Name: AI as a Strategy: Lead with Intelligence in 2026
Uploaded: 2026-03-14T06:36:05.429Z
Duration: 1 h 8 s
Description: AI as a Strategy: Lead with Intelligence in 2026

Transcript for "AI as a Strategy: Lead with Intelligence in 2026": Hey everybody, welcome to today's webinar. This is AI as a Strategy and my name is Justin Britt. I'm going be the host along here with Florian, who will introduce himself shortly, but I wanted to start out quickly with a couple of housekeeping tips. Number one, please use your Google Chrome or Firefox for the best experience. Audio is going to be streamed through your computer. There's no dial in. You can download the slides via your resource widget on your top right. The widgets can be resized to fit your screen and this recording is going to be emailed to you within twenty four hours. Along the way, we're going to have some questions that you can ask Flo and there's a Q and A box that you can type in and any unanswered questions are going to be addressed offline. So I think that's it for the housekeeping and let's get on. So welcome to today's webinar from Deltek. It's called AI as a Strategy, Lead with Intelligence in 2026 as we kick off the new year. We're hoping to give you guys insights on what professional services leaders are doing in the realm of AI in the new year. So we've got a quick here's our agenda. It's short and sweet. We've got a short introduction. We're going to have a fireside chat between myself and Florian, and then we'll do a quick Q and A where we will go through some of the questions that you have sent us either beforehand or as as things pop up during our webinar. So, without ado, I'm going to stop here and just let me introduce myself quickly. My name is Justin Brick, I'm a director here at Deltek. I'm the owner of a design and professional services firm and I've been in the professional services industry for almost the last twenty years in various roles and in that time, course, AI has been a tool which has bubbled up recently, which we're using all that I'm using within my team all the time. And over here is our main speaker Florian. Florian is a partner with the Visible Authority. And Florian, I'm going to let you take a few minutes to introduce yourself. Thank you, Justin. So I'm not going to spend too many minutes on it. Yeah, my name is Florian. I'm a partner like you said, just repeating this, at a very small boutique firm called The Visible Authority. We specialize in helping already successful or established consulting firms primarily increase their performance, increase the success by refining, focusing, narrowing the value proposition, then from there sort of intentionally redesigning and refining certain aspects of the firm. Everything from client success journey, service packaging to how delivery gets done, but it all starts with the value proposition like this dreaded question, what business problem do we actually solve around here? For whom and to what end with which outcomes and so forth? Some aspects of that might be relevant for today. Like you, Justin, I've spent the majority of my so called career. I think I'm getting close to sixteen years in 2026. Professional services industry, I've been working inside and outside consulting firms, IT services firms, including very large ones like the Lodden Accenture and many, many smaller and mid sized ones here around Europe, because I'm based in Germany. Happy New Year everybody, just if I may say that at this part and then back to you. I don't think this is the seventh is the last day you're allowed to say that. Oh, is that so? That's the rules. Don't know. Could you just make that up or? I'm just kidding. I'm just kidding. Happy new year to you as well. And everybody online with us. Well, awesome. Well, thank you for that. We wanted to set the scene quickly, you know, when it comes to AI and professional services. So, you know, in a market where every firm is racing to keep up with AI, professional services leaders face a new strategic question. How do you harness AI as a true differentiator without diluting your value or eroding client trust? Yeah. That's the big one, isn't it? That's the big one. That's the big one. That's kind of the overall question we're going to be tackling during this. And I'm going to we're going to start this off here with just a series of five questions that I'm going to pose to you Florian and you've already kind of taken a few minutes to think about those, but let's start out with this first one with how has AI already changed professional services and in your opinion, where will, its impact accelerate next? So what's coming next? Yeah, so let me preface this by saying, I mean, I'm from my background, I'm a business development guy, so I'm not technical, I'm somewhat technical, but not technical enough to speak to the inner workings of AI, but I certainly have some views on the questionnaire and people can see the four bullets that I'd like use to talk about it. But let preface those four by saying that just came to my mind before dialing into this. It's interesting to think about when we say how has it impacted the industry already, right? What timeframe are we talking? Because I think sort of algorithmic means of automation have been with this industry for quite a long time. And I think if you work for one of the big four or larger firms that have offshoring or nearshoring capabilities of delivery centers, I mean, it's always been part of the game to say, where can we sort of standardize package work? I don't know, managed service providers will do this as well. And then once we have a certain systematization and the stuff we do for clients, can we deploy some means of automation, right? For job routing, ticket resolution, data crunching, whatever it is. I mean, long before Chetzky and the like came around, you had process mining software, which some people might be familiar with, right? Hit the market like Celonis out of Germany is a name here. It's effectively software that you dispatch that in the network of a company, it comes back with like a graph of the company's processes and oversimplifying. So that already automated what used to be hands on junior consultant work. I mean, we used to fly people to places to interview the production line manager, figure out what the process was. Suddenly a software does it. I mean, that's been part of the game for a long time and I do think, maybe to give people their self confidence back a little bit, the industry has always coped quite well with those But then going back to your question, I think that you can contradict me, how I read this then was AI as we discuss it now, right, today, in the era after ChatGPT. Suddenly we have those chatbots, those LLMs, we have agents, we have the latest iteration of that sort of thing. And so I'll answer through that lens, if that makes sense. I hope that's fine. So I think what I've seen from our perspective, what AI has done to professional