Subscribe
Share
Share
Embed
The Margin is a podcast from MGI Research that explores the evolving world of business monetization. Hosted by MGI Managing Directors Andrew Dailey and Igor Stenmark, the show features candid conversations with founders, CEOs, product leaders, and industry experts at the forefront of pricing, billing, and revenue operations. Each episode dives deep into the strategies, technologies, and trends shaping how companies generate, capture, and grow revenue—from subscription and usage-based models to AI-driven monetization. Whether you're in finance, product, or IT, The Margin offers practical insights to help you navigate complexity and drive growth in the digital economy.
Andrew Dailey: Welcome to The Margin, a podcast exploring the forces shaping business monetization. I'm Andrew Dailey, Managing Director and Analyst at MGI Research. Is artificial intelligence an existential threat to professional services or a once-in-a-generation revenue opportunity for industries that are human capital intensive, namely professional services? AI threatens to commoditize and even eliminate many businesses, and at the same time, AI is a powerful productivity tool, coming at a time when services industries are struggling to attract and retain key talent. To help chart a course through this kind of fog of hype and fear, I'm joined today by Dan Brown. Dan brings decades of experience leading product and strategy teams at companies like Microsoft, where he was general manager of corporate strategy and development, and more recently at Certinia, a leading services as a business software company, selling into some of the largest professional services businesses in the world with deep expertise in professional services, automation and experience working with some of the large services companies. Dan is here to share his perspective on AI and where Gen AI fits into professional services and where it doesn't. Dan, welcome to The Margin. Dan, we're really excited to have you because you're really in a unique position in that as a business, you get to see inside of a lot of different services, organizations, from Big 5 to some of the largest companies that have service components to their businesses. And equally, you're running a software and services business as well. Right. And presumably, getting the same, pressure from the board and investors about what are you doing about AI? Right. How are you cutting costs? How can we replace people? So let's open it up with a fairly broad question. Which would be is generative AI, existential threat to service businesses or is it an accelerant to the business, or is it some crazy mix of both?
Dan Brown: Yeah, it's a great question. And I think it's it's a it's probably a little bit of both and maybe the existential threat is not manifesting itself in the way you might think. I think when we talk to our customers and we see this internally with our board and our own operations, there's just this massive amount of pressure to do something with AI. And it's a little reminiscent of digital transformation a la 10, 15 years ago when that started, to really become the parlance of what a company needed to do to be better. But it's way less clear. And so you have a lot of professionals that are going, “I know I need to do something with this. I just don't know what.” And what's the existential threat in the near-term of that is taking your eye off the ball in the sense that while there's still a whole lot of stuff that you could be doing that is descriptive or even just mission critical transactional software that can make your business tremendously more operationally efficient or better deliver your services or whatever. So I think that's just one thing to keep in mind. And I think there's so much pressure in software and professional services companies to do more with less. Gen AI is sort of this. It's like a peanut butter. You should be able to apply it everywhere. So I think that's kind of part one. Part two is much, much like the rest of the technology surface area. And I'm going to throw Gen AI and sort of more if you can even call it traditional machine learning. There's so much technology to that companies, any kind of company, it doesn't really matter if you're bricks and mortar or if you're a knowledge worker company or whatever. There's so much technology required to be competitive and to be good, but I actually think the professional services organizations are growing right now. I mean, if you think about—like we look at companies and embedded PSO or otherwise, and they have hundreds of pieces of software that they're putting together to do their job. And they it's a really difficult effort to make those things work together. So what do they do? They naturally call on professional services company. So we see a lot of high growth professional services companies that are pulling together technology that can either be at the at the stack level or at application level or both. And they're really like they're high growth. So that's in the near term. And it's not necessarily a Gen AI, although I think it's being accelerated by Gen AI the opportunity for Pro Serv companies is really growing. But I think the last one is and I’m sorry to be long winded here. I think the real question is, is there a future where Gen AI is doing some of the work that a professional services person does on a regular basis? Certainly, Gen AI is going to help with a lot of the more mundane things. The way I think about this is when we automated no transaction system of record business process, you started looking toward knowledge work. And so Gen AI gives you a way of automating not process work, but project or knowledge work, status creation, QBRs, mining unstructured information and injecting prompts. All of those things are going to help knowledge workers stop doing the mundane. Will Gen AI get to the point where I can mine tasks and projects and use that to do more of the work? There are a lot of professional services firms that believe that's the case. I'll tell you the thing that we need to understand is when machine learning has this concept of the stable world principle, which is if the past looks like the future, the predictive power of machine learning is good in that environment. That's a broad statement and there are lot flaws with that statement. But if you take that as, as a principle, Gen AI is still built on that as good as it is. It's still based on a history corpus of language. And if that corpus doesn't include certain concepts, for example, we changed our name a year ago. If you're using ChatGPT 3.5, it doesn't know anything about what Certinia is. And so the question is, if you have new work that isn't based on the past, isn't based on a stable world, it's hard to see how Gen AI in its current form would actually do things. And human beings are really good at that. We can generalize very, very quickly. We can think in counterfactuals. We can ask five whys. That's not something you see today in an AI platform. That was my short answer, Andrew.
