The impact of AI on firms' thought leadership

Welcome to the Future of the Firm podcast. I'm Emma Carroll, Head of Client Voice here at Source, and I'm really pleased to welcome Elizabeth Bolshaw to the podcast. Liz is EY's global content strategist. Together, we're going to be exploring the impact of AI on firms' thought leadership.

So Liz, a really big welcome to the podcast, first of all, and it would be brilliant if you could introduce yourself and maybe a little bit about your career to date to our listeners today. Thanks so much, Emma. I'm really excited to be here. So yeah, I basically head up the editorial function within our global insights team.

And at EY, you know, we're a federation of many different firms, but we have a global hub that is responsible for most of our high impact and high investment thought leadership. And that's where I sit. I came here from an interesting route. I started in academic book publishing, very different kind of fields, very collegiate, quite slow, quite intellectually challenging.

And it taught me how to be, know a little bit about a lot of things. And I then segued into, I'd always done some writing and I segued into the FT. And I spent eight years doing a range of things, but a lot of interviewing and so the ft then had a program called women at the top which was all about female leaders and it ranked them every year class of 50 and i got to interview all of them which was an amazing experience absolutely wonderful experience um and anyway ey called me asking me if i'd like to join them and i i think i refused something like 17 times saying that i no way on earth would I ever work for an accountancy firm, which just shows you how you need to be very careful what you deny.

I'm very glad that I did, in fact, join EY. I joined as an analyst and then later came back into this role. And EY is a fantastic place to build a career. And I'm delighted to be here now.

And we've split our conversation today into four really big areas around AI and thought leadership. So just to run through those for our listeners, they know what's coming. So we've got democratisation of information. We've got evolving client expectations, how content is discovered by users and also trust.

So the first of those democratisation of information. I'd love to get your take on that, Liz. What does it really mean for thought leadership production now? You see, the first thing is that the barriers to entry and thought leadership have lowered very dramatically.

largely large language models have enabled non-specialists to create really very good plausible reports in a matter of minutes. Not only is the writing quite good, the structuring quite good, but with a decent prompt you can also explore some synthetic data to underpin an argument which is quick and relatively cheap to do. Your data visualisation if you're using for example anthropics claude can be done again to a very professional standard and all of that is available to anybody um and that's a massive change and of course what that's led to is a huge increase in content volume and we've seen that over the last 18 months to two years um i think it's i think i'm right in saying that in december 2024 Four, machine-generated content online exceeded the amount of human-generated content.

So this huge tsunami of content, some of it quite good, some of it not, has really changed the game. It's made, you know, slop was the word of the year, I think, both for The Economist and Merriam-Webster last year. And headlines on slop continue to evolve. I mean, it's extraordinary how much people are using slop as a kind of AI backlash.

You know, there's a lot of feeling now against it because some of it is, of course, low quality. But that's not true of all of it. And LLMs themselves revolve so dramatically from, you know, when we were first looking at Jatchi BT to now, we've seen an immense and very accelerated professionalisation of some of those tools. And they're much more sophisticated than they were.

But really interesting you pull out that word slop there. Is there any other way you would say it's like changed the competitive landscape, particularly maybe in professional services? I think it definitely has, because I think one of the things now you need to do is you need to be constantly thinking about how your own content will stand out against this background. It's a new context in which we're all working.

So I think things that have always been very good to do, like collaborations with academic institutions, collaborations sometimes with clients and certainly with alliance partners, have helped us, I think, elevate our quality, which is one of the things you've got to do all the time. And to ensure that our clients are seeing something that is both differentiated, but also credible. okay and would you say that anyways ai is actually raising the standard of thought leadership in firms i think without doubt it's raised the floor so if we look even at sources own quality rankings we can see a convergence you know there used to be a long tail of material that was really poorly structured sometimes even misspelled i mean that you know that really there's no excuse for that anymore and I think most producers of thought leadership are using tools to ensure that there is a level of competency across everything.