services, especially in the last year and a half, two years, is that's my first bullet it has certainly created lots of expectations in the market, because these tools seem to be so magical, right? When ChatGPT three or whatever was came out, people are like, oh my god, that's general purpose tooling, it's almost like magic, which we all know now, probably not entirely the case. But it creates an expectation in the market that has impacts on clients' view on consulting, Clients' willingness to pay, and it gives rise to all kinds of questions of the why should we pay you two and a half junior hours for this, right? You will do this with Copilot in ten minutes, is a real thing client Absolutely happening in the industry. I mean, you talk to people who are we had discussions with people in cybersecurity space where sort of tier one support or certain steps there just get automated. I mean, that's another field where automation has always been part of the to be fair, right? But clients' expectations really are, they're asking painful questions now about what's the workload here? Does this have to be done by a human or are you using a machine anyways and overcharging? So there's discussions, those are difficult to navigate for firms, because often it's not the case that the bot can do it. Clients think, oh, they probably do this in fifty minutes with ChatGPT, but it's actually much more nuanced and difficult than that, even if you might be able to use an AI here and there, someone has to double check it, there's more work to it and so forth. I talked to a couple of people in the big four space in the tax area, which is not something I was super familiar with, but they have this problem for real because they often have retainer agreements with their clients. If you're a multinational corporation, you have some tax issue, you just dial up the tax guy on retainer and ask your question. Willingness to pay for those types of services, for those retainers has gone way down because the clients expect that the junior analyst puts a prompt into perplexity, gets the answer and that's it, right? Why can't they be faster? Well, turns out, it's not the case. You still need the junior expert to do the work because international tax law can be complicated. And I think that's my second bullet here, is the rock in the hard place, right? Because sometimes the client's expectations about what could be done with AI is up here, then firms experimenting find out that the tooling is not quite there yet in terms of what's been promised, right? So you are, and I empathize with this a lot, if you're responsible for the P and L, the market expectation is for you to reform your group using AI, but then if you implement the AI, it turns out you have to make tough decisions between how much can we lower the quality or how do we govern this, it's not as easy at all. And I think there's a clear divergence, you see this in the market, there's the stories, my LinkedIn feed is full with how AI is going to kill consulting, you see it there. Sure. That reminds me of something I just heard today, where ironically I think one in five people have a job today that didn't exist twenty years ago. So we're getting very wrapped up in, oh my gosh, you know, how are we going use this? It's going to take over, it's going take my job. But the reality is like we, you know, as an industry, we find a way to cope with that and, you know, there's a lot of opportunity there. Just because something didn't exist doesn't mean it's not going to be fruitful, you know, in the future. Absolutely, you're not gonna hear me say that this stuff is useless and it will go away week after tomorrow, absolutely not. But I think my bullet two here is also true is that a lot of those technologies have been overhyped to be blunt about it, they have been presented in ways, they came with suggestions for how to use them that actually do not give you the best value out of it. I made this joke last week that Chateappity in many ways is like a Swiss army knife of knowledge work, which means no matter what you're trying to do is always the second best tool to use, right? That's what a Swiss Army knife is. The moment you find your Philips screwdriver again, you put the bloody Swiss Army knife away and you screw the screw in with the Philips, right? But that's a very specific criticism of the large language models, right? Because they are presented as this, oh, it can do everything. And then if you really work with it on something serious, can, I mean, Deloitte found out the hard way, right? They've been in the press a couple of times for the results they've gotten out of the machine. The reality on the ground is not as easy. And I think that is another reality firms need to contend with, so that would be my bullet number two. That said, I do think there's lots of value to be found if you really know what you're doing. So if you don't have those super broad, let's write a report or whatever use case, I mean, you can save absolutely a couple of hours by having the bot generate a written report if you give it very precise prompts and so forth, but that's maybe not the very best use case, but the value is there and it increases with the extent of expertise you can bring. I mean, you and I talked about this beforehand. If you give both a carpenter with twenty years experience a hammer and you give a hammer to me, right? The value of the hammer in my case is capped at a very low level because I can't do carpentry, right? The professional carpenter will be able to do very interesting things with the hammer indeed and I think the same is true with AI tooling, which that's what it is, it's an enablement tool and the value you get out of it is limited by the experience you have. That's a constraint that technology comes with, which then if you flip that means if you are a very focused, very expert niche firm that has a very clear idea of how you go about solving problems, and then there's a step down from that, because of your IP, your systematized way of consulting clients and all these things, you have all kinds of ideas where you could plug in some AI enabled automation and that really helps, then yeah, you can find tremendous value in that and you can get small teams do a lot of stuff that would have required many more people beforehand. And we see this a lot. I mean, there's still caveats with the challenges the technology has, the AI models and all that stuff have, but it is true that in some cases expert teams of just a couple of people or small boutiques can go up against larger competitors in bits and stuff because the constraint of capacity, oh, we just have 15 senior people, we don't have an 80 person bench. A part of that can be mitigated by technology if you use it cleverly, right? So that's another impact there that's leveling the playing field a little bit. One more nuance before I move to bullet four, there is another play which I see and this is James O'Dowds, a