IS: So, it sounds like professional services are not dead.
DB: I don't think they're dead.
IS: Gen AI isn't going to kill professional services at the core. And we just did a we just did some work earlier. In the year where we did some survey field work, we went out to a lot of the large professional services firms and asked them, how do you guys sell what tools, what tools they use to sell and and who sells? Who is responsible for selling? Is it your sales force which you have? Is it the partner in charge of each practice and it's remarkable how I would say parochial approaches in most very large and very successful companies where there’s minimal, if any automation, a tool of choice is always excel there. They do need to get a quote to a customer quickly. But if it's something small 50, $250,000 project, yes, we can do it and excel if it's something more complex we take for a time, but it's remarkable how we professional services areas struggle. So its own cobbler's children syndrome, and how little adoption and how little – we were kind of shocked by how little power do sales organizations really have within professional services? It is really a star. Still very much a star culture.
DB: Yep.
IS: They're partnerships together, collectively sit down at dinner and decide what we're going to spend. Oh, and we can tell above sales, and this is what we want.
DB: Yeah.
IS: And give it to us. Our parties have really very little say. So it sounds like it's not going to kill it. But there's some mundane stuff that can be really addressed pretty well by a generative AI. And we're going to probably improve automation. Where do you think Gen AI has likely started hinting at that? Where can it fall short in terms of its ability to impact professional services as a business and also the delivery of IT, automation of it and so on?
DB: Yeah, I think that we—and I like to speak about this in, in a broader, broader context of, of machine learning and AI because I think they have we, we, we seem to have forgotten that machine learning is a very powerful tool and are focusing on when the the qualification applies to both. And that is there. There is a risk of just taking it as given that the response is. So like if you're using AI or as, as a way of mining information. So you talked about sales using information, say, in your CRM in an opportunity and prompting the unstructured information to help you summarize. This is stuff we do, by the way, with our our tool. And using quotation templates but similarly using information that is unstructured in our PSA tool to create status is and so on. So those are kind of the two main scenarios for gen AI. And you can speak about prediction and categorization for machine learning. The challenge is do you take it at face value that it is accurate? And you can prompt your way to, to, to greater, accuracy with Gen AI, but—
IS: Probably, probably a lot. I mean, imagine you're getting an audit opinion from an AI bot. Are you going to submit that to a shareholder to ask the regulators and say I passed the audit?
DB: That's right. So yeah, I mean, you're hitting the nail on the head like there are compliance reasons. There's quality consideration. In some respects I think there's quite a bit of danger in just taking it as it is. And what I don't think that if you're a client partner at a fortune—a at a large GSI or an embedded PSO at a fortune 500 company, and you've been in this industry for an extensive period of time, I don't think you're going to do that. But if you're an early-in-career professional that hasn't been around professional services and you're doing and you're you might take things at face value that you really shouldn't. And so one wonders what happens in five years, how how the interchange between technology and folks that are building up with who are growing their careers with gen AI, how does that work? I think that'll be really interesting. How attuned are they to what is real and what is not?
IS: So, it sounds like things that require an opinion, even opinion and negotiation, I have a hard time imagining that you could sell to a machine or a buy from machine. Right? Similarly, if you are, let's say you are in financial services and a company comes to you and says we have a deal on the table, we need a fairness opinion, it'd be difficult to pay half $1 billion for fairness opinion written by a machine. I mean, they can make an assist, they can get a lot of data, lying data, but ultimately a human I guess has to write that opinion piece.
