TOP LINE
Welcome to our first AI and thought leadership special report. This study explores how generative AI (GenAI) is being used in thought leadership across technology, telecoms and professional services in North America and Western Europe.
While it might not be the most exhaustive study published this year, it highlights a single, major trend that leaders across these sectors consistently recognise.
You’ll find different views on the most effective use cases for GenAI, the tools in play and even the long-term impact on roles and teams. There are also – inevitably – differences in what respondents want to achieve from AI adoption.
But one finding stands out clearly: the area in which leaders believe AI delivers the greatest benefits is speed. AI enables organisations to produce thought leadership and other long-form content far more quickly (although not necessarily to a higher standard).
And that leaves us with a conundrum. If you move faster, what are you going to do with that extra time you free up? Is that time noticeable? What will you want to do with it? What will your employer want you to do?
Read on for our full findings – and what we think that extra speed and time could deliver for content leaders and their teams.
What do we mean when we say AI?
In this report, generative AI (GenAI) refers to AI systems – particularly large language models (LLMs) – that perform functions such as text generation, research, image creation, data analysis, brainstorming and more.
5 key findings on generative AI in B2B thought leadership
- Almost all respondents report they are using GenAI in some way for their thought leadership and other content – only 5% do not use it at all
- Speed is the most cited benefit
- Speed ranks above improvements in quality or lower cost
- Western Europeans are more cautious, but in some ways more advanced, in their uses of AI than North Americans
- The key benefits of AI usage are more related to workflow acceleration than to improvements in writing or editing quality
INTRODUCTION
Why did we do this research?
We set out to find two things: how senior leaders in the
technology, telecoms and professional services sectors are using AI for thought leadership, as well as for content and insights more generally.
Secondly, we wanted to know what their positive and negative experiences have been so far.
We think we unearthed something clear. But also something challenging.
The answer to our first question was simple: almost everyone is now using AI in some way for their content.
While our survey only went deep on those who are using it, we encountered about 5% of the people we approached who answered that they aren’t using it at all. (We filtered them out, so they played no further part in our research, but it’s worth knowing that number.)
Of the rest, meaning those who make up this study, we found a mix of views on the best use cases for AI.
Yet these senior leaders responsible for thought leadership in their organisations identified one clear benefit: improved speed.
And, as in our conclusion, we need to ask what we’ll do with all our extra time if this speed is enduring. Because we have a choice: do we do the same kind of work that we do now, only better? Or, do we fall foul of Jevon’s Paradox*, and increase the volume, not the quality?
While we’ve all heard of ‘AI slop’ – low-value, AI-generated content that’s easily spun up and flooding the internet – there is a risk that the most advanced B2B producers of thought leadership could add volume by seeing an opportunity for more, more, more.
Or we could use our extra time to serve our audiences better. To stand out. To win with the best thought leadership.
* The theory named after Industrial Revolution economist William Stanley Jevons, who recognised that more often than not, any gain in efficiency just leads to more of a thing being done.
Who did we survey?
To see full details about those we surveyed, go to our Methodology section at the end of this report. But it’s worth saying up front that we spoke to senior decision-makers at large companies in tech, telecoms and professional services. Half of them were in North America, half in Western Europe.
Why this mix? Because it broadly mirrors our agency’s client base, and it’s a mix of companies and individuals who are generally advanced in their use of AI and not afraid to share their hopes and fears, as well as what’s
already gone well or poorly.
While there is a clear thread through this research, there are contradictions that we’ll get to. These show both a contrast by region, and also respondents’ personal weighing up of AI’s benefits and its downsides.
WHAT'S YOUR USE OF AI NOW? WHAT WILL IT BE IN 12 MONTHS?
You can see that over three quarters are beyond experimenting and over a quarter (28.0%) answered that they “use it extensively across many content workflows”. This last answer was selected by 36.0% of those from Western Europe, but only 20.0% of North Americans.