colleague in The UK who runs a firm, the name of which I've forgotten, but they are working on the report where they look at the way very large firms find success with AI and in their research, they have found that AI is just the next iteration of what I've already described. So they find leverage in the new technology by augmenting employees with it, but those employees are in labor arbitrage geography. So you take delivery center in The Philippines or in Manila and you invest there, you invest in the people and the infrastructure, you train them, right? So now you have a bunch of real well versed, whatever it is, software experts sitting there. If you equip them with purpose built or purposefully designed AI tools, you get a much more leverage out of that. So that's the other place. So either you're a small boutique really expert and you can leverage your expertise with the computer capacity, or you are a company that's able to invest building out those large engines for delivery and you make AI part of that. It's also a good way of getting your money's worth out of the technology. All of that as always, right, caveated by the fact that there's still risk in those solutions as well. You know, there's all kinds of questions around legal claims, copyrights and all these types of stuff, governance, data security. I mean, I was in a position to make major investment decisions around this stuff, probably what would cost me some sleep is that a lot of the services you can hire, the compute seems to be still subsidized, right? So, I mean, many of those large model providers lose money because they give the compute and the tokens to you for less than it costs them to crunch the numbers. So what happens if the subsidies run out, right? Are you embedding a technology at the core of your business where the prices could increase tenfold sometimes down the line, right? Right. It just happens in every other sort of cloud SaaS, whatever. But those questions aside, I mean, there are ways to find competitive edge and the two I'm seeing or two I've seen emerge over the past one and a half, two years were those two, genuine expertise and then add computer capacity to punch above your weight, or you play with the big guys and you really, really get even more leverage out of those outsourcing delivery engines you've built over decades anyways. All of that said, if you bear with me for one fourth point here, I do think and that's my subjective opinion, right? People can disagree with me here, but towards the end of 2025, I got a little bit fed up with the navel gazing of the industry. I'm like, oh my god, is AI going to destroy us all? I think there's a danger there of sorry, let's start again. I mean, there is a market reset, but it's there for other reasons than just AI alone. We had the Covid bubble, huge demand for lots of professional services, had a bit of over hiring, we now have cooling demand, we have very challenging geopolitical times, slowing growth in many economies, with a bunch of reasons compressing the industry, it's not just AI. And then there's a tendency to, you know, overhype the AI thing and discuss the technology and sort of move around in circles, tearing our hairs out. I think there's a place to discuss AI as a tool to build the consulting or service business you want to build and then that's a good way of having that discussion. But if all you do is sort of doom scroll through the overhyped news releases of platform vendors threatening to kill consulting, I think that's not a healthy entrepreneurial stance to take. And I think just I would just want everybody to be aware of there's a sort of a risk in the news cycle to just distract us from the real jobs this industry have. And I think the real job is just to figure out how can we help clients be more successful, fix systems that are broken, like the things consultants do, right? Then once you have good answers there, you can step down into, okay, where could AI play a role? Which is a different lens than being scared or over impressed by the headlines of who built the next tool to replace consultants. Don't see that. It's certainly important and it's everywhere, but at the same time, you know, I kind of think of AI and automation, you know, in the same bucket where, you know, you used to have a chatbot used to be, you know, that's been around for fifteen, twenty plus years, right? But you know, now it's AI, you know, it's not that much of a different iteration. Of course, there's, it's a little bit more technically advanced at this point, but you can, like you said, you can kind of get lost in the shiny new object, if you will, if you think of everything like that. Yeah, it's just not good enough for that. I'm sorry to say, like it's been hyped a lot and I get it, right? If you are one of those service vendors that has invested other people's billions, you need a good story which industry you're gonna disrupt, and I don't have the number here how many billions worth of consulting industries, but we are going to kill consulting and steal that market share from those guys, right, who are on the call here. That's a good story to have vis a vis you investors. I just referenced to the question here on the slide, as far as I've seen up until now sorry, the tech is not there yet. It can be helpful if used in very specific ways, but it's not going to kill consulting anytime soon if I am supposed to make a prediction. Yeah. Well, awesome. Those would be my big four. Yeah. No, that's awesome. Thank you for the insight there and interest of we want to move on. We got quite a few other topics that we want to touch on. So let's go forward to our next question that we want to tackle, which is how does an AI forward proposition reshape service design, trust and differentiation in professional services in your opinion? Yeah, I mean, is sort of the key question everybody has and I would like to say, so I've alluded to some of the stuff beforehand, right? The perspective needs to be, you got to address this or you got to come at the AI question from a question of what business issue do I solve? What urgent, expensive, existential struggles do I help my clients overcome? And then you've got to figure out where could I leverage AI to either do that more efficiently and more cost effective ways, whatever, or where could I use AI to maybe improve the quality of the output or the advising, whatever it's multiple angles there. So that's I think the first bullet I have. But when we say AI forward, I think we have to be aware that AI exists and we have to be aware of how those narratives and stories, killing consulting, all that stuff, how that influences clients' perceptions and expectations and questions and so forth. But it doesn't mean we need to lead with the AI, I mean, honestly, in our work with clients, we always advise against it's like if you have headlines on your website, like reimagining your organisation with AI and so forth, that is