What I think so far it hasn't done is really raised the ceiling. I think what we're not seeing is the very best pieces get better and I think that's a great challenge for all of us is how do we use AI tools not just to create a sort of minimum viable product but actually to you know hit it out of the ballpark And I think that the great challenge for all of us working in this field Yeah And thanks for bringing up Sources report there For those of you who haven read it yet it our Quality Ratings of Thought Leadership report which we produce every year And this year, we really did a deep dive into the kind of idea of a sea of sameness of all of this thought leadership.

And Liz, I'd love to throw that image back to you. And do you think AI is contributing to that? of course it is because you know the fact is that all of us are using similar tools the same tools um we are we're all looking to um i think we're all looking to explore similar challenges big business challenges and many of us have similar client bases so the question is you know how do you distinguish yourself from that sea of sameness competitive landscape particularly at the top end And when I'm thinking about this and thinking about innovation, I always love to think about Proust's great quote, which was that the real voyage of discovery is not in seeking new lands, but in having new eyes.

And I think, you know, that's a great thing to think about. I mean, if you look at a business example of this, the great US firm Corning that introduced heat resistant glassware, which we all know as Pyrex. You know, all our grandparents' covers were full of Pyrex and they still produce Pyrex. They have evolved really cleverly to produce the screens for smartphones and now are gaining enormous traction in the high density cabling needed for data centres.

so they they're sticking with their knitting they they know they know this particular field very well but they're applying it to a completely new context a context that in 1915 when parox was first invented could never have been foreseen and i think innovation even when it comes to thought leadership is about that it's about knowing where you stand and where your particular expertise is and applying it into this new context a context that's moving very fast as we all know and is increasingly volatile.

So, you know, you're looking at something like supply chain, and you're just not taking the easy option. You're really thinking carefully about what your USP is and what new view you can have. Exactly, exactly. New eyes, not new topics.

Yeah, lovely. Thank you. And our second topic was evolving client expectations. So I'm really wondering, what is it the clients want different from thought leadership now?

interestingly i i i represent ey on the global thought leadership institute which is a relatively new not-for-profit we we are attached to the apqc in the us um but it does include many of the big firms that are involved in this and we decided that we would do some research last year with about a thousand cxos across i think about 20 countries and similar number of different industries and we were exploring exactly what client perceptions of AI and thought leadership were and what we found was intense cynicism so 85 percent of respondents said that they would take a dimmer view of a piece or indeed a firm that was actively using AI in its thought leadership you know 69 percent had a negative view of LLMs in anything 95 percent didn't like the idea of synthetic data.

So the way I see this is, I think that is a response to some of the slop that is out there, not necessarily the best use of AI. And I think as professional services firms, we have a responsibility to explain the advantages of using our AI tools. You know, why have we used them? If it's simply to save money, then that is not a good answer.

But if actually we can show that the use of AI has really improved and increased the sophistication of the eventual output, then I think we've got a good story to tell. What we did also find, however, is that many professional services firms now on their websites are offering a chatbot that allows you to explore a topic through a number of different reports. So you don't have to read one report. You can say, okay, I want to know about the impact of volatile oil prices on, I don't know, plastics.

And if that very niche topic appears in 15 different places, that will be summarized for you. So curation, discovery, really finding your own way through things, I think is something that AI is really allowing us to do. And certainly clients are embracing that with open arms. um internally i think we're also able now to um for example have a repository of all the data from all the studies we do and again offer that to internal staff for ways of looking at something so providing you know really bespoke advice for clients that doesn't just look at one through one lens of one report but can actually wander through all the work that has been done that i think those kinds of curation tools um are really really valuable um and we are all using them all the time in our daily lives in fact yeah it's so true and i was really interested when you were talking about the chatbots as well because it enables people to pull stuff together but they're pulling it together from their trusted advisor not from everything across the internet yes exactly and i think that that that question of trust and credibility is is i mean i know we're going to talk about it a little later, but I think it is just such an important part of this field though.

It's always been important, but in the past, I think you could rest on your brand. You know, you could say, well, you know, here we are, we're a big four auditor, trust us. That's less and less viable now in the new context. And we have to kind of really prove our worth at every time, every milestone.

