In this episode of Thought Leadership in Practice, iResearch Services CEO Yogesh Shah speaks with Liz Bolshaw, Global Content Strategist at EY and board member of the Global Thought Leadership Institute (GTLI). Together, they set two of this year’s biggest studies on thought leadership side by side: GTLI’s Promise and Peril, which surveyed 1,000 C-suite consumers of thought leadership, and iResearch Services’ Thought Leadership in Practice, which surveyed 1,000 of the practitioners who produce it.
Liz shares what the data looks like from inside a Big Four firm: how EY decides where to invest, why executives are wary of AI-generated content even as they use AI every day, where EY draws the line on AI in its own work, and how the firm measures return on thought leadership in a zero-click world.
This conversation gives thought leadership, marketing and research leaders a candid view of how one of the world’s largest producers is handling the shift to AI, and what it means for firms of every size.
Key Takeaways
- Keep humans on the hook: EY uses AI for data analysis, white space analysis and hypothesis testing, but final writing and output are off-limits. Liz argues that AI looks in the rearview mirror, while good thought leadership needs original, forward-looking ideas and someone who takes responsibility for them.
- Explain your AI use, and sell its value: Executives distrust content they think is machine-made. Producers need to be transparent about how they use AI and explain how it has deepened the insight, rather than simply noting that it was used.
- Actionability is the gap to close: Promise and Peril found that executives rank actionability of recommendations fourth, while producers rank it seventh. Every piece should tell readers what to do next.
- Niche players can now punch above their weight: With generative engines citing specialist titles and expert sources, smaller firms with deep expertise can occupy white space and challenge the biggest brands. As Liz puts it, be a specialist guide in your village rather than a generic tourist.
- Measure each piece against its own goal: EY weights reach, engagement, downloads, tier-one media citations and mentions in earnings calls differently depending on what each piece is meant to achieve.
Yogesh Shah in conversation with Liz Bolshaw
Yogesh Shah: Hello, Liz, and welcome, listeners, to the Thought Leadership in Practice series. With me here is Liz Bolshaw from EY. Liz, thank you very much for joining us today. We’re going to spend a lot of time talking about thought leadership, so for people who don’t know you, would you mind introducing yourself and telling us what you do at EY?
Liz Bolshaw: Thanks very much, Yogesh. It’s a great pleasure to be here. I lead our editorial division in the Insights part of EY, which is the global part of the firm responsible for most of our research-led thought leadership. My title is Global Content Strategist, but I think that’s slightly overblown, really.
The main thing I do is editorial. We liaise with the business, with our research team and with our web and digital team, so we sit in the middle of all of those things.
Yogesh: Interesting. Fantastic. Let me introduce myself: I’m Yogesh Shah, CEO of iResearch Services, and I have the pleasure of talking to Liz today. I think we’ve known each other for many years now. So let me set the table for this conversation, Liz.
This year, the GTLI, the Global Thought Leadership Institute, on whose board you sit, produced a study called Promise and Peril, in which you spoke to 1,000 C-suite consumers of thought leadership. A few months later, we at iResearch Services produced a report called Thought Leadership in Practice, in which we spoke to 1,000 practitioners, or producers, of thought leadership.
These are two big studies about thought leadership, and there’s been no real opportunity to discuss these two big data sets together. I wanted us to take the insights from these two studies and talk about them, so that people can understand what they really mean.
Both studies will be in the show notes if you want to download the reports. But let me go straight to you with a question. For people listening to this conversation today, what’s in it for them? What can they hope to find in this conversation that they might not find in the reports?
Liz: I think, Yogesh, it’s the lived experience of those of us who are involved in thought leadership, whether you’re at the beginning of your career, and you’ve come in with AI very much a reality, or whether you’re like me, long in the tooth, and you’ve seen how AI is already shifting the landscape. It’s shifting the landscape for our clients and our audiences, but also internally.
We always think of ourselves as “client zero” with this, so we’re going through a lot of the pains that our clients are experiencing. I would hope that in this conversation we can get under the skin of some of those data points and explore what they really mean in practice.
Yogesh: That’s fantastic. Fair enough. So let’s begin: how would you describe the culture at EY around thought leadership?
Liz: We’re a typical professional services firm. We’re a Big Four firm and a big global firm, and I would argue we’re the most globally integrated of the Big Four. I think we’re a partnership of about 500 different firms, so it’s not an uncomplicated organization. The culture is highly collaborative. We are expected to make decisions after heavy collaboration with each other, and we do.