a dangerous thing to do in many ways, because it's not linked to a real client issue and the history is full of great jobs, now was it of technology focused practices, right? When I started at Accenture, they still had a mobility practice and then they had a cloud practice. And then I think they even had a blockchain unit and a metaverse practice. It's always the case that either technology dies and then becomes irrelevant or technology becomes successful and then it gets pervasive, right? So this is not a good play to say, oh, we are AI forward in that way. But if you don't make AI the promise, but you make AI the means of delivering against your promise, then I think you can build something that is quite interesting indeed. If people read on LinkedIn and elsewhere, there's a lot of talk about how the industry moves to outcome selling and outcome focused contracts and so forth. I have a bit of a differentiated opinion. I do think that you need to link two outcomes. You have to establish a clear through line of, look, we are specialists in solving this problem. We typically in an engagement with a client like yours have this and that impact and we have a good track record of enabling them. And then this impact over time tends to bring about or yield or influence those business outcomes. Establishing that type of through line, it's not about giving guarantees of 20% efficiency gains or your money back. Because truth be told, depending on what service business you're in, chances are you don't have a lot of control about that. I mean, might be different if you're an MSP player in the IT services domain and it's about ticket and resolution times and stuff. You might have a greater control over outcomes, but a consulting firm usually relies on the client's contributions as well, right? So you are not in control. So I think that is what we have to think about when we talk about outcomes selling, outcome oriented contracts. There just needs to be clarity between the expertise you bring, the delivery method you'll apply, maybe where you use AI and what the outcome results. I don't have a heart doctor, but let's presume I had one, they don't promise me that I'm going to live to be 100, right? They say, hey, we specialise in heart health. If you have a problem, here are the therapies we'll apply to make that go away. And typically our patients, that's the logic. And I think if you are always working on building those things, so having clear relevant promises, then building the engines to deliver against those and that is where the AI has to come in. And if you read about layoffs among the big firms, retraining acquisitions, I think that is what the large, well funded players in the industry are doing already. They're already doing this and saying, hey, we have built this delivery engine. If 20% of the pyramid, the junior roles, the delivery centers could be replaced by computer. What does that mean for our operating model? What does it mean going forward? I think that is what an AI forward strategy is and that involves looking at the tools, what can they do, what can't they do, where do they make sense, where can we fit them in? And the other aspect, and this is again a little bit more hypothetical, my bullet point three here, and you alluded to this when we talked earlier, if machines can do a lot of the, I think about this, I think a credible way where you can deploy these systems is what I think of as the prep work, right? If your professional service is at its core about enabling a decision or arriving at a recommendation or creating a plan, which again, that's not the same as the IT services colleagues maybe, but if that is the thing, then a lot of the steps that lead up to that decision or recommendation are things like collecting the data, normalizing it, crunching the numbers, analyzing things, forming a hypothesis, testing that, building a model, running simulations like all these things arguably are susceptible to machine learning, I should be careful, or even AI in the chat GPT sense, although I recently gave chat a spreadsheet to analyze and results were horrible. It's not designed to understand numbers, I'll tell you that. But anyways, so if those things can become automated and delivered by machines, then the willingness to pay the value in the consulting relationship will move to the other stuff, which I think is an interesting shift if you think about it, because those hard parts, the quant stuff, the do the numbers, build the models' part, when McKinsey would hire physicists to do it, that used to be the very expensive, very respected niche of consulting, whereas the softer stuff, leadership alignment, change management, implementing the advice, that used to be something a Bain or someone else might subcontract to another firm. I'm exaggerating a little bit, right, not to offend the Bain guys, but that comes after we've done the hard part, the heavy lifting. Well, if now the heavy lifting goes into the computer, willingness to pay value and the actual decisive impact you can make on a client will maybe move to the other end of the chain. And I think you can see this to a certain extent with some of the big players, again, that are very technically capable or they're very strong analytically, thinking more about those components of can you stick around to help facilitate the change? Do we maybe need some communications people or leadership trainers, my goodness, on the bench to really deliver the full package? This goes together with the outcome orientation, right? The client wants the outcome, you've got to bring all the capabilities and if you've been good at number crunching in the past and you now do that in computer, well, maybe you need a different set of skills at the back end of this. So I think that's another aspect of AI forward is the people stuff. It's always been important, always, but you could get away with maybe leaning more heavily on the technical things, but I think that will become a lot more important and this is something I would think about the more, the further I am on the more technical and analytical spectrum and we see this in our work, we work with many clients that let's say they used to be an expert containerization Kubernetes IT shop, it's very technical, they are in the infrastructure and so forth. Suddenly they get questions about, hey, how much of this migration could be done by a bot and then could you spend some hours training our people, right? So the soft part, I think I made the point already, I'll stop, but that'll come to the fore much more and maybe that's an area for reflection, development, maybe investment even. That ironically has nothing to do with AI tooling, right? But once the machine has given you the data and the insight and the recommendations, what do you do with it? That end becomes important, I think. And then, yeah, the fourth part I think is sort of just a summary and then we can move on with all of that. Services