AI Returns: Separating Value from Hype

Transcript

AI Returns: Separating Value from Hype

The most important things that happen in the history of our species is technological progress.
No AI project should get to the end of 12 months and someone says, "Oh, well, that hasn't worked. It's all about tracking it, modifying it, changing it, adjusting it all the way through until it does work and starts to deliver benefits.
AI doesn't necessarily remove people. It changes where you can use people to create value. And that value, is around how do you drive a differentiated customer experience.

The companies who get it right are going to solve the answer to the question, if you could automate everything in your life, where would you stop?

For the past few years, the corporate world has been boldly surfing the initial wave of AI excitement. Boardrooms worldwide have poured hundreds of billions of dollars into artificial intelligence fueled by grand promises of economic revolution. We were told productivity would skyrocket, costs would vanish and businesses would effortlessly scale. But as the fiscal years roll over, executives are searching for the next wave of provable returns and exploring what they will need to do to catch them and surf them to the beach of productivity gains. There's a multitude of data out there tracking the hype versus reality. A year ago, MIT's the Gen AI Divide State of AI in Business 2025 report found a 95% failure rate for enterprise generative AI projects, defined as not having shown measurable financial returns within six months. With the speed of AI evolution, what does it look like now?
According to McKinsey's, the state of organization's 2026 report, while nearly 88% of organisations are actively experimenting with generative tools, still over 80% have yet to see any meaningful bottom line or EBIT gains. While the figure is still high, it shows there has been improvement and there are opportunities to focus on.

The challenge for this next generation of technology, specifically autonomous Agentic AI, is to prove it can deliver measurable, repeatable business value at scale, but unlocking the value requires a total architectural overhaul. It means completely re-engineering the internal human workforce and ultimately altering how the customer experiences an organisation from the outside.
Welcome to the Next Five Podcast. I'm Tom Parker and today we're cutting through the AI noise to find measurable business value. In an episode we are calling AI returns, separating value from hype. Joining me on this journey are three leaders at the very heart of this topic.
First, we have Giles Brian, general manager for CX at NICE. Giles, welcome to the show.
Pleased to be here with you, Tom.
Next is Chris Herbert, customer service director at Openreach. Chris, thanks for joining us.
Thanks for the invite, Tom.
And finally, Zach Kass, author, podcaster, and former OpenAI executive. Zach, it's a pleasure.
Likewise, thanks.
Well, Zach, let's start with you. Having been on the inside of the AI frontier, you've watched this wave build from the ground up. As mentioned at the top of the show, McKinsey put an 80% failure rate for enterprise generative AI initiatives when measured by short-term financial returns. On the question of cost, PWC's 2026 global CEO survey showed that only a quarter of CEOs say costs have decreased due to AI, while 22% report an increase. More than half, about 56% say their company has seen neither higher revenues nor lower costs from AI, while only one in eight, 12% report both of these positive impacts. What's behind these early stage results, Zach? Is the issue technological, operational, or is it simply that returns are harder to measure on such a nascent technology?
In the avoidance of saying obvious things, I think we should call it a couple of problems. The first is people got so excited about AI that they were willing to do non-economic things that would obviously skew a lot of the return on investments that we would've normally expected. The second problem is no one knows how this technology works and it's moving so quickly that the experts aren't in the room. Everyone is trying to figure this out in real time. So it's a bit inevitable when you have companies willing to spend any amount of money to get ahead and employees who don't actually know how the technology works that you'll have a gross overspending and we've seen that. But the third issue I think is a principle one, which is that companies don't actually agree often internally on what they should build and where they should go.