DB: Well. And I think there's–I agree with you and I think there's a really important point. So there are types of consulting that I think can be disrupted by Gen AI, that we'll probably see that happen. And that is—you've kind of hinted at that. So let's say I want to go in and I want to assess the value of a company. Traditionally, a lot of that work is retrospective. I'm going to do a survey of of all the companies that exist, I'm going to look at a bunch of economic data. When you think about that, as long as you have a generative AI tool that has consumed relatively recent data, you probably can get a lot of that done. In fact, there's actually there's some really good discussions on doing generative forecasting with economic data. Like there is a large language model is large because it's consuming a corpus of data across the web and the internet. If you can get that broad of economic data, one wonders if you couldn't start doing forecasting and evaluating things economically. Really broadly, like there is a corpus of data that might be used in that way. So when you look at—if you're a firm that evaluates companies and your point of view is retrospective, I think it could be disruptive. You could be. However, it's the prospective—it's the scenario planning what could be how do all the players in prospectively operate and how does this asset, how is it going to live or die in that moving marketplace? That I think is hard to do generatively. It requires scenario planning and requires counterfactual thinking and I think those aren’t done easily.
IS: Can models get to that point? I mean, but one of the questions I got in preparation for this event today, one of our clients sent me a really funny short note saying, “are you are you including industry analyst firms in your definition of services?” Which one will be killed off as a result of Gen AI?
DB: Exactly.
IS: Which I've said probably many realistically but yeah. We don't know which ones but probably quite quite a few.
DB: Yeah I think we're all sort of all sort of wondering when does the Hari Seldon from the Foundation series, like when, when does the, the psychohistory absolute forecast of humanity, when is that going to be in the machine? I think it's safe to say that's a ways away. When you look at most companies and their ability to consume technology, to your point, back to one of your original points, I think it's low. That doesn't mean that that there aren't going to be disruptions coming where technology is—there was I believe this article is in The Economist. When is there going to be the billion-dollar startup with one person that use technology writ large, presumably lots of gen AI to create a company. Maybe then you know we're in a very, very different world.
IS: Some Nvidia guy is talking about it. He says, “you know, we'll we'll build, have AI, build ships, will build the AI, will build the factories, would build the ships and build the tools that build the ships in the factory.” Like his vision is really infinite. The question is what kind of torpedoes that vision. Is it power, is it regulation, is it environmental with a bunch of stuff?
AD: And so that's another thought. When the cloud came out, we went around and interviewed dozens of companies, from data center businesses and others and said, what could stop this train? Right. And this was before everybody really saw the power of of cloud computing. And basically the answer was, wow, there could be some massive hacks. There could be a fundamental breach of or loss of trust, right? And everyone will say, oh, we've got to rethink back to our own data centers. Well, we've had some major hacks, a lot more than have reached the headlines, in fact. And yet here we are pushing—
IS: Yeah, you could have shortcomings. Remember, we talked about it where people said, “oh, we will be shortcomings in security and this and that,” and a lot of people in financial services were poo-pooing the cloud. And guess where we are today? And I think we at one point we had our own kind of epiphany where we said, “look, people will go to a cloud regardless of the shortcomings. It's going to happen. Economics are going to be just too compelling. Simplicity, ease of use, agility that they can gain will outweigh any other risks.
AD: Yeah. What do you think could fundamentally derail it? Or is it too late?
DB: I think—this is not a statement about derailing. It's a statement about momentum. You if you go back on the history of artificial intelligence, it's had these fits and starts. There was sort of perceptrons and the first checkers automation back in—before the 70s, then you had more logic and, and expert systems and that didn't quite work. Then you had the first wave of connectionism and neural networks and that sort of stopped. And then roughly 10 to 15 years ago, you had this blossoming of of deep learning, a reinvention of backpropagation and so on. And now you have this Gen AI. So you've had its ups and downs. Eric Siegel just wrote a book recently, and he’s sort of warned that we're at this Gen AI top of the hype curve. I do think there's going to be some deflation of because the expectation be more and more to be honest with you. And I know this is sort of not like the most interesting statement. What is what is gen AI becoming? It's becoming a really sophisticated NLP tool. That's great. That's productive, but that's not redefining. Now, I'm not saying there isn't something redefining coming, but in its current rendition, it's hard to see that. So I think there will be some deflation based on the expectation. So that's, that's that's number one. I also think, we sort of chatted about this a little bit before, one wonders how people are going to make money over the billions that they're spending on Nvidia chips, training up—what is it that Microsoft said it took a couple hundred million dollars to train up Chat GPT 3.5. Is that sustainable? Companies like Google, Amazon, Meta, and Microsoft are really shifting massive amounts of R&D spend on AI. At some point they need to monetize that immensely. Now, the market capitalization of these companies is off the charts. So almost that pays for itself, but that's not sustainable. You're going to have to come up with some EBITDA and real cash flow that is associated with this. When does that happen? A better search is good. There's a lot of money in a better search. But do you need to have something more? How is Meta going to justify that? So I like to me that's going to put some pressure on this after that. But one wonders whether or not compliance is going to be the thing that puts the brakes on and I don't know, I think bureaucracy of politics may not be fast enough to control it, but there are certainly many in the AI community who are calling for this. I mean that's happened for almost ten years now that the community wants restrictions on what were true generative AI, not just textual or general artificial intelligence putting the brakes on it.