and delivery, you'll have to rethink those from the promise you make, what's the business issue we solve, what perspective do we bring to that, then how our delivery engine connect into the client's operations? Once that's clear, okay, how do you package that in units you can actually sell? And that's all I will say. I know there's a lot of talk about value based pricing and these types of things. Like I said earlier, I'm a bit skeptical there for the reasons I mentioned, but also clients are not terribly good at procuring them, right? I have to measure success, I have to measure risk, then they have to negotiate, it's easier to buy time and materials, okay. So as long as these two contrasting poles are true, which they have always been and still are, think, yeah, you just got to be smart about where your services, what they look like and how delivery works. Oh yeah, sorry, my notes was one thing. I think the better you get at selling an impact of a result, that's very clear. And the more risk you take off the table for the client, the less they care about how the sausage gets made. And that's how you probably can stay away from discussions of does a junior do it or does Copilot do it and what's the hourly rate, right? Because you say, hey, this is the program, here's what you had at the end of it and it has a sticker price of $50. And then you can work internally and increase your clever use of AI to bring the margin up, but you don't have rate cut discussions. One of the quotes we were talking about earlier that I always think about as a marketer too is that, you know, when everything is, this is something I heard recently, when everything is AI, we got to figure out ways to lean into our humanity. And that's the way that you're going to help. That's the way you're going to differentiate yourself as, you know, I think the word of the year was AI slot, right? I think that was what, Miriam Webster put out. So AI is everywhere, but there's a reason that, there's a reason why, you know, if you look in Google, the, Reddit topics or Reddit reviews are ranking higher and higher because there's actually people behind that, you know, making those comments and it's not just coming from AI. So, you know, finding out ways to highlight and bring out humanity is is going be super valuable for us as leaders. Yeah, two things I really want to click on because what you said there, think it's exactly right, but I can make it even more practical. But I think what we mean when we say bring your humanity in a consultancy or professional services environment, what that actually means is bring a perspective and be opinionated. And not just for being opinionated sake, but have a point of view as a company about here's what's broken, if I'm in supply chain management, like here's what's broken in supply chain management, generally speaking, here's where typically you would have opportunities as a client and here's how we think you should do better supply chain management, and then that is your perspective, that is a way of being a person or a human as an organisation, you bring specific perspective, that creates some difference in the market, because there might be other people or other firms having other opinions that are also viable, right? That's how you get to the differentiation point. And I think that's a way to make it a bit more practical when we say bring your humanity. Have a distinct perspective on your client's business and the issue you solve there and how that should be done properly. I mean, my heart surgeon, go back to that example, might have the same, right? They might have perspectives on diet and better ways to do a surgery than the next colleague over and now I have a choice. I might prefer one or the other, but there's a difference between the two, which I think that is what you're alluding to. And that's another reason to not be with the AI this AI ad ad because the other firm down the road can buy the same model as you then the differentiation The is gone, stuff is just like electricity. It can't be the promise, like I said. And I love you bringing up the slop because that's another argument for not positioning as the AI shop. Because there is discussion of a bubble, there is slop, there is outright fraud and crimes being done with this technology. So it's all still the Wild West. We don't know how it ends, but if it ends badly and AI actually becomes synonymous with SLOP, you've just positioned as the SLOP consultancy. That's not a place you can So it's early to make those kinds of calls. Yeah, pushing the metaverse as many of the large consultants have done, like incredibly hard in some cases, has cost a reputation. Just go to a couple of the enterprise plans and ask about them. Ask about the jokes they make. But, I mean, that's the it's not gonna cost you your your company. I don't think it does that much, damn it, but it is real. Like, if you get too associated and too excited about something that's then turns out to be false, and you were supposed to be the adviser guy who gets the stuff as things. Right. It's okay to say, let's figure this out. That's a different stance than saying, we are the AI transforming your business boutique. Mean, that's a risky gambit. My take. Sorry. We I think we want to move move over. Is that fair? Yeah. Yeah, I think we should move on for sure. We have a couple more we want to get through. So, let's move on to our next question. So, when does AI truly add value for firms and when does it just create noise or skepticism in your opinion? Yeah, so there's two sides of the metal, right? It's one thing, what do you recommend to your clients? Consultants also help with implementing AI, that's one part of the game. Other part is what do you use internally? I think the bullets here are true for both cases and I can move through them quickly because you alluded to punches out there as well. So I think, yes, it is true, capacity for certain tasks can come out of the computer. Like ChatGPT can do email reasonably well up to a point, okay? As long as you're aware of the points and you can sort of put it into guardrails narrowly enough, you can find value there. If you use very specialized ML algorithms like we had for decades, I don't know, visual inspection in industry, in a production line or whatever, those work very well, like they're very proven technology. So certain capacity for certain has can come from within a computer and if you are very smart in using them, and I think that the narrower you define the scope and the clearer you are on what it needs to be doing, the more value you can find. That's also my second bullet point, right? The more specific the use case you could put it to, the better and the tighter governance than there are rules, all of this stuff. There is another aspect of this, which I didn't fully flesh out on the slide here, but we've done this in our firm and I've given the advice to some friends around consultancies as well. There