That’s a great positive and a great strength, but it can also slow us down. As the world moves faster, that’s a challenge we’re really meeting at the moment: exploring how to do things faster. That’s undoubtedly something AI can help with.
Yogesh: Wow, that’s complicated. Did you say 500 firms? I’m guessing “fast” is part of the day-to-day lingo, in every conversation.
One of the insights from our report is that thought leadership professionals wear around 2.7 hats, so they take on multiple roles and responsibilities. One of the stats says that 81% of people who work in thought leadership have more than one responsibility. Is that reflective of EY? Or can a firm like EY escape that?
Liz: I think bigger organizations, and obviously we’d be one of those, and they’re certainly represented on the GTLI advisory board, probably have greater role divisions than smaller companies. If you went to a thousand producers, and I think I’m right in saying that about 80% of thought leadership is produced by about 15 organizations, you’ve got a lot of smaller players.
In that circumstance, you’re likely to need to work across a lot of different skill sets and roles. We’re still broad. We still have to understand the importance of roles that are not, in my case, strictly editorial. This is my second time at EY; I’m a boomerang. My first time round, I was an analyst, very firmly in the research community. We have to understand how analysts approach things, and we also have to understand the business and the commercial realities.
So there are quite a lot of things you need to understand to do the job well. That doesn’t mean you’re doing everything. But I think almost anybody in thought leadership needs to be a broad person, curious about lots of different things.
Yogesh: So curiosity is one of the big skills in thought leadership.
Liz: 100%, in my view.
Yogesh: From an inside view of EY, what made you join the GTLI board? What interested you?
Liz: Well, I was actually one of the founding members, and you, Yogesh, had a part to play in this, because we were all at a conference you held in London. After I spoke, I think it was on ESG, Anthony came up to me and said, “We need to talk.” I knew Anthony by name, but I’d never met him before. We had a conversation about how there was a real need for a body like this: one that looked across the industry, set some standards and showcased best practice.
So I was delighted to be part of it. Of course, it’s grown since then, and we now have Jeff Pundyk from Deloitte, people from McKinsey; it’s a very broad church. We’ve got lots of very eminent people, far more eminent than me.
Yogesh: I’m sure you’re equally distinguished, to be on that board. Thank you very much. I can’t believe it’s been, what, four years or more?
Liz: I don’t think it’s been that long. Maybe three years.
Yogesh: I’m glad we’re doing something for the community. I want to jump into the research and insights. What we’ve seen is that thought leadership is messy, in a good way, because, as you agreed, people play almost three different roles.
Almost half of the people we have insights from manage one or two additional responsibilities, but they don’t hold responsibility for strategy and budget. About 39% have a dedicated thought leadership function, or a structured place for it, and of those, only about 19% are responsible for both budget and strategy. I’m sure EY has a much more structured way of managing this, but what’s your take?
Liz: I don’t think it’s surprising, and it certainly speaks to some of the challenges many people face in this space. At EY, going back to the collaborative culture, strategy is set very much by our general executive, which is the equivalent of our board, and that includes our priorities.
Those priorities tend to be set for two or three years, and our insights development mirrors them to a large extent. We think about things in several ways. There are services where we’re number one or number two in the world, and we want to maintain our position in those. So it’s very important that we continue to provide our clients with really good frontier research in those fields.
But like most firms, we also have a growth strategy, and we’re responding to changes in the environment. Where we don’t yet have that reputation, I think thought leadership has a particularly important part to play. So we look at strategy, and the money follows the strategy for us, very substantially.
Yogesh: That’s good. Let’s compare that to the standardization the GTLI introduced. How did it come about? What was the need for standardization in the first place?
Liz: It started with a recognition of how disparate the market is. As we’ve just said, you went to a thousand producers; there are a lot of people. If you look at the way those organizations do their thought leadership, even the definition of what thought leadership is varies a lot between firms.
Many people produce things I would consider content marketing and call it thought leadership, and others produce what I would consider academic papers, which have absolutely no place in—
Yogesh: Or a couple of sustainability documents saying, “We are a sustainable company, and this is a thought leadership paper.”
Liz: Exactly. So when you start looking at things, there is a real need. It was also a challenge to ourselves, because it’s very good to be forced to think: how would we define this, and what kinds of standards would help us benchmark where everybody is? And then how can we promote best practice, train people and do all the things that come out of standardization?
Because we sit under the APQC, and they are experts in standards in a way that I’m certainly not, I think we had a head start.
