AI at Work: Beyond Productivity

Generative AI isn't merely accelerating workplace productivity—it's fundamentally reconstructing how value is created, expertise is defined, and organizations develop. My conversation with enterprise strategy expert Andreas Welsch reveals the profound shifts happening beneath the surface of today's AI revolution.

TLDR:

• 75-90% of C-level executives believe their companies approach AI strategically while only 33-50% of employees agree
• AI delivers up to 43% productivity gains and boosts work quality for 68% of users
• Employees can reach 60-80% expert-level performance in new tasks with AI assistance
• Less than 10% of companies have deployed Gen AI across five or more functions
• Successful AI scaling requires aligning with business strategy and measurable KPIs
• Viewing AI as a cybernetic teammate rather than just a tool changes how we implement governance
• The biggest risk for businesses is not adopting AI at all
• Start by auditing existing tech stack for AI features already available from vendors

We unpack the dangerous disconnect between C-suite confidence and workforce reality: while 75-90% of executives believe their AI approach is strategic, barely half their employees agree. Andreas exposes the critical misconception that "AI doesn't apply to our business," asserting instead that AI's relevance spans every function—the key lies in discovering where it creates meaningful value for your specific context.

The transformation extends beyond the impressive statistics (43% productivity gains, enhanced work quality for 68% of users) to something more fundamental: AI is democratizing expertise. Employees can now perform at near-expert levels in unfamiliar domains within days rather than the traditional "10,000 hours" of practice. This doesn't eliminate the need for deep expertise but fundamentally changes how we think about team composition and skill development.

Perhaps most fascinating is the conceptual shift from viewing AI as a tool to seeing it as a cybernetic teammate, particularly as agentic AI emerges. This perspective change demands new governance frameworks—Andreas suggests we might look to existing human resource practices rather than reinventing the wheel. The greatest risk for leaders isn't implementing AI poorly but failing to implement it at all, as competitors capture value and markets transform around them.

Ready to transform how your organization approaches AI? Discover practical strategies in Andreas Welsch's AI Leadership Handbook and learn how to align technology with your business strategy for measurable outcomes that drive real competitive advantage.

Links to Andreas content:

https://www.intelligence-briefing.com 

https://www.aileadershiphandbook.com 

https://www.linkedin.com/in/andreasmwelsch 

Podcast: “What’s the BUZZ?—AI in Busine

Chapters

• 0:04 AI's Transformative Impact on Work
• 0:56 CEOs' Misconceptions About AI Adoption
• 2:27 Real Productivity Gains and Expert Leveling
• 6:36 Moving Beyond Pilot Mode to Scale
• 8:22 AI as Cybernetic Teammate
• 11:09 AI Strategy for the Next 12 Months
• 13:52 The Biggest Risk Leaders Face
• 15:24 Episode Wrap and Key Takeaways

Transcript

Kieran Gilmurray: 0:00

Generative AI doesn't just boost productivity. It redefines value in the modern workplace like nothing before it. It's levelling skill gaps, reshaping expertise and upending how businesses grow. Today we unlock the upside and avoid the hidden risks. Hi, my name is Kieran. I'm a globally recognized authority on AI automation and digital transformation. My guest today, Andreas Welsh, is a leading expert in enterprise strategy, known for helping global businesses turn emerging tech into real business advantage. Andreas, great to have you here.

Andreas Welsch: 0:32

Kieran.

Kieran Gilmurray: 0:32

Thank you so much for having me Excited to be with you, andreas. We're going to jump straight in. Look 75% of knowledge workers now use Gen AI, often without company oversight because the companies ban them. Often without company oversight because the companies ban them. What's the biggest misconception of CEOs that they still have in regards to AI adoption?

Andreas Welsch: 0:56

I think one of the biggest misconceptions that I frequently hear is AI doesn't apply to our business. I think they couldn't be further from the truth, because AI does apply to any business, to any role in a business. It's just a matter of figuring out where does it add value in our particular business area, in my particular role? There's an interesting study that was conducted at the end of last year that was published in March of this year by a company called Writer and they found that between 75% to 90% of C-level executives think that their company is approaching AI actually strategically. That leads to another point, right Overconfidence. We think we're doing this actually really, really well those that are on the other side of the spectrum.

Andreas Welsch: 1:35

But it's actually turned out that employees said about a third to half of the employees that were surveyed said well, we're actually not doing so great with our adoption, we're not doing so great with our AI literacy. And just because you have some guy working on chat GPT in your company doesn't make it strategic, right. Just because your CIO is rolling out co-pilot to everyone in the company, again doesn't make it strategic. And just because you turn a blind eye on adoption also doesn't make it go away. So biggest misconception is it doesn't apply to our business or we're doing really, really well. I think the truth is somewhere in between, and it's all a matter of teasing out. Where are these individual things that we can already be doing without breaking the bank?

Kieran Gilmurray: 2:21

well, I'm starting to see stats you know like. Like Genai has delivered up to 43% productivity gains. It's boosted work quality for 68% of users. Well, are those figures real, and is this just a faster way of doing some or the same things we've always done?

Andreas Welsch: 2:41

Look. I'm convinced that they are real because people have always looked for ways to improve their work and how they get their work done faster and better. So to me those are absolutely valid stats. But does it mean just doing the same things faster? Maybe initially to get more accustomed to these tools, to warm up to how do I use these ai tools? How do I use them differently? How do I use, maybe, different tools in in my line of work in um becoming more proficient? But I think once you have reached that level of proficiency and comfort of how do I work with these different tools, then you can also think about well, what are the, the ways in in which I can use these tools on top of what I usually do? Right, do I change my workflow? Do I change my approach in bringing these tools? And so I think initially it's very well part of doing the same things faster. But then that trajectory of how can I do things differently opens up as well.

Kieran Gilmurray: 3:47

Are we also talking about doing things smarter? Because I've seen in one other study and there's no shortage of study, it seems, where you know it helps, some employees reach, you know, 60, 70, 80% expert level in terms of performance and tasks they'd never done before. How does that change how we actually think about expertise recruit, hire or develop?

Andreas Welsch: 4:11

So I looked at this topic at the end of last year and explored it as part of an article in my newsletter, the AI Memo, and I saw these studies too and I thought, well, that's interesting that even if you're not that proficient in a certain domain or area yet, that you can reach a higher level of proficiency or a similar level as experts very, very quickly just by using these tools. And just recently I experimented with using AI to generate code to build an app. There's this trend at the moment called vibe coding, where anybody can use AI and just enter a prompt in your native language and say build an application that does x, y, z. My case was a business simulation that I wanted to build, and I saw very, very quickly that, yes, it does make me a better coder because it writes code with within a matter of seconds, and minutes even. But I still need some level of expertise, especially if that thing hits a wall, If there's a roadblock, if I'm trying to improve it beyond what AI has given me.

Andreas Welsch: 5:16

And so I think, while it does help you perform at a higher level of proficiency initially, there's also still that idea of having to practice for thousands of hours. Right, Malcolm Gladwell said, you need to spend about 10,000 hours honing your craft before you become a real expert, and that's five years of full-time employment. So I think, yes, it does make you more proficient initially, but you will still need to continue practicing and you will need to ask good questions to those that are already in expert roles how are you doing this? Why are we doing this? What are the different angles and aspects? On the other hand, also, if you are an expert and if you have already a level of knowledge, it's another booster for your productivity right? So, wherever you are on your journey whether you're just starting out or you're already pretty proficient or an expert AI helps you perform a lot better and improves your productivity there.

Kieran Gilmurray: 6:14

Oh, I like the idea of a booster, but I've seen that less than 10% of companies have deployed Gen AI across, say, five or more functions. So how can we get more companies putting on the rocket boosters and moving them from pilot mode to scaling with confidence?

Andreas Welsch: 6:36

That's the age-old question, isn't it? It doesn't go away. We've seen this with machine learning, we've seen this with RPA, we've seen this with different technologies, and I think what's common across the adoption of all of those and maybe the slower adoption getting out of that pilot mode is that you actually need to align your projects and your initiatives to your business strategy. What is the largest strategy that we're trying to pursue and to execute over the next 12 to 36 months? How are we planning on growing revenue? Are we looking to cut costs and where does technology allow us to do that better in new ways?

Andreas Welsch: 7:16

What are the KPIs the key performance indicators? What are the PPIs the process performance indicators that we should be looking at and align our projects to? Because if we can make it measurable, we can also have a different conversation with our business stakeholder. All of a sudden. We're talking on eye level and we're helping them achieve the goals that they already have. So make it measurable, align it to your business strategy and then it really doesn't matter that much whether it's gen, ai, agentic AI, machine learning, rpa or any other kind of new technology and innovation that you're bringing to business, but starting with the technology first is like having a hammer in every problem is a nail.

Kieran Gilmurray: 8:00

I've seen lots of nails in my career that are people trying to whack them.

Andreas Welsch: 8:03

So you never know, this could change things.

Kieran Gilmurray: 8:06

Shifting gear. Andreas, you've spoken about AI acting more like a cybernetic teammate than a tool. What did you mean and what are the implications for team structure and decision-making authority inside of businesses?

Andreas Welsch: 8:22

So I've recently had lots of great conversations with experts and also with quite a number of evangelists on that topic of is AI to be seen as a teammate or is it just another kind of automation, another kind of software, another way of doing things like we've always done them in business? And I personally believe that if we adopt a mindset of AI, and especially this new topic of AI agents and agentic AI, if they become virtual employees, if they become digital employees, it shifts how we view them right. It shifts how we interact with them. All of a sudden, if I need to ask that agent for some information or to do some research for me, it's a different mindset. Also, how we put guardrails around these agents that can act just based on goals like go research, what are the?

Andreas Welsch: 9:15

The current market trends for mid-size manufacturing companies in in the southern parts of the us, for example?

Andreas Welsch: 9:22

All of a sudden, you know we we say here the guard rails. Here's what I want you to adhere you by, and similar to how we would guide our employees right, whether it's codes of conduct, ethical behavior, if you think of finance, accounting standards, ifrs. So if we apply these concepts, we also get to a point where we say well, actually, there are parts of the business that have already figured it out for human employees, right, human resources, for example, but where are they? They're not having a seat at the table. At the moment, it's technologists redefining what the future of work actually looks like, but again, we've already created codes of conduct, we've already implemented performance reviews, we have different roles and standards and descriptions. So why are we reinventing that and why is IT or technology function doing that? So I believe, despite all of that hype and promise, if we adopt that mindset of these components, these agents are similar to humans, then we should also think about how do we treat them like humans when it comes to these guardrails, when it comes to these standards.

Kieran Gilmurray: 10:32

So now we're talking about agentic AI then and not everybody will be familiar with that term, so I'll get you to explain it in a second, but let me phrase it within a question. So a lot's changing. You know we've had GPT, chat GPT since for years, but known since about November 2022, 2023. So, if I'm a CEO, how do I assess you, I assess my current strategy, how do I know not just what's happened in the last 12 months, but know where to put my bets when it comes to AI, gen AI, agentic AI, rpa or digital technology in the next 12 months? Things are moving fast. Where do I bet, andreas, to give me the best opportunity possible to lead my organization to some form of success?

Andreas Welsch: 11:20

I couldn't agree more. They're moving incredibly fast. I mean, we were just talking about this before we went live and we said there's just an overwhelming amount of information and news and announcements and product releases coming out on a weekly basis, let alone within a month or within a quarter. My recommendation usually is again, look at your business strategy. What are the big goals you want to hit in the next 12 to 36 months? How can technology help us achieve these goals? And then work your way backwards Again. You don't have to build everything from scratch. I think there was a big misconception during the machine learning hype six, seven years ago, where everybody tried to build a platform, everybody tried to build models and they realized well, we don't have the data we don't have good data.

Andreas Welsch: 12:05

We actually need a lot of data scientists. They're expensive, they're pretty rare in the market and then, when we put all of this together, maybe our priorities have changed. Our stakeholder has moved on. So what do we do? Again, you don't need to build everything from scratch. Now. With gen ai, generative ai, things have gotten a lot easier because these models are more plug and play and they are pretty powerful in themselves.

Andreas Welsch: 12:27

But my recommendation usually is look at the tech stack that you already have. What are applications that most of your company use, or maybe one business function predominantly uses? Do an audit. A lot of people like the word, but take an inventory or take stock of what applications do you use. Do these vendors have AI features built in and I would say most vendors have done that over the last two years Are we using them already?

Andreas Welsch: 12:53

If not, what's stopping us from using them? Are the features delivering meaningful, measurable value? Can we tie that to a KPI where we see this improves when we turn on this feature? Are there additional costs associated with it? Are those add-ons? Are those additional subscriptions, higher level tiers and the like? Is it worth the investment? But that will lead you to a path of success a lot more quickly compared to starting at the very foundational layer of do we have data? Do we have the right data? Do we have complete data? So you already leapfrog what what many companies are still trying to do. But look at where ai is built into your applications and how we can turn that on very easily and quickly well, let's flip the lens for a second then, because we're very pro-AI and geni-AI.

Kieran Gilmurray: 13:44

But what's the one risk CEOs are not talking about today, but really should be?

Andreas Welsch: 13:52

I would say that the risk is that they're not doing AI Again, that they think I'll put my head in the sand. This will go away. It's a fad that will pass, and I think it's not going away. It's not going away anytime soon. It's not going away anytime at all, because we see that you know, it's not just the startups or the big tech players talking about this. But now we are crossing that chasm, to speak in innovation management terms, where this early majority has tried it out. They are getting value.

Andreas Welsch: 14:25

Other people in the market are seeing oh yeah, well, it seems like there's something to it, so we better try this out too, and so I think there's just more momentum building on one hand. On the other, also, if you find these use cases, if you find these scenarios that you can align to your business strategy again, that makes it tangible and you have proof points in your business as well to build that momentum. So, again, to me, the risk is to say we're not doing AI, it does not apply to us, it is going away. I don't want to worry about this. The reality is many of your competitors competitors. They're not thinking like that. They are investing. They are looking at this because they see the opportunity. So it's here to stay and it's a matter of starting and figuring out what are the little things that we can already do without breaking the bank but then making meaningful difference in our business the bank but then make a meaningful difference in our business.

Kieran Gilmurray: 15:24

Wonderful, andreas, thank you so much for all of your insights. Uh, where can people follow your thinking and your latest content to get more of this wonderful, wonderful knowledge?

Andreas Welsch: 15:35

thanks so much, first of all, for the opportunity to to join you today. If you'd like to learn more about how you can turn technology hype into business outcomes, I would recommend pick a copy of my AI Leadership Handbook so you can learn about the nine aspects of making AI tangible in your business. You can also follow me on LinkedIn and get more information on our website, intelligence-briefingcom.

Kieran Gilmurray: 15:59

And I can highly recommend that book, having read it and bought it myself. Look, folks, today's conversation has made it clear AI isn't just about speeding up work. It's about redefining work, about who does it, how value is created and what CEO or leadership must or should look like. It's also about mindset, not just tools, that should shape your success. So, look folks, please follow along. I welcome you to future episodes of AI, Unscripted. Please, please, please, do buy that book. It's exceptional. Please do follow Andreas and I wish you every success in your AI, your agentic AI, your autonomous AI, your gen AI journey and everything else in between, Until next time. Thank you so much indeed. See you all soon.