might even be a function of AI and take this with a grain of salt because I made this as a half joke, but I said if you have a task in your team or your workflow that you can hand over to an LLM without further refining of rules or prompting or creating a custom instance, if you can do that. Ask yourself whether it's worth cutting the task overall. My go to example is, as probably every consultancy, we have experimented with those fancy meeting recorders everybody has these days, right? The AI that hops in the conference call and writes a summary. Well, that really told us the thing, which is, A, the AI has made lots of crucial mistakes and you can't rely on them for higher stakes and dense meetings, that's one thing. The other thing is you can see the click rates on those, right? That's the brutal truth of who's reading the memos. No one does. So it's probably actually worth not having a full recording of the call at all, but do the good old metrics, you know, decisions, take like a very short thing, summary, executive level, you could put in an email and send to the right people, you save everybody some time and work and that's another lens of specificity, right? It's the question of This meeting could have been an email. Yeah, and now that you say, hey, an AI could, and if you look at a task or a job in the chain and say, hey, the AI could do this. If it's valuable, fine, put it on the app, have it run there, but it's also the moment in time where you pause and say, we need this in the first place? Right, that's one. The real winning combination, bullet number three here is, and this will remain true forever, experts plus machines. A senior engineer who really knows what they're doing, equipped or enabled with capable AI tooling will deliver significant improvements in I think both quality and efficacy, which I care personally a bit more about, but then also efficiency. And that's why I would always look at those augmentation cases if you're talking AI in a broader sense. If it's just killing rote work, automating stuff that I need to click three buttons or whatever, that's also valuable, but I doubt that you need AI for that, right? There's probably other automation ways to get rid of that stuff. And then the last point here is if you create noise and I see this a lot when you say, hey, we have this AI division and everybody gets like a meta set of meta glasses and they do expensive executive demos, I file that under noise because that's light on investment, right? But if you really sit down, you really craft a strategy for solving a client issue and then as we said before, you built the delivery engine, you invest in your people, you invest in the carpentry training much more than you invest in the hammer, that sort of thing, then there will be a return. I think there's sometimes a tendency to do sort of surface level AIs of putting lipstick on a pig so just you can float with the hype. I'm not saying anyone here on the call does this, but it does exist in the industry, that's just you might as well not do that, right? Either be serious and do the investment or not and I want to call out that some of the stuff is expensive. Those models can be expensive if you want to run stuff locally, you want to get your data in order, so for a typically investment or asset light industry, consulting, you can start by a notebook, bring your smart friend along and you have a business, so we're not very capital intensive in that sense. I think that is something that might change for a little bit there as firms do invest in the skills and the tooling they need for the next phase, whatever that is. I don't know if that answered the question fully enough or if you No, I think that's great. Is there any specific AI use case that you are often hearing from, some of the leaders that you work with that you just feel like you don't really need that? Tons of. So LinkedIn is full of people who confuse the artifacts of consulting with the process of consulting and the value is in the process. So if I see another, oh, look at my magic slide writer, it'll kill McKinsey. My brother, the value of McKinsey is not in the slides. Like it's not, they need to produce the artifact. It is true, it is true. And that takes time and there's maybe some value in automating that to a certain extent. But to believe for a second that's what justifies the price tag of the teams, you have no idea what you're talking about. And honestly, I know I'm exaggerating because people love their decks and the clients insist on them, but it's one of those cases where you could ask the question, should we get rid of the process altogether? Because if there's an over fixation of the deliverable someone taught me this while I was working at Accenture, they said if the client is obsessing over the deliverables, the appendix needs to be longer, you need more charts, yada yada. What that means is the communication during the project has not been good enough. It's a signal of uncertainty, they're not clear on the decision they can't support, so the deck must be better. Okay, so there's a lot of that in consulting where people build AI solutions for problems that come out of the complexity that exists because the consultancy has not yet managed to focus its value proposition and its strategy. I mentioned the tech builders, there's others, there's tools for knowledge management and mining, right? Solutions that go past into your graveyard of PowerPoints and help you figure out which of those you could recycle. Well, my answer to that is if you need that for every new RFP and the bot has to sift through 260,000 presentations, why is that the case? Why does your practice not have one sales stack and that's it? And then people's heads explode, but the answer is you can't do that in your firm because you've never made the investment of sitting down and saying can we narrow what we do around here to the extent where we could have a brochure that covers 80% of the cases that come through the door? And instead of fixing that, because that is very difficult, they buy an AI solution, which they don't need. So, to flip your question, if I may, because that's also a good answer to this question maybe, if you buy technology that tries to take the cost out of or replace the artifacts of consulting, not the process, that's a waste of time and money in many cases. If, however, you can find something that ties into the process of consulting and I mentioned that beforehand, collecting the data, putting it together, cleaning it up, crunching the numbers, building models, that is the process, that's where the value is generated. If you can find solutions that do that very well, I mean, build the business case and go all in. Salon is Germany's second software success case after SAP for a reason, because this process mining stuff actually works and saves tons of time. It doesn't create the deck, for sure not, but it replaces the process step having to interview