Yogesh: I’m sure it can’t have been easy to reach a standard, because you want the entire industry to put its weight behind it. When you compared notes with like-for-like companies, were there disagreements? Were there areas where consensus was difficult to reach? Can you talk to us a little about that?
Liz: You know what? Strangely, there weren’t, and I can be totally honest here. One of the great pleasures of the GTLI is that we’re a very collegiate band of people and terribly well behaved. We very rarely try to get secrets from each other, or commercial advantage, or engage in one-upmanship.
I think we were all very aligned on the boundaries between things like thought leadership and content marketing, and on the quality standards that make a piece of work really stand out. The difficult thing is how you do it, and I don’t think we disagree about that. I just think it’s hard. It’s like saying, “I know what an elite marathon runner needs to do, and I know what they look like.” That doesn’t mean I can go and run a marathon. Believe me, I can’t. Those are different things.
Yogesh: Do you have a number for how many people have done the GTLI training so far?
Liz: I don’t, because I don’t sit in that part of it. I know we’ve got a session coming up in London, which I’m going to be joining, and I’m really looking forward to it. I think it’s probably somewhere around 100, and it’s growing all the time.
The first cohort went through at our symposium in Houston, and I think there were about 30 or 35 people. Then we did something in New York, but I’m not really on top of that, so I don’t have the numbers.
Yogesh: That’s fine. So, Promise and Peril, the report you produced: as I said, it covers almost 1,000 consumers of thought leadership, and it also includes around 350 producers. You also did a cut of about 15 top producers, which happen to be the GTLI board members. What really surprised you in this data? And what are the promise and the peril, according to you?
Liz: The Promise and the Peril are all about using AI in thought leadership. I came up with the title, and I’ve noticed everybody else is using it now, which is good. The promise is that AI allows us to do many things much faster, and I would say much better, than in the past. We can now analyze enormous data sets that would have been impossible to analyze before AI without spending a huge amount of money.
The peril, of course, is the other side of it: generative AI slop, and the fact that we’re now living in an era of immense content output. There’s so much noise in the market, and some of it is, unfortunately, weak, poor quality and sometimes erroneous.
All of those things are perils, and they don’t just harm the organization responsible for them. They harm us all. And that brings me back to the report: what we’re seeing is real resistance among clients to accepting that anything AI-produced is valuable. Personally, I think we all have a responsibility to be transparent about how we’re using AI, and also to explain, in a positive way, why we’re using it.
That means explaining how it has augmented and deepened insights in ways we couldn’t have achieved without it. I don’t think any of us are doing that particularly well. We’re transparent, certainly: we have strong methodology guidelines, and we have to say how we’ve used AI. But I don’t think we necessarily talk up the decision to get there, or where the future lies, because it’s moving so quickly.
We were in the field nine months ago with this research, and I reckon the landscape has changed in those nine months. It really has.
Yogesh: I’m sure. I remember one of the points discussed at TLFT, without naming the Big Four firm involved. There was an issue with an AI-led study in Canada, and there was one in Australia where a firm had to pay something like half a million dollars back to the government, if I’m not wrong.
That’s not only a bad reputation for those professional services firms, but also for an industry like ours, because, as you rightly said, we have to hold the standards.
Liz: Absolutely. What all of us have done as a result of those cases, and there are other instances, is to really strengthen our governance around thought leadership creation and make sure things don’t slip out. These things can slip through the net, but we certainly have very strong governance now, and I think most people do.
Yogesh: I know we’ve spoken a lot about AI, and I want to do a deep dive on it. But before we do: when you compared the top-line Promise and Peril data with the cut of the top 15, to make an apples-to-apples comparison, did any insight or stat surprise you?
Liz: It surprised me a little how much smaller producers, and obviously not all of them were very small, just slightly smaller than EY, were struggling with budget, for example. That was the number one obstacle to doing great thought leadership: budget, skills and people, which are all budget-related.
For us, by contrast, it’s definitely the complexity of the organization, and what you started with: establishing a strategy and really sticking with it. So there were some surprising things. The other difference was the way these organizations look at success and measure return on investment.
Smaller organizations tend to look for lead generation and commercial payback, and they want to track that very specifically. Bigger organizations tend to focus more on brand and reputation, so the beginning of the funnel rather than the end, if you like. That perhaps isn’t so surprising. What surprised me was that there were so many stark differences, more than what the differences actually were.
Yogesh: When you say smaller organizations focus more on lead gen, I think we agree that’s not wrong. It’s also a strategy that can be delivered through thought leadership. Brand awareness and credibility are important, but so are lead gen and demand gen.
