Everyone knows the story of Kodak. An engineer from the company developed the world’s first digital camera, but leadership suppressed the innovation because they feared it would cannibalise their traditional, highly profitable photographic film business. Other competitors developed the technology and, when Kodak finally recognised the disruption, it was too late. The company filed for bankruptcy in 2012.

It’s a cautionary tale for other businesses that are reluctant, or unable, to give up legacy business models in the face of emerging technologies. The AI revolution seems likely to create many Kodak moments if companies fail to adapt. The industry I’ve worked in for more than 20 years, thought leadership, is one that looks to be at risk.

There are some interesting parallels between the disruption of these two industries. Like Kodak, thought leadership agencies have benefited from a highly profitable model over many years. Producing high-quality research-based content is not easy, which has allowed these incumbents to benefit from high barriers to entry and a wide competitive moat. But these advantages are now under threat as the cost of thought leadership production falls dramatically and as AI provides easy access to capabilities that were once scarce and expensive.

Just as Kodak was reluctant to disrupt itself, so there are institutional barriers to change among thought leadership incumbents. These businesses have built up sizable research, content development, design and development teams, whose work can now be done, at least partially, by AI. Sure, AI will not replace the entire value chain any time soon and there will always be a need for human oversight and input, but there is no doubt that some aspects of the work these teams do look vulnerable. Like Kodak, these incumbents will want to protect profitable, legacy business models and hang onto established teams for as long as they can, rather than go through the costly and emotional process of transformation that will allow them to compete effectively with more nimble, disruptive start-ups.

There are parallels with the product, too. Anyone who remembers dropping off a film to be developed will remember that it took days to get the prints back and, when you did, a good proportion of them were not the high-quality photos you had hoped for. Digital photography made this whole process instant and largely foolproof. In the same way, AI and other emerging technologies are short-circuiting the thought leadership development process, turning a process that clients are accustomed to taking five to six months into one that can be completed in a matter of weeks.

The argument from thought leadership producers has been that AI cannot replace the work that they do. Data from public large language models (LLMs) is by definition derivative of other sources, and therefore is incompatible with one of the core tenets of thought leadership, that it needs to add to the conversation. But this argument is based on a false premise. It confuses generative AI, which simply rehashes existing data, and analytical AI, which enables AI to be used as an engine to process new, proprietary data. Disruptive market entrants recognise that thought leadership must still be based on proprietary insight and that gathering this must be part of the production process. But they also recognise that the way these insights are processed can be significantly accelerated, creating much more rapid time-to-market, greater efficiency and reduced cost.

The thought leadership speed problem

One of the biggest problems hampering thought leadership production is lack of speed. Creating this kind of research-based content is incredibly time-consuming and it’s not unusual for a project to take five months or more. This highly methodical process is robust, but it’s a major drawback because the research that underpins the campaign may be three months old by the time it sees the light of day. There’s always been this trade-off between rigour and time to market, and rigour has generally been seen as the more important goal.

The first obvious problem with this is the loss of topical currency. Thought leadership production cycles are simply not in sync with the pace of business. When we think of the topics that are of most interest to audiences, it’s the fastest-changing ones, such as AI adoption, robotics, data and sustainability, that are most in demand. I can well remember working on a campaign on the “future of work” in early 2020. The research had been conducted and then COVID hit, leading to office closures globally and the partial shut-down of the global economy. Suddenly, our well-crafted research felt unusable given that it predated this upheaval of the business context. Other, less extreme, events happen all the time, meaning that research is often out of date by the time a report is ready for publication. Another problem is that other, more nimble competitors may be able to get their perspectives out more quickly, while the incumbents are bogged down with lengthy cycles.

Long production cycles cause other problems. Very often, workflows are manual and sequential, which means that a huge amount of senior executive time is consumed. The opportunity cost of all these hours providing input and reviewing outputs, rather than earning high fees, can be significant, and is often not factored into the budgets of thought leadership production. Stakeholders also often lose enthusiasm from lengthy projects and may be more wary to get involved in future if they feel that contributing is going to be a big demand on their time. This lengthy process involving many stakeholders and a committee-like approach also often waters down a bold thesis into a safe, generic consensus.

Slow production also means delayed results. Companies want rapid impact from marketing - stronger pipelines, more rapid conversions and increased demand - but this is difficult to achieve if a campaign is stuck in an endless loop of review processes.

Traditional thought leadership production processes and workflows entrench these problems. For example, most thought leadership relies on sizable surveys of hard-to-reach executives. End-to-end, this can easily take eight weeks, taking into consideration research design, programming and testing, data collection and analysis.

Projects are also often held up because the production team or agency cannot find time with senior leadership, whose input is essential in order to shape a campaign idea or perspective. And when they do find time with them, those executives can sometimes question the original approach and move the goalposts, causing projects to get stuck. Leaders may also be needed to review content further down the production line, but if they are travelling or busy, this can cause further delays. Legal and compliance reviews are also notorious for slowing projects down as they debate potential exposures or seek to water down provocative points of view.