As Josselyn Simpson, Vice President, Global Editorial Director, Heidrick & Struggles told us during her interview for our ‘Sense-maker sessions: Content leaders’ podcast series: “Everyone will soon, if not already, need to be AI fluent in the way that maybe 15 years ago there was digital fluency.”
We then asked our participants a broad question: what percentage of their team’s published or otherwise public-facing content output (across all formats) now involves GenAI in some capacity.
GenAI specifically is in use to varying degrees. As the chart shows, over three quarters (78.0%) are still using it for the minority of their output. Those in the top quartile output sliver make up just 2.0% of responses.
However, towards the end of our survey, we asked about how the use of GenAI for content might change for them over the next 12 months.
You can see over a quarter (32.7%) answered ‘Increase significantly’, the majority of respondents (61.3%) said ‘Increased moderately’ and just a small proportion (6.0%) answered ‘Stay the same’.
Clearly, AI is widely in play, and still growing inside the majority of companies.
However, this was the first area to show some major differences in the Americas/ Europe comparison.
Our participants in North America were more likely to be in the 1–25% category for the first chart (50.7% of them versus 24.0% of the Europeans), while in the second chart, the numbers saying they plan to ‘Increase significantly’ were 24.0% for North America as opposed to 41.3% for Western Europe.
WHAT'S YOUR USE CASE?
We then asked about common editorial tasks and whether GenAI is currently in use.
The charts speak for themselves but some areas we’d highlight are:
- AI is commonly in use for research summaries and idea generation/ brainstorming
- But it’s unpopular for SEO tasks, visual content and compliance reviews
While there were some slight variations, the findings were largely aligned by region and job title.
Similarly, we asked about content types they have created entirely or in part using GenAI.
We weren’t surprised to see GenAI commonly employed for blog posts (57.3%) and LinkedIn posts (52.0%) but less commonly used for other social posts (9.3%) or video/films (24.7%).
GOALS
What were our respondents’ goals, specifically when using AI for content creation?
This is where speed really arose as a key priority, sought by 60.7% of our sample, and it featured in a number of the qualitative replies we received.
However, there were also some surprises in responses here. Only 22.7% say they use GenAI to reduce costs and just 20.7% are using GenAI for experimentation or innovation, a number that’s largely aligned across regions.
And there is a potential contradiction in ‘speed’ being our number one area of benefit, yet only 42.0% state they seek to use GenAI to reduce team workload.
Speaking to us as part of our ‘Sense-maker sessions: Content leaders’ podcast series on YouTube, Lucia Rahilly, Global Editorial Director & Deputy Publisher at McKinsey & Company, told us: “So, everybody’s using AI to see where they can accelerate movement through the chain. And that’s a really good thing to me because… the faster we can get good ideas to market, obviously, the better.”
QUALITY
We then asked the six-billion-dollar question:
For this question, we gave only four options, and you can see it’s a split decision. But as we sit here in 2026, only just over three years after the GenAI revolution began, almost half of respondents told us the quality of GenAI-produced content for thought leadership is either ‘Good’ (40.7%) or even ‘Excellent’ (8.0%), even if the most common answer was ‘Adequate’ (44.0%).
The mixed feelings about GenAI were apparent in some of the qualitative responses we received. One respondent said that on the positive side, “The main value has been in accelerating research synthesis and converting insights into usable content more quickly,” but that “AI-created content is too generic or [exhibits] bias that may dilute or risk unique brand identity”.
Tools
When it comes to the AI tools in play, there were some nuances and differences by region.
When asked which tools our respondents rely on, they told us:
- Public LLMs (e.g. ChatGPT, Claude, Gemini etc.): 78%
- Enterprise/Secured instances of these LLMs: 54%
- AI features in software (e.g. in Microsoft, Adobe, Salesforce etc): 65.3%
- Proprietary/Company-built AI tools: 35.3%
They were free to tick as many of the four boxes as they liked. What’s clear is that almost a quarter don’t rely on public LLMs at this point. And, of those who do, almost none solely rely on them.
That said, for a sample made up of tech-centric companies, we were a little surprised that the number against the last category, Proprietary/ Company-built AI tools, wasn’t higher.