Liz: Absolutely. Marketing colleagues always talk about the customer journey and the funnel. I much prefer to think of it as a flywheel, because you go round. You might start by reading a piece of thought leadership you’re really impressed by. You then see who’s written it and think, “That’s interesting. I’ve heard of these people before. Maybe I should check them out.” You then go deeper into the organization itself, find something else, and that persuades you to go round again. So I think we’re in a different world now.
And clearly, with AI answers from Google and others, we’re seeing our work surface in completely different ways, I’m sure often anonymously.
Yogesh: You make a very good point. Something I read recently uses a playground analogy: customers no longer move in a linear way through awareness, interest, desire and action. It’s like a playground. You come in through one door, you go on a slide, a seesaw, a swing, and you can choose to play on anything.
If your content isn’t where those slides and games are, I think you’re missing the most important aspect of the customer journey. Thought leadership plays a role everywhere, but so does marketing, and I think they go hand in hand. So I think you’re very right.
Liz: I totally agree. We’re human beings, and we know ourselves. We doom scroll, don’t we? I certainly do. On Instagram, you start off with something, you think, “Oh, that’s interesting,” and an hour later you’re thinking, “What on earth have I done for the last hour?”
We’re living in a world with an awful lot of signals coming in, and you need to play in lots of different spaces. We certainly do. Social media is really important for us. We think in terms of owned, earned and paid media, as a lot of people do: you own your website, you earn media attention, and sometimes you pay for things as well. But you need a mix.
Yogesh: Is there a balance?
Liz: I think there’s always a balance.
Yogesh: What’s the right balance that companies should keep in mind?
Liz: The difficulty is that it depends on the individual piece; each piece is different. If a piece has a strong policy angle, for example, we’d probably judge its success largely by owned media. We really want to be picked up by the EU, by policy directors and analysts, and to be quoted. If we’re doing something on corporate reporting—
Yogesh: Close to my heart. Very interesting.
Liz: —we would want to be picked up by CFOs, because they’re the main audience for it. We probably wouldn’t worry so much about being quoted in something like the FT, because corporate reporting is quite a practical subject, and we want to make sure practitioners are actually taking it on board.
Yogesh: True. When I was going through the report, one of the uncomfortable stats I found in Promise and Peril was around credibility. Producers ranked credibility second, but consumers of thought leadership, mostly C-suite executives, ranked it eighth: dead last.
That surprised me a lot. Are we chasing the wrong metric? Does credibility not matter enough? What’s your take?
Liz: I agree with you; I think credibility does matter, and in an interesting way. 80% of the thought leadership those same people consumed came from firms they already used. They already buy those firms’ services. So I think there’s a bit of a mismatch there.
There’s another stat in the same question that I think all thought leadership producers should take to heart: the difference in the importance of actionability. I think I’m right in saying producers ranked it something like seventh, fairly far down the scale, whereas consumers ranked it fourth.
Making sure you don’t just write an interesting paper, but also say what readers should do with it, is very important. I know it’s something we’ve focused on as a firm, and I think a lot of other people have too.
Yogesh: Do you measure credibility, or trust, at EY?
Liz: We measure our brand. We have something called the Brand Survey, which we do every two years, I think, and it’s a huge piece of work. We also have other programs, such as Voice of the Client, where we analyze all the conversations we have with clients and their feedback, and it all goes into one big pool.
So yes, we’re very focused on our brand, reputation and credibility. It’s number one. We’re a Big Four firm. If we’re not credible, that’s a major problem for us.
Yogesh: As you said, you’re a major firm. Let me segue back to the framework the GTLI has now defined: the maturity framework for thought leadership practitioners, which starts with ad hoc, then moves to emerging, then defined, then strategic.
I would say EY, Deloitte, the Big Four and many other firms in the GTLI’s top 15 sit in the strategic tier. What was the reason for building this framework? What was the intention behind having people measure themselves against it?
Liz: I think there are two things. Certainly, it’s for people to measure themselves against. Benchmarking is something we should all be interested in. But the other thing was to show that it’s a journey: you can move from an ad hoc position into the next phase. You don’t necessarily have to jump straight to strategic. You can say, “We know this is the next step on our journey towards becoming a strategic player.”
Like any maturity framework, it’s designed to give people a sense of where they are now and where they could go next.
Yogesh: We used this maturity framework in our Thought Leadership in Practice study, and we asked people where they would rate themselves. The result shocked us. Almost three-quarters of organizations don’t call themselves strategic. Only about 8% were ad hoc, and about 39% were emerging.