AD: You have you have this conundrum to where organizations are saying, we've got to get everybody in the organization to understand and embrace and learn how to use generative AI tools. And at the same time, companies, rightly or wrongly, are saying any use of any outside Gen AI is absolutely prohibited.
IS: Yeah, we've seen this already. We've seen security—
AD: The internet is going to take over the world, and no one can use the internet.
IS: Nobody can use a website. But that's 30 years ago. That's exactly what was going on. They would go to appointments with clients and people would say we don't have any websites in the company. And we would say, really? Let’s open it up, let's see it in the meeting and it was very embarrassing.
AD: But there's this bigger problem, which is that the absorption, the human absorption of technology and organizational absorption, right. And the change management that goes with it. And where do you see from what you're looking at in terms of professional services and services organizations where where are the areas where the absorption level is high and where there's zero? Where is it going to really soak in and make a difference, and where is it going to roll over the top and not make any difference?
IS: Services companies can benefit from this probably meaningfully. We don't know if the lob is going to be distributed evenly between the small companies and large companies that provide services like the Accentures of the world and PwC. They could benefit disproportionately. But there's also a sense that some of the smaller companies can see this as an opportunity to compete more effectively, kind of have a more level playing field, because we all have access to the same tools that the big boys will have access to.
DB: I think that the great thing about the current state of generative AI is that it actually is pretty simple provided you're plugging into a platform, a large language model and a GPT platform to plug it in. So we use the Einstein GPT and Copilot. We've worked with Salesforce quite extensively with—you can kind of think about this as layers. So there's an interface and a trust layer that Einstein GPT provides for you. You can plug in your own versions of your large language model and your GPT. So you can use OpenAI. You could use something that is, that is, that is, private like a Molly from Databricks. So once you do that, it's actually quite easy to prompt away and again, prompting on how do I extract information that is sitting out on the customer engagement platform, which is constituted of our products and Salesforce products and whatever other products you have on their platform, or using the information that is there to create information like a QBR, or status, or an escalation report, or whatever. It's actually pretty straightforward. Now, what we've found is once we build up these prompts, we can package them. So we can package these prompts for, “hey, here is the summary from a customer success perspective. Here is a summary that is a project relay.” So it's quite democratized.
IS: And so it’s like a programing language basically.
DB: That's right. Yeah it is. And it's metadata driven. So it's high level in the end. You don't have to be a Python coder or to go do this stuff and we're trying to make that easier. And I think that every company can make use of it now. It still has the dangers that the AI has in, in terms of is it is it is accurate is you expect like we're trained to assume when I typed numbers into a calculator and they come back, they're perfect. There's never going to be an error. Well that's not–you need to set your expectations differently.
IS: Predictability is an issue. Yeah.
AD: Dan, you just touched on another interesting question which is where is the money going to come from. Yes. Right. Like you're getting in your role. You're getting pressure to do more. Yeah I'm guessing that your budget didn't just get increased 30%.
DB: That's right. I think that's the tricky thing. And I think it goes from the platform providers all the way up of how do you monetize and is the value proposition high enough? Knowledge workers are expensive. And so if you can save 10% of a knowledge worker’s time so that they can do higher value work like cranking out a status report is a very low value add activities. Yeah. So how you monetize all the way up the stack obviously in Pro Serv, get the stuff that when you go any new itemize your billing, it says “we're charging you $500 to generate a status report.” I'm being facetious. Obviously. We're advising you about where the where the market is going. People are willing to pay high values for those versus the mundane. And I so I think in a Pro Serv organization, that's ultimately how you monetize. It's just going up the value stack. Although, Andrew, this is pretty interesting. We did have a conversation with a customer that said we're preparing for a future where we have digital professional service agents that we're going to charge for. And when I heard that, I was like, “wow, that's pretty cool. But you do realize that now you're a software company?” But anyway, it was an interesting conversation.
