“Why are we doing this?”

It was a somewhat disarming question. The person posing it was a marketing director, and the question was asked in a meeting about planning the next in a series of thought leadership reports that the company had produced for many years.

“Well, we’re doing it because it helps position us in the market, strengthen our authority and hopefully unlock commercial opportunities,” replied one of the marketing director’s close colleagues.

“I know what we are supposed to get out of this,” said the marketing director. “But how does it help the audience? What’s it really for?”

It was an important challenge and one that doesn’t always get enough attention. As a consultant working with thought leadership producers and the former CEO of a large agency in this space, I have had dozens of conversations about the value of thought leadership to the company producing it. There is rightly a lot of attention paid to how these investments translate into tangible value, whether that is improvements to brand equity or reputation, or commercial success.

But this focus on measuring value, while crucial, tends to overlook another important factor. Discussions are too often framed around the value for the producer rather than the value for the audience. Why do audiences consume all this content, and what do they get out of it? How do they use it in their business and what lasting memories does seeing or reading this content create?

There is, of course, a broad quality spectrum when it comes to thought leadership. Some of it is excellent, while a lot of it is pretty poor. Better examples with good distribution campaigns will be remembered, shift perceptions of the producer and might influence future buying decisions. Those at the lower end of the quality spectrum will largely be ignored.

Edelman’s 2024 B2B Thought Leadership Impact report found that 73% of decision-makers say an organisation’s thought leadership content is a more trustworthy basis for assessing a company’s capabilities than its other marketing materials. This is what “good” thought leadership should do. Its job is to create an association between the brand and its chosen area of expertise and to create trust in capabilities to do a job. Audiences don’t even need to look too closely at the content itself. They just need to form an overall favourable impression by recognising that the producer has gone to the effort of researching and publishing content on an issue that is relevant to their business needs.

The value of great thought leadership

But surely “great” thought leadership ought to do more than this? The ultimate goal should not only be about strengthening associations and memories but about shifting markets, changing conversations and influencing corporate decision-making. How often does that happen? I suspect it’s quite rare but it’s not unheard of. My old company FT Longitude found in one piece of research that 76% of business leaders say thought leadership can help them make better business decisions. I’d highlight that “can” is an important word there. Although I don’t have data on this, I suspect that most thought leadership reports are not being passed around the boardroom table. It’s wishful thinking to expect that they are.

So if you want your thought leadership to be truly influential, what do you need to do? The traditional answer to this question is that you provide insight that audiences can’t get elsewhere. You explain complex issues and cast a light on what the audience’s peers (including their competitors) are thinking and doing. This provides useful context that helps leaders to navigate difficult decisions. The very best thought leadership should also disrupt current modes of thinking, challenge accepted truths and bring new, valuable information to the table.

AI as an input to decision-making

That was already a difficult thing to do but the advent of AI has made it even harder. One of the strengths of traditional thought leadership was its data. It was unashamedly skewed towards the left brain, highly analytical and quantitative. Audiences valued that aspect because this information was difficult to find elsewhere. Yet these strengths of thought leadership are precisely the areas where AI is also highly effective. The concept of “proprietary data” is becoming increasingly elusive and that reduces the value of a lot of thought leadership.

In the past year or so, there has been a lot of research into the potential for AI to not only support but replace traditional decision-making. In other words, to perform the role that was, until recently, helped by decision-making inputs such as the best thought leadership (along with many others, of course).

In one simulation of the automotive industry (1), researchers from Judge Business School found that LLMs outperformed human decision-makers when it came to product design, supply chain and cost optimisation. The models generally did better than humans in controlled environments when the situation was predictable. They were, however, less successful when they needed to deal with uncertainty, ambiguity or “black swan” events.

Another study from Cornell University (2) looked at the extent to which LLMs could match humans in different creative tasks. It found that these tools could rival human performance in structured problem-solving but were less effective at creative writing and presenting a diverse range of ideas. And although the LLMs were good at generating ideas, the researchers concluded that human input was still essential to assess their novelty and usefulness.

Although LLMs are far from perfect, the implications are clear. They will become increasingly important as inputs to strategic decision-making that could leave more analogue alternatives, such as great thought leadership, out in the cold.

So should thought leadership producers simply give up and accept that their work will, at best, create associations between their brand and certain business topics, as opposed to genuinely influencing the audience? I would argue that they shouldn’t throw in the towel, but they do need to recognise that their work requires a different emphasis. I’d make three recommendations.

Don’t rely on data alone. We’ve seen from numerous academic studies and real-life experience that LLMs excel at presenting and processing vast quantities of data and evaluating decisions based on these inputs. Thought leadership studies that over-rely on data will simply not be able to compete. Data is still important as a foundation but doesn’t have the intrinsic value in these studies that it once did.

Be more human. LLMs are currently poor at evaluating the emotional dimensions of decision-making. Thought leadership producers need to dial up this element of their work, exploring the complexities of decisions and how emotional factors can influence them. There needs to be greater emphasis on the psychological drivers of different situations, and more acceptance of how these can lead to a range of outcomes.

Accept uncertainty and ambiguity. Corporate thought leadership producers have traditionally shied away from too much focus on uncertainty, worrying that it makes them look less confident about a particular set of recommendations. But we have seen that coping with unpredictability remains a key weakness of LLMs. So, if producers want audiences to use their content as decision-making inputs, they need to embrace ambiguity and the different implications that it brings. Don’t downplay it or pretend it doesn’t exist - celebrate and emphasise it.

‍

‍

(1) Mudassir, Munir, Ansari, and Zahra (2024) "AI can (mostly) outperform human CEOs", Harvard Business Review

(2) Sun, Yuan Yao, Li, Zhang, Xie, Wang, Luo, and Stillwell (2024) "Large language models show both individual and collective creativity comparable to humans" Cornell University

‍