Does that surprise you? Should people strive to be strategic? And are we saying that organizations that are ad hoc, emerging or defined, because they’re also publishing thought leadership, aren’t producing great thought leadership? What’s your take?
Liz: No, I would never say that. It doesn’t mean an individual piece can’t be good work. However, if you’re investing as a firm and you believe thought leadership is an important part of your strategy, I think you’re likely to get much more back from it than if you suddenly say, “Let’s spend $50,000 on this particular piece of work,” and it just goes out on its own.
That’s going to be increasingly important because of GEO, because every piece of work is now judged by LLMs in the context of how much else surrounds it. So it’s a different universe we’re in. But no, I certainly wouldn’t argue that. In fact, I think with AI, a niche player can have a very strong influence, and I’m all for challenger brands.
Yogesh: Tell us more about that. Does it make you uncomfortable, as a brand like EY, that a niche player can challenge EY on a topic?
Liz: I love a challenge, Yogesh, so I’m never frightened of that. But it certainly keeps us all on our toes. One of the things the research showed is that none of us can sit there comfortably and say, “I’m a huge brand, so nobody can touch us.” The answer is that, first, our clients perhaps don’t care so much about the brand, and second, challenger brands in this new universe can find white space and occupy it.
One of my favorite sayings is that you can be a specialist guide in your village rather than a generic tourist. And I’m all for being a specialist guide.
Yogesh: Fantastic. And I think it’s equally important that emerging and defined thought leadership producers double down and do more of that. If I understand you correctly, they should use the right tools, but they should also be in a position to challenge leaders like EY, Deloitte and others by coming up with a topic, an idea or a hypothesis that could really shape the industry narrative. That’s so important, and it also challenges you to do better and smarter.
Liz: 100%. I’m a real believer in quality wherever it exists. As you say, it raises all boats, as long as people are doing really good things. We look at our competitors all the time, not to follow them, but to get ideas and see where things are moving. I learn a lot from our competitors.
Yogesh: Absolutely. When we were talking earlier, you mentioned that EY runs its own brand survey to find out how clients really digest thought leadership. Why did you start it, what’s the purpose behind it, and what have you found?
Liz: That’s led by our brand and marketing team, so it’s not something we in Insights spearhead. It isn’t aimed specifically at our thought leadership, although that is a big part of it. It’s also about the client experience and how we’re perceived in the market.
It’s absolutely critical for us to know that, and to know where our, I wouldn’t say weaknesses, but our opportunities for improvement are.
Yogesh: Have the questions around thought leadership surprised you? Have they met your KPIs? Is there any color you can share with us?
Liz: One thing for us is that the actual quality of what we do tends to be very highly rated by our clients, exceptionally highly rated, I would say, which is great. Where I think we can do better is in how we take those pieces of work to market: in other words, how our client-facing teams actually use those pieces of work and present them.
Some people call it the “last mile,” and I think some of our competitors might be cleverer at it than we are. Just putting something on your website isn’t great, and neither is sending someone an email saying, “We’ve just done this new piece of work. Here’s a link.”
There are more sophisticated ways to use that work to really generate deep, engaged conversations with our clients. That’s probably an area we’ll focus on more in the future.
Yogesh: Do you think thought leadership should take center stage in marketing strategy? Within a big marketing strategy, there’s content, distribution, demand gen and advertising. Where does thought leadership play a role? And how do we differentiate it from content marketing?
Liz: I think it plays a massive role, but they’re all important. I would say thought leadership is an expression of the collective cognitive abilities of your organization. That’s what it ought to be.
Yogesh: That’s a nice definition of thought leadership.
Liz: The brains of the organization should be visible in your thought leadership. If you’re really good at one thing, make sure your thought leadership reflects that. If you’ve got a star, as some of us do, in supply chain or something, make sure that person is involved in the thought leadership.
So those things are important. But other things are important too. If you’re playing at the big end and hoping to win $100 million engagements, advertising and all those other things play a part in the mix.
Yogesh: So true. Let’s talk a little about AI. I sometimes joke that AI is more about “actual intelligence” than artificial intelligence. One of the stats we’ve seen is that around 85% of executives use AI, yet 82% of them think AI is going to replace thought leadership, perhaps by 2027 or 2028.
What do you make of this dichotomy in the research? What’s your take on AI, and specifically on synthetic research? I don’t know how much EY is investing in or using synthetic research, but it would be interesting to get your opinion.