Obviously, there's just so much going on at the moment. There is just so much volatility. So how important is speed of production now? It a really interesting question And my own view and it not necessarily the firm view is that we have to get much better at operating at two speeds So we have to be able to do shallow and fast respond really quickly to something that has happened that clients expect us to have a point of view on, and also deep and slow.

So one of the problems I think that AI is potentially doing is a kind of cognitive atrophy where you stop thinking because things are easy and quick. So if we look back to when the internet really became part of our lives, there was a great and very influential article titled, Is Google Making Us Stupid? Which was really bemoaning the fact that people no longer had to memorise information and recall it, because Google obviously is the end of your fingertips. You can do it all the time.

I think now the question of cognitive atrophy is what AI is raising as a potential danger, which is that you stop critically thinking because AI can do so much of it for you. And all I would say is I don't want to be using the cliche of human in the loop, but the fact is that thought leadership is a contact sport. You need to engage with it in a very active way. And I think it's very, very important that any of us involved in this continue to use our critical faculties because that's what clients expect.

Thank you. And discovery was where I wanted to head next. And SEO, it was always seen as really essential to getting our content across online and discovered. But now we've also got generative engine optimization.

so how are you going about balancing those two and thinking about the combination there yeah so so generative engine optimization is already really changing the way that your content is discovered um there are several things that i think are important in this you know the first thing is the question of gated content so many particularly marketing departments are very keen to gate certain content because obviously you can collect subscriber information and many much of that is useful.

However, gated content is invisible to search engines. And in terms of Geo, increasingly, I think there is an argument to say that all of your content needs to be open and available and readable. One of the things Geo is doing is making us think carefully about LLMs as readers, not just producers. I mean, all the effort, all the sort of talk has been about NLMs producing and writing and doing all of those things, creating content.

But actually, NLMs are also our audiences now, and it's a very important audience. So we have to think carefully about how they operate. And there are a number of things that I think actually play very well to quality. One of them is the number of sources and citations.

How often is that piece mentioned by another credible source? How many links in and out have you got in your content? What's the context in which that piece sits? So are there other pieces of a similar quality surrounding it on your site?

Is it obvious that you're more of an expert in supply chain because you've got, you know, seven really good pieces there, or is it just on its own? Some of these things will build into how much, how fast you're ranked and how quickly people can find you. So it is changing the way that we think about content. I would say in a positive way, that it's certainly impacting discovery.

And we're seeing, you know, I think across the board, everybody is seeing a lower click through rate, because it is certainly harder. So that business of competition that we talked about earlier about competition coming from volume is also coming in terms of discoverability. So interesting and I was really intrigued when you were talking about links in and I'm wondering how you're having to go about thinking about links in and whether there's any extra work streams in you know getting people to put links into your work.

One of the things we've noticed is things that we wouldn't have worried about in the past so if you think about earned media which is certainly important to us and it is to most professional services had we would have worried about tier one media you know general business financial times all of those wonderful places but now also very specialist titles that may have quite small niche audiences are also very important because they are rated highly by geos so you know if you're quoted in you know sheep shearing weekly that may be an important thing now in the way in a way that it wasn't in the past so I think one of the things that we're all learning is the advantage of being an expert in a niche is increasingly vital and not being a generalist because it's in broad generalist terms that LLMs can do a job probably as well as anybody but it's where you need to go narrow and deep that actually you need that expertise that is beyond what is generally available in summaries.

And we talked a little bit about trust earlier, but it's such an important topic. And I'd love to think about clients first. And is there anything we haven't talked about, about the concerns that clients have around trusting AI generated content? I think the most important thing is transparency.

I think both being open about how you've used AI and also explaining why you've used it, how it's enhanced what you're doing. And I think that's vital. I think also there was a time when you could rest on a legacy brand, the fact that you're trusted in the marketplace. I think that's becoming less true.