In terms of the tools our sample uses in their companies, they told us they could be happier.
SCORING AI
We then asked about eight specific areas where our respondents graded GenAI’s capability in relation to thought leadership.
You can see that GenAI again scores well in terms of speed, or ‘time taken’ as our chart says here. It is also strong in terms of cost.
However, less appealing is AI’s effect on ‘originality/differentiated points of view’, ‘emotional intelligence/audience understanding’, accuracy, ethical/compliance issues and working with proprietary or
sensitive data. Brand and tone alignment came out fairly neutrally.
Here are some comments our respondents made:
• “Biggest benefit is organising ideas and exploring new perspectives.”
• “AI often struggles with technicality or highly specialised subject matter.”
• “The GenAI authoritative tone can cause team to skip critical decisions.”
On that last point, accuracy, it’s apparent that hallucinations are still a problem. This is where LLMs fabricate facts, even down to including third-party citations for made-up reports and other references. This was born out in comments such as: “Output appears convincing but has slight errors” – which perhaps underplays some of the risks.
Even well-constructed content can be based on outdated information. We heard: “We have seen instances where the information is outdated, so we still need to validate everything against current sources.”
Other downsides
‘Hallucinations’ comes near the top of our list of errors and challenges. However, heading the list of negatives was ‘generic or lowest-common-denominator content’.
We have seen this time and time again in our experiments or those of our clients. Here we heard comments such as: “[GenAI] can produce content that lacks emotional depth or storytelling impact.”
Only 3.3% of respondents (just 1.3% in North America) told us they have had ‘No significant issues’.
AI'S PLACE IN THOUGHT LEADERSHIP
Thought leadership offers one of the highest returns on investment in B2B marketing and is the gold standard for insights and content.
This part of our study asked about areas of a flagship thought leadership report that could be confidently automated with GenAI today.
Which parts of the workflow could you confidently automate with GenAI today?
To start at the end, hardly anyone replied that they thought leadership should only be human-led. That shows how far we’ve come in a relatively short period of time with GenAI.
And you can see that AI scores well across many areas, most notably topic ideation and background research summaries, something that we’ve experienced first hand across our work for companies in these sectors.
In general, across this topic, North American respondents were more bullish about AI usage but they were particularly more positive, by 5–10 percentage points, in these two areas.
Where AI scored more weakly was for creating graphics and visuals and – in line with previous comments – fact checking.
This section of our study also underlines that using AI is about so much more than writing and editing.
HUMANS ONLY
We then concluded with another way of framing AI’s impact and opportunity. We asked: Do the following areas still need human involvement?
Humans still very much lead the way when it comes to arguably the biggest differentiator, that is ‘Developing original ideas and perspectives’.
Similarly, human endeavour outweighs AI for ‘Industry expertise and applied insight’ and ‘Structuring arguments and narratives’.
However, AI scores well for ‘Deep research’ and even ‘Tone and brand voice’.
On that last subject, qualitative comments from participants were varied. One told us: “[AI] requires extensive editing to ensure accuracy, originality and brand voice.”
Another mentioned that AI helps in this area. “Helps maintain consistent tone and voice across all content,” they said.
However, we were surprised to see such maturity for AI in terms of executive and stakeholder sign-offs.
And while AI represents a risk for compliance, you can see that it can also be an aid – only 14.7% of the time did our respondents say compliance is an area requiring the most human involvement.
AI'S FUTURE ROLE
What were our participants’ final comments on AI’s role, especially in relation to thought leadership and how their teams are structured?
Already, fully 18.0% now see AI as a ‘core creator of content’. Meanwhile almost exactly half (49.3%) – the largest answer to our question – consider AI a ‘collaborator alongside humans’.
The rest see AI much more in support or part-time roles but together they make up barely a third of all replies.
When it comes to AI displacing humans, most tell us that jobs are safe where they work.