people to draw a process map with a click of a couple of buttons and exaggerating, but that's the thing. Nice. Some good insight. Thanks, Flo. Appreciate that. Let's, in interest of time, think we got a couple more. So I'm going move forward to, that was I the last one, right? So very quickly because I covered a couple of those. So there is task based support. You can do it. My brother is a software engineer. He says most of the vibe coding apps are not, they're bad, but he uses them regardless because he uses them as Google Translate, right? So he works on 15 projects, I don't know, there's 50 different programming languages as he switches from one to another, he wants to do a thing, but he doesn't have the syntax of that particular programming language in his head. So he uses an LLM to say, I'm trying to build this function or whatever, right? I'm not a coder. What's the syntax for that in this language? And it gets a decent answer and I think all of us have instances of that where we say, yeah, actually that's a tiny part of my daily work where it helps and I think those exist everywhere. Specialized machines, I've seen some interesting use cases, for example, in the realm of application modernization, so IT service companies trying to modernize legacy mainframe apps in a bank, for example, often very drawn out process because the clients might not even understand the business architecture behind that system anymore because the people who built it twenty years ago, twenty five years ago are no longer with the company. So that information is lost. Well, now those are not OpenAI elements, but specific generative AI solutions exist to crawl the code base and come back with a hypothesis to tell you, okay, we think this was the business architecture. So again, that collapses time in the process of building towards the modernization. And then these things can also support you in refracting code or whatever, so there are interesting solutions of that sort, they're not ChatGPT, but they're dedicated tools like that. And then the third one, should really not be talking about that because I'm not technical enough about it, but there are now junior companies playing with layering or nesting, I should say, this idea of agents and then context knowledge or rule based works and then another layer of agents. And so what that does is you build a system where an agent, like it's the digital junior consultant computer, thinks about operating model stuff or ingest some data, generate some idea, takes that into the next layer. That layer has been trained in the frameworks the consultancy uses and it roots out the bullshits, excuse my English, that the junior AI just created. And then those turn around and run it through governance layer for functions. So these types of sort of multi layered systems, they seem to come up quite a bit here and there and they seem to be good at what they do. And then my last word here is whenever people say agentic AI, I immediately take a step back and become critical if not cynical. Those things will eventually become good. I think there's great stuff if I understand correctly happening in research with agents being able to take on longer tasks and all that stuff, but a majority of the things that are being marketed to me as a buyer, AgenTic Systems and they play with a couple of them, maybe it's me and I'm too dumb, but I can't get them to do what I need them to do with a reasonable amount of effort involved. Maybe people here on the call can correct me now, but that would be my take on the real world stuff. I apologize if that wasn't the most specific, precise representation, but those are the things I see. No, I agree with you. A lot of the agentic AI that I deal with is not quite there yet. I would concur. Although of course we see the value. I think we got one more question here and then we'll try to get to a couple more, submitted from the audience, but wanted to ask you what critical questions should leaders ask when shaping an AI driven business strategy? You already touched on some of these things, but Yeah, I think that's just a synthesis in what we said, right? I would always wonder, are we deploying AI for AI's sake or are we doing executive entertainment as I used to call those demo workshops? Are we solving a real issue here, right? Mainframe modernization is a real issue, not knowing the business architecture and having to guess is a real issue, deploying AI on that, if it works, great solution. Next thing is, is this about doing the right thing? That's sort of the same idea but a lot of the AI plays seem to be around efficiency and coming back to my example with the slides or the call notes, sometimes because there is a resetting in the market now, or we are struggling with these questions, I think there's an opportunity to challenge the actual thing itself and not just deploy it and forget about it, I don't know. So it's a good time to take that pause and ask, hey, is this about making the thing more efficient or should we actually be discussing whether it's worth or not doing the thing? What is that? We even have all the inputs? Given the situation with data today in most consultancies and also client organisations, I sometimes chuckle when they talk about deploying AI for this and that. Like example from our world, we sometimes, that's part of our process as we refine value propositions, we analyze past successful engagements of clients and we try to contextualize the work the firm has done. And then we ask them, hey, great job there on the SAP implementation. What was the overall business context for that and how did that turn out? Was the client pursuing integration, growth strategy, market expansion, was it and how did the KPIs for that turn out? And you'd be surprised about the number of consulting firms that A) don't know because they never asked and B) they routinely do not do baseline measurements for higher level business KPI as they embark on their projects. And then they later come back and tell us that they struggle with case study curation and all that stuff, because they don't have the data in the first place. So if that's your setup, fix that first before you discuss using AI for case study generation with me, okay, just to give one example. And then last but not least, did we review all the risks? That's probably too much to ask. I don't know if someone can review all the risks, but from compliance issues to potential price hikes to, dare I say, geopolitical risks, right, would you build on a Chinese API today? Would you build on a US API? I don't know what it is, but so many firms and clients seem to be keen on making this a core enabler of their business workflows and processes. I think some rigorous vetting has to be done and honestly, we have engaged with a couple of very large corporations where that seems to be very much underway. Very tight governance, strict rules, you're not allowed to bring your AI recording thing to