---

AI didn’t invent bland thought leadership. We’ve been watering down expert opinions in the name of consensus for years. Now that producing another polished draft takes minutes, the real challenge is making sure there’s still something worth saying.

---

Professional services firms are paid to have a point of view. Clients hire consulting, accounting, and advisory firms because they want people who can look at a complicated situation, make sense of it, and tell them what they think should happen next.

Put that same expertise through a content review, though, and things can get weird. I’ve watched teams get hung up on a single word because it might sound too judgmental. I’ve seen people ask AI whether a sentence sounds “nice enough” because they’re worried that being too direct will put someone off. I’ve watched a strong take get watered down because the brand is supposed to stay neutral.

None of those decisions feel dramatic in the moment, but together they can kill the story before it ever gets published. A strong recommendation becomes a consideration. The sharpest sentence is the one everyone spends the most time trying to soften. A clear point of view turns into something the whole category could comfortably agree with.

Then we publish it and wonder why nobody cares.

AI gets blamed for a lot of generic content now. I think we’re giving it too much credit. Professional services firms were publishing interchangeable thought leadership long before ChatGPT showed up. There were smart people involved. There was real expertise behind the work. The final piece could still sound like something any competitor might have published.

If anything, AI has removed the last excuse not to notice how much of that sameness was already coming from us. Producing another perfectly competent version is now almost effortless.

The trouble starts when generating another option becomes easier than deciding what you actually believe.

The draft arrives before the decision

There’s good evidence that AI can make writing faster. In a 2023 MIT study of professional writing tasks, participants with access to ChatGPT finished faster and received higher quality ratings. The assignments were short and didn’t require the company or client context of a full thought leadership piece, but the productivity benefit helps explain the appeal.

And I’m not arguing against those benefits. I’d happily take the time back. A six-week production schedule has never guaranteed brilliant thought leadership. Slow content can be every bit as bland as fast content.

But some of the friction in that old process forced useful conversations. When a substantial piece required interviews, expert time, writing time, and a meaningful chunk of budget, somebody had to make a case for doing it. Why this topic? What does the firm actually have to contribute? What would we tell a client who asked us about this tomorrow?

AI changes the order of operations. Now you can have a convincing-looking draft before the team has properly answered those questions.

Once that draft exists, work starts gathering around it. Facts need checking. The SME wants another example. Someone has a note on the introduction. Legal needs to see it. Design is expecting copy on Thursday. By then, questioning the premise feels bigger than rewriting the copy. And rewriting the copy is suddenly very cheap.

When another version takes minutes, indecision barely looks like a cost. Version three becomes version four. Someone asks to see the headline both ways. Another reviewer wants the original paragraph with the new ending. What looks like iteration can actually be a way of avoiding the editorial decision. Eventually, everyone is tweaking the copy while the real question remains untouched: What are we trying to say? 

Some of the time AI saves will absolutely go back to teams. Some of it can just as easily disappear into more versions.

How useful advice becomes “a balanced approach”

Imagine an advisory firm is publishing a point of view for operations leaders considering a major technology investment. During an expert interview, the SME gives the content team a clear recommendation:

“If nobody owns the broken process today, buying software probably won’t fix it. Sort out the ownership problem first, or you may end up moving the same confusion into a more expensive system.”

A reader can do something with that. They can look at their organization and ask who owns the process, whether the problem really is the technology, and whether a purchase should wait.

Then the review comments arrive.

“Can we soften this? We don’t want people to feel stupid for buying already.”

“We should be careful not to alienate companies that have invested.”

“We have some of these issues internally too.”

And one I’ve heard in different forms plenty of times: “Should the brand really be taking a side here? We want to stay neutral to the industry.”

These are reasonable concerns. They could also make the advice better. Why do capable companies buy before ownership is clear? Maybe everyone is fed up with the current system, and a purchase feels like progress. Maybe the technology genuinely can solve part of the problem while the organization works through the rest.

Now there’s something interesting to explore. The expert can explain which problems should delay implementation and which can reasonably be worked out along the way. The advice gets better because someone challenged it.

There’s also the easier route:

“Organizations should take a balanced approach to people, processes, and technology.”

Everyone can live with that sentence. Nobody had to make a difficult editorial decision. And the reader has learned almost nothing.