Liz: There are two parts to that question. Let’s start with synthetic data. EY has experimented with it, but I would say we’re at the cautious end of using synthetic data. Some of our experiments have thrown up real issues, and we’d want to be very careful about resting full hypotheses on synthetic data alone.
We tend to run parallel studies, so we have real people and we also have some synthetic data, and we compare and contrast. We work very closely with a company called Aaru, which many people will know. We’re learning a lot from them, and we’re certainly piloting a lot and will continue to do so.
But I wouldn’t say we’re in the vanguard of synthetic data use. We’re cautious, with good reason. On AI more generally, what we’ve seen, and it certainly came out in our research, is that clients are nervous about taking something they feel has been entirely machine-generated and making decisions on that basis, and they’re probably right to be.
However, they’re getting better and better at using AI tools themselves. Most of us are the same: if you’re in control of it, you tend to feel more confident about its use. If you’re not in control, because somebody else has been using it, the barriers come up and you’re a bit more worried.
So that’s what we’re seeing. We’re very much in an interim space. First, AI models are getting better and more intelligent. Second, people are getting far more used to using AI, and agentic AI, in their workflows, so they begin to trust it a bit more.
But so far, I have not seen a piece of purely machine-generated thought leadership that I think holds a candle to something with human fingerprints on it. That may change. But one of the main reasons I think it’s the case is that, by definition, anything an AI tool produces is backward-facing.
It’s looking in the rearview mirror. It can tell you a lot about the past, but it isn’t very good at predicting the future, and by its nature it isn’t very good at having innovative ideas. That’s where human intelligence can help.
Yogesh: It’s only as good as the data you feed into it.
Liz: And that data is backward-facing. Some of it may be perfectly good, but it may be two, three or four years old.
Yogesh: So help us resolve this irony. The data tells us that around 80% of executives want their teams and juniors to use AI, yet the same people wouldn’t want to read a report generated by AI. As thought leadership practitioners, how do we find that balance? What advice do you have for us?
Liz: I think there are two things. First, we need to be transparent, and actually do a good job of talking up where we’re using AI and why it’s in the reader’s interest. We need to be very clear about the advantages.
Second, AI can provide interactive tools and personalized journeys through a whole portfolio of thought leadership, which a lot of our clients want, and we have to make sure those are available. On curation, personalization and aggregation, AI is absolutely fantastic, and I think we all need to recognize that.
Yogesh: So how would you define the difference between AI-assisted and AI-generated? Would curation count as AI-assisted? And when you say AI-generated, correct me if I’m wrong, do you mean a whole piece of thought leadership generated by AI?
Liz: I think AI-assisted means you’ve used AI in some part of the thought leadership, but you haven’t literally given your AI tool a prompt and told it to write a 4,000-word report without reading it. I would hope none of us would do that. It’s not a good idea, and even though AI is getting better, that doesn’t change.
Yogesh: Please don’t do that.
Liz: Please don’t do it. It should come with a health warning. If you want a quick heads-up about something, that’s how I use it in my daily life, and I think it’s fantastic. If you just want to know the top ten hotels in Venice, say—
Yogesh: Or France.
Liz: Wherever. That’s all fine, because it’s not critical. But I always use the same example. Suppose somebody told you, Yogesh, “We can give you a broadly probable answer to what’s in your current account at the moment, based on two years of money coming into and going out of your account. We know that on this particular date in September, you normally have X in your current account.” If I said that’s what your bank would give you, I don’t think you’d be very happy.
Yogesh: No, because this is about money. I need to know exactly what I have and what I don’t have.
Liz: Exactly. And it doesn’t take into account the fact that you bought your wife a very expensive present because you’d had a row yesterday. There are always these individual things, and the same goes at a corporate level. People are making really big decisions, multimillion-dollar decisions. Probability isn’t good enough. You need as much certainty as possible, and I think the combination of AI and humans is how we can arrive at that.
Yogesh: A lot of people talk about humans over the loop, in the loop, above the loop, whatever it is.
Liz: I like “on the hook.”
Yogesh: On the hook?
Liz: I think human beings on the hook is what it should be, because in the end, you have to take responsibility for it.
Yogesh: Absolutely. So how do you use AI professionally at EY? From a thought leadership standpoint, where would you not let AI touch thought leadership at all? What are the lanes, and how do you stay within them?