So how can thought leadership producers avoid having a Kodak moment? I would suggest a few important steps.

Integrate, rather than just bolt on, analytical AI. Many marketing teams and agencies see AI as a bolt-on that accelerates existing tasks and processes. The Salesforce State of Marketing report found that only 32% of teams have fully integrated AI into their workflows. Employees are often left to figure it out themselves, with the same research finding that 70% of team members say their companies have offered no formal training in the technology. AI gets treated as a cosmetic feature that enhances existing processes and, at worst, all this means is that generic, poor-quality content gets created a bit more quickly than before. AI is just a glorified copywriting assistant.

A more fundamental approach means rethinking workflows from the ground up, building multi-agent systems that can process steps concurrently rather than relying on a lengthy linear process. Consider, for example, the content development stages. This typically relies on a core asset that is developed first (usually a long-form report) with various spin-off content assets such as videos, articles and presentation decks being produced afterwards. Often, a last-minute change to copy on one asset means reworking the others, too - a tedious, lengthy, manual sequence. Instead of this linear process, an asset builder agent enables parallel production. It acts as an orchestrator, loading LLMs with raw manuscripts and brand design guidelines to automatically generate a variety of web native publications. Last-minute changes can be updated on the core manuscript, and flow through automatically to every content asset.

Explore alternative research methodologies. For decades, the traditional online survey of executives has been the default way to gather insight for thought leadership content. But it is expensive, prone to quality issues, and also incredibly time-consuming. Fieldwork alone can often take many weeks. Other options are quicker and potentially more reliable.

AI-led qualitative interviews are one option. This approach can gather insight more quickly than online surveys, and often generate much greater depth because a more open, probing questioning approach is possible. Synthetic data can also be used as a “digital twin” to mimic traditional survey respondents. Results are very promising in this space, although I have so far noticed a reluctance on the part of thought leadership producers to use this beyond topping up traditional surveys or testing hypotheses. AI also allows for "passive mining": ingesting thousands of data points from earnings calls, customer reviews or analyst reports to identify trends and pain points in hours rather than relying on extensive desk research taking many weeks. All of these approaches can be quicker, cheaper and more innovative than traditional surveys.

Gather insight asynchronously. Marketing teams often struggle to get the hour in a senior stakeholder's calendar that enables them to get the foundational insight they need. This creates delays that may mean production can stretch out for many months. These same stakeholders may also be needed to approve drafts later in the production process. But this is again subject to delays. If an email sits in a business leader’s inbox for five business days because they are travelling, that's a week's delay to launch that hadn’t been anticipated.

The first problem can be addressed by a more asynchronous approach to insight gathering. Rather than download their expertise in a 60-minute interview that may take many weeks to organise, SMEs can answer targeted prompts via voice notes or custom internal AI bots, allowing producers to capture expertise and use AI to synthesise it into outlines and messaging. The second problem can be addressed by a shift in how the SME input is requested. Rather than sending a 20-page report and asking “what do you think?”, consider excerpting relevant points for approval, using an AI script, that clearly links back to the initial input they provided. And, rather than expecting extensive written comments, allow them to leave voice notes that can be transcribed and converted into clear changes to the document that address their feedback.

Move away from rigid formats like PDFs. The PDF has been the mainstay format for thought leadership over many decades. Although it is portable, sharable and widely associated with B2B thought leadership, it has many drawbacks. Producers find it difficult to analyse how audiences are engaging with them, and audiences experience friction downloading them and navigating to the content that is relevant. They are also fiddly and time-consuming for design teams to work with. Simple, last-minute changes from legal or senior stakeholders can throw out page layouts and require extensive redesign.

Shifting to web-native, modular layout platforms can eliminate a lot of this friction. Writers can drop copy into pre-designed templates that adjust automatically and dynamically as text is revised. Development also no longer needs to be a bottleneck. AI front-end generators (such as v0, Lovable, or Claude Artifacts) allow teams to paste a content brief or wireframe prompt and create production-ready code in minutes, entirely bypassing traditional web development backlogs.

Thought leadership buyers no longer want to wait months for campaigns to drop, just as photographers got bored waiting for the film to develop. B2B audiences and buyers demand instant, highly relevant and easily digestible insights, not findings that may be many months old by the time they are shared. Thought leadership producers can either hold onto the comfortable profitability of the darkroom, as Kodak tried to for several years, or they can embrace the digital revolution and undergo the necessary transformation. Those that refuse to adapt will eventually find that their competitive moat will disappear. The services they used to offer will migrate over to disruptive new entrants, or to in-house teams that have mastered how to use these tools themselves.

‍