So I think increasingly, partly because of the way that content is discovered, it's less important that it's come from you, but more important what the actual quality of it is. And that's a great opportunity for challenger brands. But it's also a challenge for those of us who do have some brand equity in the market that we have to continually, I think, continually prove that we're worthy of that brand and that trust. Fact checking internally critical using LLM detection tools the importance of having human in the loop importance of having I think real world examples you know being able to cite client work being able to show how an idea is proved in practice.

Those things are increasingly important. And what are you specifically doing at EY to build trust around, you know, your use of AI? well we're certainly strengthening our governance we have very strong governance at the global level um so our global thought leadership has had quite strong guardrails around it for a while now but we're also um looking to deploy those more actively across you know our 500 plus firms across different countries um i think the other thing is ensuring that you know just double checking at every level, your sources and your, there are no hallucinations, there are no, you know, sources that don't exist.

I mean, many of them, and most of us have had examples of this, sound very plausible. But if you don't check it, you won't know. So I think that old fashioned, very old fashioned way of fact checking is, again, absolutely critical. Oh, thanks, Liz.

It's been really interesting talking through those, you know, four really important factors and I'd love to turn the lens inwards for the final part of the podcast and talk more about EY and EY's use of AI in thought leadership production so what are you most excited about that you're doing at EY there? So well I think there's lots of different things we've evolved about I think we started about 18 months ago so we've evolved a proprietary tool that web crawls for white space, does it in a very user-friendly way.

So it uses an LM to do it. So you can just use natural language as questioning. It evaluates new proposals. It will make suggestions for improving them.

It will critique a hypothesis. And we've also built in our own quality metrics, which are very similar to sources quality metrics, but just marginally different. And you can put in a draft or a final draft into the model, And it will tell you how you've scored against those important quality things. And that's, again, a wonderful way of improving the output at every stage.

So we're certainly excited and we use that increasingly, I think, across everything we do. It's sharpened up our research plans, it's sharpened up our proposals, and hopefully it's polished the final drafts. So those are really good. And I think the other thing that for us is exciting is using tools like AlphaSense to really look at a range of almost real time news across 350,000 sources.

So you can you can go very quickly deep into something and get an excellent summary of all the things that have come in overnight even. and then create a deck out of it. I mean, there are all kinds of outputs that, again, are very fast to do. So I think now we have an opportunity to be much more sophisticated in terms of the analytical methods that we deploy, the different data sets that we might look at, things that are in the public domain.

You know, one of my favourite things is LinkedIn's job titles, data that they provide. You can really see shifts happening in real time as new job titles become more important and there are more of them. So there are all ways of looking at what's happening in social media, what's happening in niche expert areas, sub stacks. You know, if you think about the kind of universe of expertise that is out there, our ability to interrogate that universe is massively improved through AI tools.

You've definitely enthused me there, so thank you. What would you say is your biggest learning in terms of the use of AI? i think the main thing is as we've said before it's ensuring that we don't all fall into this sea of sameness and i think that is a a real danger um and i think for smaller firms um it's all about finding a niche and exploring it build a reputation in a specific area be the best in that area. Don't be a tourist in everybody else's field, but be an absolute expert guide in your hometown.

And I think that's a really, that will always win, I think. Okay, that's a brilliant piece of advice. I'm going to ask you for a final one on our final question. So I'm wondering about thought leadership teams who are less evolved on the AI journey.

Is there any kind of unexpected piece of advice you'd offer them? Well, going back to the GTLI research again, we asked about challenges that thought leadership producers had. And smaller producers all came through with lack of budget, lack of people, lack of everything, but all really resource-related. And what I would say is AI tools can really help you get over that.

You need fewer bodies to do a very good job. You can really deploy AI to give you the very similar advantages to much, much bigger organisations now. And I would encourage any small firm to really look at AI tools from the perspective of doing things fast, doing things fairly cheaply and actually being more agile than perhaps some of those really large firms, such as EY. Liz Bolshaw, thank you so much for joining us on Future of the Firm today.

I really appreciate it. Thank you so much, Emma. I love this conversation. if you found today's discussion interesting you can find more episodes on Spotify Apple Podcasts or anywhere else you get your podcasts to find out more about how we're helping shape the firms of the future head to sourceglobalresearch.