AD: I think in the short-term services organizations, particularly the sides, are incredibly adept at extracting value from the market. Oh, yeah. There's a short-term window where they're going to take advantage of the tools and take all the margin.
DB: Absolutely, right.
AD: And the question is when does it get commoditized? And that and that margin of opportunity just goes to zero.
DB: The management consulting world is one that I watch closely. When I was at Microsoft in the strategy team, I spent about 12 months there, before coming to certain year. And you had a lot of ex-management consulting people in there. And the thing that they were incredibly adept at was taking a problem and cranking it through the lens of strategy. And so with very little information, they would scrub it. And then, of course, you get these beautiful presentations and graphs all going up into the right. And you were like, “wow, that's amazing.” And then you the secret sauce is they've just got this this, this like inventory of content to go use. One wonders how much of that and, as you all know, it's really high price consulting. So one wonders how long they can survive at the prices that they currently do. The Economist talked about McKinsey letting go of people. I think that was a couple months ago, how they got overextended post-Covid. I don't think that's done. I think there's some exposure in those types of professional services organizations. They find stuff that is that is higher value add.
IS: I think I think it goes back to this issue of like leveling of a playing field. And we have one question here from the audience. One question describes, “So will large size and BPO providers being able to differentiate or will Gen AI become the next ML automation of the past decade.”
DB: Yeah. Yeah.
IS: Hopefully doesn't become the next blockchain.
DB: We'll see. Yeah, exactly. I think it's a great question. The signal that I see and I keep coming back to–it's not answering this question exactly, but every single company is being bombarded with outrageous expectations for predictive improvement and and EBITDA. Part of that is this sort of strange economic climate that we're in where we that there's technology growth sector is off the charts, but everybody else seems to be under sort of just borderline recessionary pressure. So there's this vice. And every company has tons of tools. These tools, we're one of them, hopefully, they're cranking out new stuff all the time. Which means that people and what company is up to speed on all the technology that they have at their fingertips? Zero. It’s just impossible.
IS: No one, yeah.
DB: And so now you're asking all these companies to use what they have, which they're not, to use what's coming, which they can't, to add in a whole different technology, which people are unfamiliar with and to be able to justify what kind of productivity gains, who are you going to turn to help you with that? I think you turn to professional services to help you put it all together. And by the way, you keep your current job and business going. I think it's really hard.
IS: Yeah. So from that perspective, professional services firms will benefit, there's no question about it. Because by the time you kind of look at this, if you were in a business side of things, you say, okay, I have to do all the things you just outlined and I have no time. My board is breathing down my neck because the competition is at the door. Activists are knocking at my window every day. Want me to get the stock price higher? What do I do? I engage a well-known professional services firm and they take care of it. It covers my behind and maybe we'll get something done. If I give it to an internal team, it could take a year or two to start up and then uncertain results. So from that standpoint yeah it is guys can benefit benefit quality.
DB: And I think your point about you sort of CYA is quite important. The pressure is coming from top down, obviously. And so at some point somebody says, “hey, wait a minute, instead of Gen AI, we could do this. Or the expectations of 10% improvement of productivity from Gen AI are unrealistic.” Who wants to hear that? And I think that's part of the challenge, what is a realistic expectation? We don't know the answer. And with the products that we're selling, we don't know yet what a realistic expectation is for prompt engineering using a GPT for these mundane tasks. We think there's we think there's productivity—
IS: Yeah, there's some early proof points, but it's not enough yet. And in many markets it's not. If you're designing a bridge, an engineering firm and a, yes, you can probably use it somewhat. But ultimately, am I going to ride on that bridge designed by a machine? Maybe. Maybe not. I mean, I want to see some proofs that this is not going to fall down. Not a Baltimore situation kind of thing. Yeah. So, it sounds like a lot of the services firms can really benefit from this, both large and small. We had another question here. I'm not sure I completely understand this, kind of a semantics here. “How will custom services GPTs impact the market?” Not super clear, but maybe you guys can take a take a stab at that.
AD: The thing I keep thinking about is that when the internet came out, there was the same kind of sense of urgency, although arguably today the pressure and the urgency is what it was when the internet kind of landed.
DB: Yeah.