Liz: As a firm, we’ve invested very heavily in AI. Our audit practice uses AI; we have a proprietary model, though don’t ask me about its capacity, because I’m not an auditor. But I know we’ve invested hugely in our tools, and it means we don’t need to sample to do audits. We can take whole data sets, which is much better.
In thought leadership, as I’ve said, we use AI in data analysis and data collection. We use it for white space analysis and hypotheses. We use it a lot. But it’s completely verboten to use it for final writing and output, and I think that’s right.
Partly that’s because I find generative AI perfectly clear but quite robotic. There’s something very dull about that kind of language. We want to engage our audiences, and we want to have a bit of personality. I don’t think that’s really possible with AI-generated content.
Yogesh: It’s not human. I’ve seen a lot of people use a “humanizer” skill or a “remove AI slop” skill, but I don’t think the result really ends up human.
Liz: You can get away from some of the tics GenAI tends to have, but that doesn’t mean you get great writing.
Yogesh: That’s true. I want to dive a little deeper and get your thoughts on ROI. It’s such an important topic for us as an industry, and measurement is very difficult. At iResearch, as you know, we define ROI by the four Rs: reputation, relationships, revenue and real-world impact.
How does EY look at ROI? We’ve spoken about credibility, trust and awareness a lot. Can you tell me how you approach ROI?
Liz: We have a really sophisticated group of people who look at our metrics, and again, it’s a mix of things. We’ll always look at reach, which is sometimes important, but not always. It depends. If we’re doing a very specialist piece on metals and mining, for example, the chances are we won’t get huge reach, because the number of people involved is small.
So reach isn’t the right metric for that piece of work. We look at engagement: things like scroll depth, and whether people click on things and engage with them. We might invite people to put their own data into a model and see what comes out. If people are engaging with the content, that’s usually a good sign.
We look at downloads, and we also look at media: how often a piece of work is cited by the tier-one media we really care about, or in earnings calls. Those are all measurements. Depending on the piece and what it’s trying to achieve, we weight those things differently. So we have quite a sophisticated model, and we score things against it.
Yogesh: Where do GEO and AEO sit? Are they a big part of ROI? How do EY studies show up in a zero-click world? What have you seen there?
Liz: I think they’re changing dramatically how we do things. We’re moving well beyond SEO and getting your tags right. How and where you show up is definitely changing. One example is very specialist titles with perhaps quite small audiences, which in the past probably wouldn’t have figured in our landscape.
Now we realize that GEO actually cares a lot about some of those very specialist titles, which means we’re more likely to be cited through them, so we care more about them ourselves. Where content is playing, how GEO is picking things up, and whether you’re cited as a source or as the main source: those things really matter now.
We’re learning all the time. We have a really good digital practice, and in fact one of our US colleagues spoke about this at the last symposium. It’s not something I’m a super expert in, but I certainly read the reports, and I find it really interesting.
Yogesh: One of the insights he shared still fascinates me. One of the goals he described wasn’t just about earned media in the big publications. It was also about industry specialization, like the mining and minerals example you mentioned.
So why not target focused publications that write about mining and minerals? AI may well pull mining and minerals data and insight from those, rather than from the Wall Street Journal or the New York Times. That opens up the opportunity to place your data across several different publications, rather than relying only on the bigger titles for earned media.
Liz: 100% true. It goes back to what I was saying about challenger brands and niches. There’s a real opportunity now to have a huge impact from a small base. If you’re really expert in a small field, the chances are you can now make a lot of noise in a way that was probably more difficult in the past.
Yogesh: Do you think AI has made measurement easier in some shape or form, through tools and technologies? What are your thoughts?
Liz: I agree. I think AI has made measurement easier in some ways.
Yogesh: But at the same time, AI has also generated so much content.
Liz: Exactly. You’ve got better tools, but you’ve also got a lot more data.
Yogesh: In terms of tools, are there specific ones you use or have access to when you produce thought leadership at EY?
Liz: We do lots of different things. We have the proprietary metrics I explained before. We use Conductor, and we also use tools like Hootsuite for social listening. So we have a whole range, as you might imagine.
Yogesh: Absolutely. Looking ahead, is the GTLI going to do a second version of its study next year? If so, what do you hope to find out? Can you share any thoughts?
Liz: I can. We are going to do a follow-up to the previous research. This time, we’re focusing just on clients. We may do some qualitative interviews with thought leadership producers to get their reactions to what we’re seeing.
It will be a slightly smaller study in terms of numbers, but again it will go to the C-suite. What we’re looking to see is whether people’s responses on AI are shifting, which we imagine they will be.