And when you actually look under the hood as often as I have, you discover that companies are not actually pointing employees at a North Star anymore. Employees are working individually on small KPIs that don't actually always point to a broader goal or outcome. And so it actually represents, I think, a broader corporate failure right now, which is a lack of courage to describe what could happen. Ultimately, the best companies we're going to see are going to start to make this technology invisible. Requiring every employee to click around and do a bunch of their own work inside of a technology hub was never going to be the panacea. The best technology becomes infrastructure.
Chiles, global AI spending is surging toward the trillions. Yet in 2025, over 70% of organisations were breaking even or losing money on these deployments. I don't want to fixate on those costs. I want to look at the organisational part of this. How do you, from an organisational standpoint, integrate AI to yield real measurable business value? And also, how long do you give or indeed have to wait as an organisation to determine whether an AI project has been a success? Is it six months? Is that fair? Is it a year? Is it two?
Okay. So for an AI project to deliver rapid immediate benefits by which we mean OPEX reduction and within a year, revenue benefits for the year, two things need to be true for the very fastest. The first is obviously that you apply the AI to something suitable, something applicable. The second, that's obvious, of course. The second is less obvious. It needs to be self-contained. It needs to be something that doesn't have upstream or downstream impacts that erode the benefits. Examples exist in our industry, for example, in the software industry, something like software testing, very straightforwards. In our client's business, something like the application of AI to inbound phone calls, another place where it works. Other issues are more complex. For example, software engineering in general, there's a question there about, are you creating the right product for the right market? Are there going to be deployment issues?
There are issues and challenges there. So essentially, when we come to the timing question, it depends upon what sort of AI project you're doing. If you are focused on or dealing with something that's self-contained, highly applicable, you should be seeing benefits inside three months, six months. If it's something that's more complex that involves a whole organisation before the cash actually starts to reverse direction, you're seeing revenue, you're seeing benefits, that could take up to a year. But something you said was interesting. It's not about waiting. Nothing in AI deployment is about waiting. Everything in my experience requires work. It requires design. It requires iteration. Everything requires you to pay attention. No AI project should get to the end of 12 months and someone says, "Oh, well, that hasn't worked. It's all about tracking it, modifying it, changing it, adjusting it all the way through until it does work and starts to deliver benefits.
And also being very clear whether you're focusing on something simple and self-contained or something that has across organisation implications.
Yeah. Chris, Giles said there it requires work. Enterprise value is heavily tied to how smooth an organisation runs. The PWC study showed that 12% are seeing positive returns. What is the most important enabler for companies turning AI capabilities into real bottom line efficiency? And what are the leaders doing that sets them apart?
Yeah, so Zach said a couple of points, and I'll paraphrase a little bit. Perseverance. Thinking differently. So I think there's some leaders out there that are looking to apply AI to get instant results. And when those results don't come instantaneously, then there's a tendency to, when it didn't work, let's try something different. So the first thing first is start with the high value business case, high volumes. That's exactly what we did in Oprah each. We took the biggest volume user case first. It wasn't the most complex, wasn't the most simple. It's something that had big volume that we could test really quickly and understand whether the implementation of our AI was working, delivering business value. If not, then tweaking it back to Giles' point, how do you tweak, iterate, learn, adapt. Once you've got that in that proof point, then I think it's about how do you rapidly scale across the business.