AD: And yet here we are today, 2024, less than—let's be very conservative, less than 15% of B2B business, possibly, what's going to play out here for services firms is the kind of frontal—the question of is it an existential or is it an accelerant is not the right question. It's what are the dark corners where there's hidden value?
DB: Oh, yeah. Exactly.
AD: Right?
IS: Everybody wants productivity, that's for sure.
AD: True. Yeah. That's it. What the and the and the other mega trend is that the demographics are not in the Western economies favors in the sense that we're not having enough children to meet the replacement rate of the population. So service organizations, everyone is going to have to look to productivity gains.
DB: That's right. At TSW, we had a customer advisory board where we brought in some of our most important strategic, clients. And I use the analogy that we both have used, which is we're sort of in the state of Gen AI especially, but even I would say machine learning—that that the internet was and I had a picture of Amazon's website in 1997 circa. Like and you looked at you're like, oh my gosh, this is like a CGI, like bad it's like somebody doing the website in the back. And everybody knew it was important to to your point, they knew they had to do something. They didn't know what. They didn't know how it was going to disrupt. They knew it was going to disrupt. So I think that the Gen AI is and the expectations were off the chart. You got to have a website. No, we don't have any website. Yeah, you do. And here we are 30 years later. And yet to your point, what's the penetration of B2B e-commerce? What's the penetration of the internet? Are we still—how much more disruption is going to have going to happen through just sort of basic internet technologies? A lot. And so we should look at Gen AI and machine learning in general as something that is going to be a 20-to-30-year trend. So I think we're at that point where everybody's kind of going, “gee, I want to do something. Not sure what, not sure how powerful, but I better get on with it.”
IS: I wanted to ask you, maybe we have maybe time for one more question. What have you guys learned at Certinia about adopting generative AI? I mean, you have this big team of doing work with a lot of clients, but you develop your own products as well. What have you guys learned? What would you do differently, if at all? What are the big takeaways, do you think?
DB: Yeah. Great question. The first thing is we've taken a principle called pragmatic AI. And essentially what it means is it has three principles. Number one, it needs to have impact. Number two, it needs to be easy to deploy because because we know machine learning. And I think even Gen AI has historically been very difficult to bootstrap in an organization. And last but not least, we want information that is coming from AI to be closed loop to be brought back into the system so that you can do your work. So it needs to be actionable, it needs to be deployable and it needs to be closed loop. And that closed loop means, “hey, if I'm if I'm doing some knowledge work over here with our or our customer success product, if I'm getting information from from Einstein, GPT or from a prediction, I can use it at the point where I'm doing my work.” So I think those principles have been very sound. They've resonated with customers. What we've learned is the same principle applies when you're doing something that is materially different with your product, and that is you have to work very, very closely with your customers. So we do an internal early adopter. We go out, we work with—Jellyfish is a really good example of a company we've been working very closely with on AI. We give them drops. We work with their people; they give us feedback. So a great example was, “hey, if you have something that is a prediction, help us with mass actions based on that prediction.” That's one of the feedbacks that we got from working with Jellyfish. So I think that that principle is really important because you don't know you don't know what the best manifestation is of how those prompts should be built up, what that interface should look like. So that's, that's that's the number one learning. The number two learning is that the technology piece is a very, very small part of the problem. And I think all knowledge worker professionals will have to have an operating knowledge of Gen AI and AI in general to do our jobs. That enablement—technology always needs enablement. This is a big enablement hurdle. It really is. That that piece is is much bigger than we anticipated and it's it's much, much more important. And it's probably not a surprise.
IS: Because ultimately AI can help potentially maybe in getting the best people, or better people, or training people up.
DB: Right, right.
IS: Maybe instead of developing a course you have AI develop the courses and categorize them and organize into exercises and stuff like that, which is always very tedious kind of thing to do.
DB: It's a very interesting area of exploration in Gen AI, which is expertise mining. When you're doing—one thing that Gen AI does that is really exceptional is its ability to consume unstructured information. And project work is very unstructured and most knowledge work is very unstructured. And so it's a great topic and question to explore in the future.
IS: All right. On that note, thank you very much Dan.
AD: Thanks, Dan.
DB: Thanks, Andrew, I'll see you guys again. Good to see all.
AD: Thank you for listening to The Margin. If you have questions about today's episode or if you'd like to schedule a call with an MGI analyst, reach out to us at insights@mgiresearch.com. You can also reach us on LinkedIn, Facebook, and X, and you can find more information about our research and advisory work at mgiresearch.com. Until next time.