That’s what happens when getting everyone comfortable becomes the goal. The team started with advice specific enough to influence a decision and ended with language broad enough that nobody has to defend it.

This is where AI brainfry sets in. When you’re staring at a dozen subtly different, perfectly polished ways to water down a hard truth, your critical faculties just give up. You aren’t editing anymore; you’re escaping the fatigue of the feedback loop. You pick the blandest version everyone can agree on, just to move on.

If judgment is what you’re selling, why edit the judgment out?

One question I’d keep coming back to in review is: Would we actually say this to a client in the room?

Someone has to stand behind the advice

Thought leadership carries particular weight in professional services because prospective clients are looking for evidence of how the firm thinks. In Edelman and LinkedIn’s 2024 research, 73 percent of surveyed B2B decision-makers said thought leadership was a more trustworthy way to assess an organization’s capabilities than its marketing materials and product sheets.

A capabilities page can tell me that a firm is experienced, strategic, and innovative. A good piece of thought leadership lets me see what its people notice, what they prioritize, and how they reach a recommendation. I can start deciding whether I trust their judgment before I ever get into a meeting with them.

That makes the person with final editorial authority pretty important. “Make it bolder” doesn’t help much when the writer is also expected to accommodate every comment in the document. If three senior reviewers disagree about the recommendation, another round of copy won’t resolve it. Somebody needs to decide what the firm is prepared to say.

That doesn’t mean protecting every strong opinion from scrutiny. Good review should make the thinking harder to poke holes in. Challenge the evidence. Push on the reasoning. Fix language that sounds condescending. Add the exception that changes the recommendation. Those comments leave the reader with better advice.

Even “we have this problem too” can lead somewhere useful. Why does a firm full of smart people struggle with the same issue? What makes it difficult in practice? What does the textbook advice leave out? You don’t need to turn the piece into corporate therapy, but there may be something more useful in those answers than in removing the point altogether.

The same thing comes up when teams use AI to repurpose existing work at scale. Plenty of content should be reused. An accurate definition doesn’t become more valuable because somebody rewrites it for the seventh time. A good explanation of a process may work perfectly well in several places. Let AI handle some of that repetition.

Advice needs more attention.

A CEO considering an acquisition and a team integrating two reporting systems a year later may need some of the same background information, but they’re trying to make different decisions. Shortening the CEO piece and making it more conversational won’t answer the integration team’s questions.

Keep the explanation that still works. Reuse the background. Save the writing time. Then spend some of it with the expert working through what has actually changed for the new reader.

That feels like a much better use of AI than producing 10 increasingly “nice” versions of the same recommendation. 

Give readers a reason to choose you

Publishing polished content is getting easier. Being polished isn’t much of a differentiator when everyone can be polished.

Your competitors can use the same models, summarize the same research, and cover the same industry news. What people can’t get from every other firm is how your experts interpret what’s happening, what they think others are missing, what they’ve seen work, and what they would actually advise someone to do next.

Professional services firms have spent years trying to get their experts into content. The next challenge is making sure their expertise is still recognizable by the time the content goes live.

For CMOs, I’d look beyond how much got published. Put a few recent pieces beside a competitor’s and hide the logos. Can you tell which one came from your firm? Does the piece reveal anything about how your people approach the problem?

And would one of your advisors send it to a client dealing with that issue as-is, or would they immediately write three paragraphs in the email explaining what they really think?

That last one is probably worth paying attention to. If the useful advice only appears in the accompanying email, there was useful advice available. It just didn’t survive the publishing process.

AI can give teams meaningful time back. Use it for repetitive work, rough drafts, and adapting background material that doesn’t need another hour of an expert’s life. Then be deliberate about where the saved time goes. Ask another question. Resolve the disagreement everyone is trying to edit around. Decide that the story isn’t ready yet.

And before the next piece goes live, check whether the advice your people would give across a meeting table is still on the page.

I think CMOs need to change the question they ask in review. Less “Is this safe?” More “Is this actually ours?” Could any firm in the category have published it? Or does it contain enough of your experts’ judgment that a reader could only have gotten it from you?

Because getting words on a page was never the hardest part. Someone still has to decide what your firm actually thinks and make sure that thinking survives long enough to get published.