What I would ask before commissioning another article

The commercial case for thought leadership is not an excuse to be vague about quality.

Edelman and LinkedIn’s 2025 report draws on nearly 2,000 global professionals, including visible decision-makers and less visible participants in buying decisions. It finds that these “hidden buyers” engage with thought leadership and value material that informs or challenges their perspectives. Strong content can also help less familiar providers gain consideration.

That is an argument for useful thinking, not an instruction to publish more frequently.

Before commissioning the next article, I would ask five questions.

1. What does the reader understand afterwards?

“Raise awareness” is not specific enough. Name the distinction, misconception or decision the article will clarify.

2. What makes this person qualified to say it?

A job title supplies context, not evidence. Look for experience, research or an observation the person can substantiate.

3. What is the argument?

“AI and the future of work” is a subject. “Our approval process is preventing us from realising the benefit of faster drafting” is a position that can be examined.

4. What would complicate the argument?

Look for boundaries, counterexamples and competing explanations. If nothing could change the conclusion, you may be producing advocacy rather than analysis.

5. Why this format?

A long article is not always the answer. The material might work better as a short note, a conversation, a practical guide or a response to a question clients keep asking.

The format should follow the job. Not the other way round.

What these experiences changed about my practice

At ISB, the task was to connect research with the questions of a wider readership without sacrificing its intellectual substance. At ELMS, it was to recognise that different audiences needed different routes into specialist knowledge. Internally at Microsoft, it was to connect organisational messages with the realities of people’s work.

I do not claim those settings were equivalent. Their differences are what made the experience useful.

Together, they taught me that the most consequential editorial decisions often happen before the prose becomes polished: selecting the question, identifying the evidence, recognising the audience’s starting point and deciding what the material can honestly support.

That is the work I bring to thought-leadership consulting.

Sometimes it produces an article. Sometimes it reveals that the proposed article has no argument yet. Sometimes it uncovers a much better question than the one in the brief.

That is not a delay in production. It is the reason to involve an editorial partner.

The internet does not need another confident explanation that change is happening. Readers need help understanding which change matters, what the evidence shows and what choices follow.

The writing should make that work visible.

Research notes

1. Anil R. Doshi and Oliver P. Hauser, Generative AI enhances individual creativity but reduces the collective diversity of novel content, Science Advances, July 2024. The experiment concerns short-story writing; its findings should not be treated as a direct measurement of business thought leadership.
2. Edelman and LinkedIn, 2025 B2B Thought Leadership Impact Report: Invisible Influence—Unlocking the Power of Hidden Buyers, June 2025. The report examines how thought leadership reaches and influences different participants in B2B buying groups.