Yogesh: Is there any particular question or insight you’re excited about or looking forward to? I know you’re very closely associated with that study.
Liz: I tend to lead the research on it. I don’t know quite how that happened; I think it’s because I offered to write it. I don’t think there’s one particular question, but there’s the general question of whether our clients are still as conflicted about GenAI.
On the one hand, they’re saying they think everything will be AI-produced any minute now, and on the other, they’re saying they don’t trust it. I suspect that gap will be closing. But I haven’t seen the data yet, so I don’t know.
Yogesh: I want to understand a bit about how Liz uses AI personally. I can tell you what I do. As a tech optimist, like you, I’ve probably connected every single connector I could. I use Perplexity, and I’ve connected everything, just shy of my bank accounts, because my wife doesn’t want AI to know about our bank accounts. If it were up to me, I’d connect those as well.
I also use it a lot for health. But I’m curious: how do you personally use AI? Has it made a difference, or improved your life or your thinking in some way? I’d love to hear your thoughts.
Liz: I use it all the time. Recently, we were remodeling our house, because my kids are now growing up, thank goodness, so we can get rid of all the plastic toys and start to think about how we want to live. With ChatGPT, you can look at a sofa and say, “That’s quite nice. Let’s put that four-seater sofa in my current room. Now show it in five different colors. Actually, no, the three-seater looks better.” You can add real wallpaper and all kinds of things. In the past, that would have meant very clunky CAD modeling, which I was never good at anyway.
Now you can just write a few prompts, and the quality of what comes out is exceptionally good. Really, really good. You just use it all the time. And at work, it’s now really hard to remember life without it. I was first in the workplace before the internet, and my kids can’t believe it.
They say, “How did you know anything? No Google? What did you do?” And I do think back to what we did in the ’90s. It was so hard, and yet we did it, obviously. I think the same is beginning to be true of AI. How did we work before ChatGPT and Copilot and all these tools I use all the time? How did we do things?
Yogesh: I remember going to a museum with my eight-year-old son and showing him a rotary phone. He said, “Dad, were you alive when there were rotary phones?” I said, “Yes, absolutely.” And he asked, “What is this?” So many things were different back then. From black and white to AI, I think we’ve seen a huge change in the world.
Liz: No mobiles. “I couldn’t get to a phone.” It’s true, and my kids do exactly the same. They constantly ask me how we even operated, and I have to say, “Yes, I know it’s weird, isn’t it?”
Yogesh: One of the things we did at TLFT this year, Liz, was invite a few students from NYU, to expose them to this beautiful industry we’re part of, one you’ve been part of a lot longer than I have. For somebody who wants to join this industry, what’s one skill you believe they should start working on in 2027 if they’re considering a career in thought leadership?
Liz: I think data analysis is a critical skill, and using AI for data analysis. We would now expect anybody coming in to be able to do that and to understand it. But alongside curiosity, I also look for skepticism. I love people who automatically challenge what they see.
People who are mistrustful are always good for me, because—
Yogesh: Why do you say that?
Liz: Because we live in a world where things can look very glossy and appear very reasonable, and still having that human skepticism is very valuable. In the end, it means you do better thought leadership.
Yogesh: So, trust but verify.
Liz: Trust but verify. Exactly.
Yogesh: This is my last question for you, Liz. We’re starting something new in the Thought Leadership in Practice series: we ask each guest to leave a question for the next guest, without knowing who that will be. I don’t know what the question is, but we have one left for you.
The question is: how do you see thought leadership evolving? Will it still exist in five years? What rules will still apply?
Liz: Oh, that’s an easy question. It’s definitely evolving and changing. I’ll be honest: I think there will be smaller teams, with fewer people involved. Some of the skills that are really important will continue to be important. Being able to understand complexity, make it simple and articulate it is still quite a human skill, and people will continue to prize it.
But a lot of the things that used to be important, like large teams of people analyzing earnings calls, you don’t need anymore. You just press a button. We have to be realistic: an awful lot of analyst roles are simply disappearing. There’s just no reason for them, and I think that will accelerate without question.
So yes, it’s evolving. Will thought leadership still be valid? In my view, yes, it will, because people still want good ideas. For me, that’s what thought leadership is about: having really interesting ideas that can change the way the world works.
Yogesh: That’s a good way to end this conversation. I told you I don’t like five-year questions, but that was a beautifully answered one. Thank you very much, Liz. It was lovely talking to you.
Liz: You too. Thank you very much, Yogesh. Bye.
Resources: