Designed for decision-makers across financial services, the podcast focuses on practical strategy, operational resilience and enterprise outcomes. This first episode sets a clear tone: move beyond hype and focus on measurable business value.
AI has moved well beyond innovation labs and IT pilots. It now influences:
Yet many insurers remain caught between proof-of-concept success and enterprise-scale transformation. As organizations begin to adopt AI more broadly, the question remains: how to unlock measurable value from it.
It is easy to get lost in technical terminology like NLP, LLMs and agentic AI. But technology alone does not create value. A clear and strategic business application is needed. Darren Rudd, who leads CGI’s UK insurance consulting practice, frames AI in insurance through four practical archetypes.
This research emphasizes that experimentation is widespread, but integration is limited. Many insurers are layering AI onto existing workflows rather than redesigning those workflows. As Darren explains, “people are applying AI to what they do today rather than using it as an opportunity to rethink the process.”
History offers a useful analogy. When factories first adopted electricity, leaders replaced steam engines with electric motors but kept the same layout. Productivity improved only when factories were redesigned around electricity. AI requires a similar shift. Real value emerges when organizations rethink operating models, not just automate legacy processes.
Even with strong investment, enterprise-wide transformation often stalls. Darren lists three common barriers.
Will insurance look fundamentally different by 2030? Thomas predicts that the more likely scenario is progressive evolution rather than sudden disruption. However, change will not occur evenly.
In these areas, hybrid AI-human models will remain essential. AI will reshape the value chain, but at different speeds depending on regulation, trust and risk tolerance.
The defining shift for insurers is structural. It will reflect a move from isolated pilots, innovation lab experiments and point solutions to AI embedded in the operating model, enterprise-wide orchestration and measurable ROI aligned to strategic objectives. Organizations that treat AI as a bolt-on tool will see incremental gains. Those that integrate AI into core workflows, governance and talent strategies will achieve sustainable competitive advantage.
While investment levels are high, enterprise transformation is not. Value emerges when AI initiatives align to measurable outcomes such as cost efficiency, risk reduction, growth enablement and client experience improvement.
Because AI transformation is human-centered, the most significant constraints are:
Functions that rely heavily on unstructured data will transform quickly. High-touch, trust-based segments will evolve through blended AI and human decision-making. AI may not trigger an overnight reinvention of insurance. However, sustained redesign of processes, intelligence layers and decision frameworks will compound over time. As Thomas explains,
Join us for our next episode, where we take a deeper dive into how financial services organizations can transition from legacy systems to a future defined by innovation and intelligence.
- Chapter 1: Moving from experimentation to enterprise value
-
Thomas Rauschen:
Every day we hear about modern technology and AI, from market movers and cybersecurity risks to accelerated value creation and social impacts. The AI conversation is everywhere, in the media, in boardrooms and client conversations. But how is it shaping the insurance industry? And what is the balance between hype, myth, and reality?
Hello everybody, I'm Thomas Rauschen and I lead CGI's global insurance industry. In our new podcast series, AI Uncovered, The Future of Insurance, we talk about the rapid business and technology transformation the industry is facing and the endless opportunities of modern technology and AI.
In today's episode, we will discuss if AI and insurance is an evolution or revolution. To shine a light on this question and setting us up for future episodes, I'm thrilled today to be joined by Darren Rudd, who's having this conversation with his clients every day. Darren leads our insurance consulting practice at CGI in the UK. And again, I'm really pleased to have you on board here, Darren, to our podcast. Welcome.
Darren Rudd:
Thanks very much, Thomas. Good to be here.
- Chapter 2: Reframing AI around business outcomes
-
Thomas Rauschen:
Great. Look, before we unpack the topic, let's discuss a little bit the different terms that are that we are hearing in our client conversations, in the boardroom, in the newspaper, you know, AI is everywhere. And you know, we talk about machine learning, and NLP, LLP, GenAI, agentic AI. There are so many terms out there, but sometimes we wonder what do these different terms all mean? And it would be great, Darren, if you could help us really shine a light on these terms and level-set a little bit here.
Dareen Rudd:
Of course. I think, actually, when I'm talking to people, I think the most useful way to think about it is rather than worry about the individual technologies, because they tend to sort of be bundled together, if we frame it more in terms of how we can apply it to make a difference. So, I tend to break it into four areas. The first one is typically around the GenAI, LLM-type of large language model piece. How can you apply it to a point solution to make a difference in terms of being more operationally efficient? So, how can you do things internally to do things more effectively?
The second way of looking at it is how do I add these types of AI, GenAI capabilities into products that I use and services that I offer back out to my customers?
The third one is how do I use AI to gain more insights? That tends to be more of the traditional machine learning, the analytics side of it.
And then the final one is the more stuff that we're talking about a little bit more at the moment, the agentic AI, which I'd say is more about how you join different components together to start running a process overall with a little more decision intelligence. So, I think if we break it into those four, that's probably more useful than trying to get into the specifics of the difference between the individual technology. At least when I'm talking to the business, that's where I find it most useful.
- Chapter 3: Bridging the strategy gap through integration
-
Thomas Rauschen:
That is really helpful. And it seems to be on your description, there's obviously a lot of opportunities around AI in the market, and that obviously is reflected in our latest Voice of Our Client survey, where, obviously, it's shown that digitalization remains the top global market trend across the insurance industry.
And before we dive into my next question, I just want to want to give you a stat from the survey that I just mentioned. It looks like that 46% of insurance executives are confident in producing the expected outcomes from their digital strategies, which is great, but only 43% of organizations have an enterprise-wide and comprehensive AI strategy in place. That sounds to me that yes, there's movement across AI, across the modernization and digitalization journey, but equally it also sounds like that is not comprehensive across a whole enterprise, but rather focused on certain areas.
So, we see a lot of companies are experimenting, testing AI, deploying it in certain areas. But maybe you can give us a view of what you are seeing in the insurance industry today and probably in the next two to three years.
Darren Rudd:
Yeah, good question. I think part of it comes to the point solutions and the experimentation is more on that how do we become more efficient—so, the GenAI side of it. I think insurance, in particular, has been using a lot of the other AI, the insight side of it. I talked about, how do I understand the data I have and what's going on for a long time now. So, I think there's got to be careful when we start to break it down in terms of which AI we're using.
But I like to use the analogy, I think, where we are now and where we're going to be, from something I read, which I thought was a really good way of talking about it. Factories originally were driven by steam engines, and that meant you had a big steam engine in the middle of a factory with a big crankshaft running all of the machines. When electricity was discovered and they started swapping the steam engine out of the electric engine, they kind of left the factory where it was, the big crankshaft in the middle, and all the machines stayed there. Took them years and years to work out well, hang on a second, I can make a smaller electric engine, move my machines around, get them off the crankshaft, and be more efficient. And I think that's where we are with applying AI, at least now. People are applying it to what they do today, rather than using it as an opportunity to rethink.
So, I think we're seeing the point solutions and the smaller POCs, people are saying, how can I use it to do what I do today? I think people have to start thinking about how I could do it differently using this technology. And I do see some of that emerging and new ways of thinking about it, but I think it's going yo take a little bit of time to drive out.
Thomas Rauschen:
That's perfect. I'd like to challenge you a bit on your observations, though. Right now, it feels like many organizations are deploying AI primarily from an efficiency perspective, essentially taking cost out, you know, in claims, in underwriting, sales, etc. However, we also see a lot of advances today in AI and data and getting more process insights. So, the opportunity is increasingly about making smarter decisions. So, is an efficiency-driven mindset really enough to take organizations to the next level, or do you think they need to shift towards a more intelligence-driven modernization?
Darren Rudd:
And it's a good challenge. I think it's like all things. There are lots of streams running. If we can make ourselves more operationally efficient, take some of the drudgery and the manual work out, that immediately frees up some money to do other things, maybe to do that intelligent refactoring. But I do think there are examples now of insurance using AI to get their underwriters, for example, to the right data faster, which then makes them more efficient, but from a top-line point of view, making better decisions. So, I think there's a mix of the two, and I think they all go together.
But I come back to that sort of analogy of the crankshaft going through the factory, we're still moving the machines around, or even deciding where to move them. So, let's make ourselves more efficient, free up some value to enable us to actually do more. And if my competitors are doing that, then I should probably be looking at how I use that, as well. Would be my challenge in terms of how I'm seeing it right now.
- Chapter 4: Three reasons why enterprise-wide AI adoption slows
-
Thomas Rauschen:
That makes sense. And if we dive a little bit deeper into what we are seeing today and again in the next few years, you know, what is stopping organizations from getting most of the AI and the opportunity that is attached to AI? I mean, obviously, what we hear in our Voice of Our Clients survey and we hear it in our client conversations, yes, we talk about legacy systems and we talk about data quality a lot, right? But what is beyond those topics? You know, what do companies need to focus on in order to make it real, powerful, and do AI transformation right?
Darren Rudd:
Yeah, I think the tech industry has been its own worst enemy in part, in that when I talk to CIOs in particular, they've been here lots of times before in a hype cycle where everybody's saying this amazing new tech is going to change the world. And they tend to be the ones that are asked to implement it without a lot of thought because everybody's demanding we need to be seen to be doing it, and then they're the same ones who then have to unpick it again in a year or two's time when it hasn't really met the hype. So, I think you get a level of tech resistance, people are excited about it, but the senior leaders are thinking, “It feels like I've been here before, and maybe we're not going to hit the hype.” We've then got the business side of it where I think again at the senior level, there's been lots of technology failures, and the new tech coming in has not met the hype, and we are in a massive hype cycle around how far you could go with AI, and particularly the GenAI side of it. I think there's lots of value that we can get, but it's got to be aimed at the right place rather than where the hype is. So, I think we've got business leaders as well going, “Hmm, I'm not actually seeing the benefits you're telling me I'm getting.”
Even now, we know that there's a gap between expectation and reality. So, I think that's a big friction point. And then the other bit, and I know we've chatted about it before, is you know, this tech can feel quite scary and it's changing the way people could work, need to work, and potentially people's roles. So, the individuals who are being AI'd or their roles and the processes again are going to be sitting there going, “Whoa.” So, I think the other area that's stopping it is people just sitting there going, you know, “explain to me how this is going to work and what does this mean for me?”
That's a really important part, I think, in terms of making this a success, more so than some of the, you know, moving to cloud didn't really change the way you did your job. This type of stuff potentially can.
Thomas Rauschen:
But that sounds to me, Darren, that AI or AI technology shouldn't be just a bolt-on to your technology architecture. I think it also what you mentioned leads to that it's not a technology transformation and initiative, it needs to be across the whole organization, which includes people and change. When it comes to how to do it right, you know, like for example, when we talk about transformation initiatives across organizations, right?
We always say from a consulting perspective, it's all great, but please make sure that you do like cybersecurity, operational resilience, you know, change management, really have it embedded into your transformation initiatives or programs. Can we make the same case for AI? Whatever transformation initiatives or modernization initiatives you run in your organization, AI should be or needs to be a layer in your project.
Darren Rudd:
Yes, that's a really interesting bit. And I don't always get everybody loving what I say on this side of it.
Thomas Raushen:
That’s okay!
Darren Rudd:
I think a lot of the failures that we've seen over the past, and you know the numbers show it. We know that the common metric is 70% of those big digital transformation, tech-led transformations fail to deliver the value that expected to do. And I've been around doing this for you know 30 odd years now, and you see the same patterns emerging. If we go in it from a tech-led point of view and we haven't really thought about the why or what we're trying to do, then we tend to get failure. It has to be thought of holistically, and I worry, particularly with AI, that we're chasing the silver bullet that's going solve all my gnarly business problems, where actually I really need to be starting with which problem am I trying to solve first, and then which sets of technology, rather than a single technology, will make that difference.
And it isn't just tech, as you've said. It's around, how am I going to bring the people on, how am I going to make that process change done, how am I even going to think about it from a data point of view and where I get that. So, I think if you're not looking at it holistically and understanding the problem you're trying to solve and who for, and whether they even want you to solve that problem for them, then you're going to fail, regardless of which technology.
I think AI may accelerate the impact or the damage you do if you don't get it right. And there's so much expectation that it's going to make a difference. I worry that this is going to be an even bigger lack of expectation at the end of it. But I know I'm a little bit contrarian there in terms of not everybody agrees with me, but hey, we all have opinions.
- Chapter 5: Predicting the future of insurance, 2030 and beyond
-
Thomas Rauschen:
Yeah, but there's also no right or wrong answer, right? So, everyone has different opinions. But I really do like your summary. Before we wrap up the podcast, let's get our crystal balls out here for a little bit, yours and mine, and think about what might the future of the insurance industry look like in 2030 and beyond, considering obviously AI. What is your prediction? Again, there's no right or wrong answer, right?
Darren Rudd:
Yeah, crystal ball gazing is always an interesting one. I actually think it comes back to the whole thread of the conversation we're having, how it's going look 2030 and beyond, like over the next four or five years, is really going to be dependent on how quickly we move from the hype expectations to a reality of how to use this new technology in the right way, and then how we prepare for whatever's coming behind it, because there are other technologies that sort of been drowned out a little bit in the AI space that are likely to be more effective than some of the current technologies, particularly in you know doing more of the human-led work rather than the summarization and the things that it does today.
And I also think though it comes down to how our business leaders and our technology leaders choose to re-engineer the business. Coming back to that steam-driven factory, if we leave the crankshaft in place and all the machines sitting there in a row, then I'd say in 2030, regardless of if you've stuck a bit of AI around it, it's going look very similar to what we do today. Maybe with just some more people pulling their hair out when their AI at all doesn't do what they expect it to do. That would be my slightly more pessimistic view. I'm quite optimistic, but I'd say if we don't get those things right, it won't look a lot different. And to be honest, I started in the industry in ‘96, and sometimes it doesn't feel like we've moved an awful lot forwards, at times.
Thomas Rauschen:
Yeah, I have a little bit more, almost like a progressive or a more aggressive view. You might spin also the case that in 2030 and beyond, you know, your operating model would be 50-60% more linear, you know, you're perfectly well integrated into all the other ecosystems that you work with, spear travel agency, car rentals, etc., changing data in real time, you know, but that's what I mean. There's like so many opinions, right? But we can only drive it forward based on our experience.
But if we step back and look at everything that we've discussed today, Darren, for me the message is actually quite simple. It seems that AI in insurance isn't short on momentum, investment, as you mentioned, also ideas. What it is short on sometimes, at least what we see in the industry is coherence.
We see a lot of experimentation, on point and solutions, a lot of optimism, of course, as well. But the real value probably will only show if we integrate AI not as a tool or new technology but really embed it in our operating model. And I think the opportunity is there, but I also do believe that it requires leaders to think differently beyond pilots, beyond technology, and beyond the constraints of how insurance has always worked. So, over the next few years, I would agree with you. We will probably see more like an evolution than a revolution. And yeah, that's my summary. Would you agree with my closing argument?
Darren Rudd:
Yeah, I do. I think, particularly at scale, I do think there are those organizations, I can see them now starting to drive innovation through using AI, but they're the ones that are willing to sort of rethink how they're doing stuff. But I don't know whether that's necessarily going to hit everywhere at scale. But I think there will be examples of that evolutionary side and revolutionary side across the pitch.
- Chapter 6: Key takeaways for insurance leaders: Embedding AI in enterprise strategy
-
Thomas Rauschen:
Perfect. Before we close the podcast, let's summarize for our listeners the three key takeaways. So, the first one, of course, AI momentum is real, but value creation is lagging. So that's what we are seeing across the industry, but we are getting there as an evolution, as we discussed. The second one: the real constraints are not algorithms or technology, but it's the foundational aspects of AI and the behavior that we need to instill across organizations. The third point will be, AI will reshape the insurance value chain or segments, but it will be unevenly. There will be areas where we see a rapid disruption, be it in distribution, software and development, or unstructured data, but equally a little bit more, maybe a slower or more pointed disruption in other segments. If you go to commercial insurance or re-insurance, where this is really a trust-driven business, I think it will evolve more gradually.
So, these are the three takeaways from today's podcast. Thank you, Darren, for joining in today, and of course, thank you to our listeners as well. We, of course, look forward to our next episode, where we discuss in a little bit more detail how insurance companies can move from legacy to innovation and intelligence. Thank you, Darren.
Darren Rudd:
Thanks very much, Thomas. Great chat.