Now to do that, I think it's about the ability to articulate the business value, not just in terms of pounds, shill-ins and pents, but also in terms of the output for the customer experience. What does it mean for your customers and the outcomes? And that might not be financial to start with. And that's a big bold move that I think we in Oprah each are certainly starting to push towards now, which is customer experience. So I think for me to summarise, it's probably about innovation. So thinking slightly differently to solving existing challenges, but then perseverance right across the organisation.
In the mid '90s, Jackie Fenn, a prominent technology analyst, created the Gartner Hypecycle framework. It chants the pattern of overenthusiasm, scepticism, and eventual realistic adoption that almost all major technological breakthroughs experience. One such stage coined by Fenn is the trough of disillusionment. I love this. The precise moment where inflated expectations crash into operational reality. Some companies will give up, but those who push through could reach the next stage, the slope of enlightenment. Feels like I'm sort of running through a field at the moment, but the slope of enlightenment is what it's called where true benefits emerge and more realistic test cases solidify and then hopefully we find themselves at the plateau of productivity where widespread mainstream adoption and real world value is proven. This is a question to all of you. Do you think we are in the trough of disillusionment and how do sea levels and boards push through these early fears not give up, persevere, and wait for the longer future returns?
Zach, let's start with you.
The problem with describing that framework is that it doesn't account for all the unique elements of AI. In principle, what we call the jagged perimeter. So whereas with many prior technologies, it was very clear what the technology was very good at and what it wasn't good at and it could apply very specifically to an industry that's not the case with AI. So saying even in the case of life sciences, well, AI is not good at solving this problem, but it is good at solving this problem. So right now somewhere on earth, someone is improving their ability to identify a new target. Someone actually recently improved their ability to improve a CRISPR output. If you are in life sciences, you're not going through a trough of disillusionment with AI. Now the promises that have been made to a bunch of people, the public perception I actually think is going to drag further.
And a lot of this has to do with very specific relationship between humans and technology, which I think has much more to do with the device and the display than it does AI itself. But I do think it's quite likely that some industries will actually say this doesn't make a huge difference, hospitality, food and beverage. In fact, they might discover that AI complicates their relationship with the customer to such a degree that they want to back away. Whereas other industries are never going to go through a trough of disillusionment, they're going to enter this new enlightenment phase. Other industries are going to be born entirely because of this technology.
Absolutely. Giles.
So everyone loves the description, soap of enlightenment, trough of delusion. They're wonderful concepts, but they are just concepts and it's a huge generalisation and it's not as if a whole technology for all the people experience that technology go through their stages at the same time. So for any organisation, it depends upon the decisions they make, how smart they are, the partners they choose, how robust they are, how much they persevere, where keeps coming up. It depends upon how they behave with regard to the technology, how they enact it. And so what I'm saying to summarise there is one business's trough of disillusionment is another business's time to roll up its sleeves, get involved and make the stuff work.
Chris.
I mean, Tom, I love the description there that you gave of this. I think expectations are now becoming more realistic. And I think for me, the challenge that organisations face is probably the application of it against those use cases that I spoke about. Maybe going after perhaps the wrong thing to start with, you then lack confidence because it didn't quite work. Now, I don't think that's the technology per se. I think it's the ambition and the use case choice of the individual businesses. I think the key thing is I think as a businesses, I think we're understanding the linkage between proactive AI agents and human agents now working together, humans and AI working together. I think that is now starting to help break down some of the barriers within the organisation in that the nervousness of is AI going to take my job, for example.
So for us in Operach, we're now looking at how do we take away the heavy lifting to allow our colleagues to focus on the right things for our customers to deliver the greatest outcome almost in unison with AI.
We've talked about timeframes before about wondering or finding out whether it's delivering value in AI projects. We've just talked about Jackie Fenn's hype cycle. Well, recently, Fenn said that AI's 18 to 36 month average timeframe to ROI overlaps the trough of disillusionment, causing a gap between companies who abandon at the trough and fall behind compared to those who stay committed and realise the benefits. Do you believe that in 2026 and 2027 there will be the largest gap we've ever seen between AI leaders and AI laggards? But Zach?
The largest gap yet? Sure. That seems reasonable. I think you can imagine that there will be companies that exist in 2026 that don't look like they should exist for much longer and that by 2028, the market will have sort of sorted itself out between companies that truly don't want to be left behind and companies that do. But I also think you should safely assume that a lot of companies are going to differentiate on the basis of how they use it in sort of new and actually quite surprising ways. Giles.
So I'm going to pick up something that Zach said earlier about choosing the wrong use case is very easy to think about AI as being just for cost savings and cost benefit. Actually, if you think about AI for a business has been primarily about getting and attracting customers, keeping customers, getting value, getting lifetime value for customers, reducing churn, that's actually where you'll see the biggest benefit related to that. Directly to your question, yes, the gap is the largest you'll ever see at this time, but actually the gap will extend because the people that apply AI to the fundamental reshaping of their business and the relationship with their customers, those will be absolute winners and others will be losers.
Chris.
I think that to answer your question, I think yes, and I think we're already seeing that, right? Both to the points of Zach and Giles here, I'll go back to the point I said earlier. I think the difference is some leaders, some businesses again stuck in trying to prove the nth degree in a pilot phase or a test phase, trying to perfect everything before then going with scale, which then enables them to unlock the business benefit. And I totally agree with you, Giles. It's yes, the financial benefits are super important. Of course they are, but actually we are now looking at how do we drive a differentiated customer experience on the back of the capability that we can deploy and we have deployed by the way. And then how do you then compliment that? Again, go back to my point with our colleagues to take the real complex things that you know what AI may not ever be able to handle or solve efficiently.