Only 5.3% see AI displacing ‘multiple roles’ (although Europeans are three times more likely to say that), and 26.7% answered ‘Yes, one role’. We have no way of telling if that role was theirs.
BOTTOM LINE
We can draw several conclusions from our study but one leads the way: that AI reduces the time it takes to create content. This is seen in one further question we asked:
Put aside that 4.7% of respondents told us that AI had added time (a figure that was 8.0% in Western Europe). We can also see how that can happen but we don’t have more data on how that happened for them.
Conversely, a solid 48.7% report some reduction in time spent on content production and 18.7% selected ‘Significant reduction’.
We should point out that only some of the time saving is about the writing and editing of content. The analysis in this report so far shows other areas where AI is making a real difference.
To take just one example, one participant told us: “AI enables faster sense making across large volumes of information.”
However, it’s clear that humans are essential for much of the content creation process and especially the judgement and strategy that goes into the best thought leadership. In short, we’d argue that the hardest parts still need humans.
Original ideas, perspectives, and industry expertise are all still better handled by people.
At this year’s Global Thought Leadership Institute Symposium, the phrase ‘Human in the lead’ (as opposed to the old AI safety motto of ‘Human in the loop’), was coined by Accenture and quickly became shared language in the room and the wider community.
As for the trade-off between AI’s speed in relation to its downsides, one interviewee said: “AI speeds up research and idea development significantly [but] content often feels too generic and requires heavy human refinement.”
And some would argue that AI’s use for the drafting part of content creation is a false economy altogether, given the amount of fact checking and editing that is needed.
So carry on experimenting with and putting AI into play. But be realistic about where humans are always going to be better. And – just as we’ve seen with areas such as tone of voice or compliance – that won’t always be where we assume humans are needed. Just as spinning up copy with AI is rarely good enough for serious producers of thought leadership.
EXTRA TIME
Lastly, if speed is identified as the number one goal and number one benefit, not just according to our modest sample but across the content world, where does that leave us now?
If speed delivers us more time, will we (as Mr Jevon predicted) just create more? Or is this a chance to really think about what consumers of our content need?
Sometimes we need to get our insights to market fast. But for the majority of our work, will we put in extra time and effort to prioritise better over more, and differentiate what we do?
The industry should. And that’s what this agency would like to see.
Who approves?
We were curious to see who signs off on AI-assisted content.
The results were largely as we’d expect, although the answer ‘Marketing/Communications leader’ was more common in North America (41.3% versus 28.0% in Western Europe).
We also worry for that small sliver of respondents who answered ‘No formal process’.
METHODOLOGY
Field work was carried out during January 2026.
We polled 150 content leaders, half in North America and half in Western Europe. We surveyed senior roles only.
Respondents came only from medium-sized or large enterprises, including 33% from 20,000+ companies. These are the types of companies that typically have in-house content teams and use more than one external agency.
Industry spread was across technology (50%), telecoms (10%) and professional services (40%) sectors.
We surveyed senior roles only, broken down as C-suite (12%), VP / SVP (e.g. Thought Leadership, Marketing, Content Marketing, Digital Marketing, Strategy) (21.3%) and Head / Director / Senior Director titles (66.7%).
Participants confirmed their roles were spread across thought leadership, editorial, research/insights, marketing/communications and strategy.
Qualitative responses were captured as free answers during our quantitative data gathering.
We’re keen to make sense of AI’s opportunities and risks in thought leadership.
Speak to us about AI and any aspects of content, insights and thought leadership.
We’d love to hear from you.
You can email us at info@collectivecontent.agency, call us on +44 800 292 2826 or find us online on LinkedIn.
With thanks to our research partner High Beam Global, a full-service market research firm.
Tech Quarterly
Our quarterly summary of top research, market stats, new developments and predictions in five key technology topics of importance to our readers and clients: artificial intelligence, automation, CIOs, Internet of Things and virtual reality/augmented reality/extended reality (VR/AR/XR).
If a major analyst report, survey or forecast has been published on any of these topics in the preceding three months, you’ll find out about it in Tech Quarterly.