calls, they built their own in house models and stuff has to run on that. So smart players are already doing this and it's a big job and I understand the idea of let's just figure stuff out entrepreneurially and I think that's fine, but before you go from test to scale, I think that risk perspective is really essential to cover. Maybe that's my German heritage and I'm a negative fancier, don't know, but those would be my four. There's probably many more, but those came to my mind. Nice. Awesome, Fleur. Well, that was a great summary of AI as a strategy and thank you for the time there. I wanna take just a minute here to remind anybody on the line, if they wanna have a question, please enter your question into the Q and A panel on the screen. And we only have a couple minutes left. We'll try to get to a couple of questions here. But for the ones that we don't, we're going to make sure that we follow-up afterwards. So we won't leave you hanging. But let me go through a couple of the Q and A here and maybe pick out a couple of the first ones that we got flow and I'll pose, I'll pose these questions to you. So why don't we start with this one? Number four. How can one best implement the balanced use of AI with the need to still engage with clients and prospects on a personal basis to find out what's really going on inside their minds? So if I read that correctly is how close to surface do you want the AI to be in relationship? And I think it always works better if the client gets a great experience around diagnosis and then judgment and results delivery and then the subsequent support and implementation. This is very generic, realize because it will be different across industries, but my point is, the more the client is exposed to how the sausage gets made, the closer you are to a commoditized model where you get paid by the hour and I mean, that has been not the best place to be in even without AI, right? If you just do software development on a specific topic, see this all the time, right? People are not wanting to pay, I don't know, US rates or Germany rates for development work, because they know that they can find someone in Eastern Europe, Asia, whatever, to do it for cheap and I think that the same logic holds true. So I would still design the experience around engaging the clients personally and then having whatever AI you're using in your delivery system in the backend running. Obviously, the exception there is if your solution or service or product has to do with AI somewhat, yeah, I would showcase that probably on your website. Maybe your website should be a chatbot experience. If you're building chatbots, maybe your website should be a scroll through flat text website, but it is a chatbot that sort of I converse with to learn about you or whatever, that's a bad example probably. But that would be the exception, but for typical services businesses, keep the back office in the back of the office. I don't know if that makes sense as an answer. Yeah, that makes sense. I'm going to go to, Jimmy asks, what book, so I'll expand to see, he asked what books are there to keep track of the changes in AI on topics like data compute scaling, secure, agentic, physical AI and open models. And so besides books, but I would just maybe post what watering holes you go to for, you know, learn industry wise for your expertise. In keeping with my story from before, I very little of it, so I'm close enough or interested enough in technology to share any book types of that version. What I do is I about, I try to think about in our business, the problems we solve and the process along which we do that. And then sometimes I have occasional thoughts of, hey, couldn't a machine do that? And then honestly, me, the sources are a bit more informal. So it's a lot of friends and network in that space. Mean, that just through the nature of my work and know lots of data science or machine learning people. And believe it or not, we get lots of pitches from startups who build stuff that might be relevant in our space. So that piques my curiosity, that's another area I go to. So to summarize, I think the socials, a couple of IT publications you might want to follow. So we have Heise here in Germany, I don't know if that's an international, they're relatively technical. A couple of podcasts and then that's it, but again, I'm not close enough to the technology or I'm not tasked with large scale implementations, I'm not the best source for pronunciation. When it comes to strategy though, I think the visible authority would be something I would be following. Well, yeah, that is if you want to sharpen your answer to the, how do you focus on the business issue and then build from there. The front part of that, yeah, we can help with that, helping you then pick the tooling to run with it. You hit that. That's different, right. So let me just get to one or two more of these flows. So, we've got one number three here. How do we position AI as augmentation rather than a replacement? So, again, I feel like I'm repeating myself, there's not too much value here, but if you think of it as can the AI step in at the level of the process of consulting, so steps along the way to get to the result. I think that's always good. Is there a human in the loop, the expert? Because if it's just the machine doing something, we're talking about automation and that has the same value as the call bot, please press 1, all interact with fully automated systems and the value they have to us is very limited. The full automation I don't think is a good choice, but the expert that is just like consulting has always been, the senior consultant or the principal has a pyramid of juniors you never saw behind them to do some of the prep work. Well, part of that pyramid can nowadays be AI, if you're smart about it. So that's one part, the senior expert that gets augmented, I think, or leveraged, that's one part. And then the other part is, the more of the expertise in your firm, the more that you can formalize and feed into the systems, right? These are the frameworks we use, this is how we analyse the kickoff work. The more you can train the model in a specific way that is specific to how you work, the better it then becomes flipside augmenting your process instead of just doing generic stuff it could do for the next firm down the road. I don't know if that was the answer Yeah, to the that's perfect. Awesome. Well, think we're buttoned up against the time. So, I just wanted to say thank you very much for your time. This has been, insightful for me and I'm sure it was for the rest of the audience as well. And I want to remind everyone that the slides will be delivered following our presentation here. They will go out to everyone and we didn't get to some of the questions. Some of them had to do with DELTECH and VantagePoints particularly. So we will make sure we follow-up after this is done with each and every one of you and if there's a couple of the.