So yeah, I think we're seeing it already and for me it goes back to get the big high value user case, prove the benefit fast, and then start to scale so that you can build the confidence in the organisations. Well,
I want to focus on how you work with colleagues because I want to look now at how you redefine the workforce. AI agents and human employees need to collaborate, obviously, redefining capability requirements and building human engagement with tech. The upside, according to McKinsey, is that 55% of leaders say successfully building AI capabilities of employees will bring exponential value. That means the human workforce has to be completely re-architected. When we automate the routine administrative tasks, we often talk about liberating the employee, but how do you actually manage that transition? What does an organisation look like when its teams shift away from execution and entirely toward high stakes judgement and relationship management? Charles, this is to you.
Sure. So liberating people from execution from administration and basically moving them up upskilling obviously requires the upskilling process to take place. Well, the fundamental thing you're talking about is a shift in productivity. So let's take nice, for example, three billion in revenue, 10,000 people, an average revenue per staff member of 300,000, respectable for our industry, but plenty of room for improvement. So what we're talking about essentially is generating far more revenue with the similar sort of people without growing their headcount and effectively that makes our company much more valuable, much more effective. What you have to do in order to make that happen is to upskill people, train them, upskill them, and of course give them the tools and make sure they have the best tools available to deliver their work.
Zach, you've argued that AI shouldn't just replace human labour but elevate it and expand human potential. You've said that closing the adoption gap could spawn a new era of human progress. Can you expand on your argument about AI's potential impact on human progress? And secondly, how do leaders build trust with their internal teams so they see AI as a collaborative catalyst rather than a threat to their job security?
The history of technology is the story of human progress on the whole. I mean, outside of the good guys winning the big wars, the most important things that happen in the history of our species is technological progress. So it's not hard to argue that technology is how we enable people to do more and therefore offer more people the opportunity for a great life. This particular moment presents new uncertainty, especially with respect to jobs. There are two separate questions. The first is around how should AI enable a better life for the average person. Very simply, it should deliver new scientific discoveries and it should drive down the cost of living. Housing, healthcare, and education and the pillars of a great society should get less expensive because of technological progress and automation. And some of these are policy problems and some of these are technology problems, to put it bluntly.
Trust therefore comes from the individual and their relationship with their friends, family, and employer and the individual and the relationship with society. And if your employer is describing a world where the shareholders exclusively benefit or where the margins improve, but they aren't describing an improved North Star, a castle on the hill, then the trust is going to wane. And similarly, if the average person doesn't observe that AI is improving their ability to provide a better life for their friends and family, but is in fact just making them more addicted to the device or more consumed by media, they're not going to be happy. And I actually think this will end up being a fairly booly and outcome over time. We will either observe AI like we do the mobile device, something that actually sort of destroyed communities in some respects, or we'll observe it quickly like the internet or electricity, something that enabled a major shift in global fabric.
Chris, this workforce redefinition that we're discussing has a direct line to your domain. When frontline customer-facing employees are freed from repetitive back office ticketing or manual data entry, what happens to their capacity and how do you effectively re-skill a workforce to focus purely on complex value creation?
Yeah. So I think AI doesn't necessarily remove people. It changes where you can use people to create value. And that value, again, I'll go back to what I've said in the previous answers is around how do you drive a differentiated customer experience. So our people within Operach, we are freeing them up more and more from doing the repeatable activities through some of our proactive AI deployments now. And I've got those people looking at the more complex cases. So where you really need that human power to go and go, right, what do I need to do for this customer? In openreach, we've driven our trustpilot scores from 1.6 to 4.6. And as you know, 4.6 is rated excellent and 1.6 was rated bad. So we've gone on a real big journey. But what I want to do is how do I get to five? How do I get from 4.6 to five?
And I think the secret to that is using the capacity in the organisation that we now have to really crack through those complex, challenging problems that we need to solve much better for our customers going forward.
Chris, I want to stay with you on this one because it is really important we look outside an organisation, how organisation react with their customers. In an era where products, software, and services are becoming heavily commoditized by technology, customer exists is emerging clearly as a primary source of competitive advantage. That 4.6 that you mentioned getting to five is going to be a competitive advantage. How can organisations use AI to deliver hyper personalised proactive journeys on the backend while keeping every interaction feeling genuinely human and connected?
Yeah, look, I think the biggest frustration, whether you're in telecommunications or retail is not knowing when something goes wrong or even before that, when something doesn't go wrong, not knowing what's next in the journey. And our business operates at scale. We do millions of interactions with our customers every single year and whilst the majority of those interactions go well, some don't go so well. So for me, it's about how do we leverage the capability to keep the customer informed, keep the customer connected to what's going on, give them a really clear idea of, okay, what's happened? Why has it happened? What am I going to do to put it right and when am I going to put it right? Now, to do that at scale, I'd need to hire many, many more people than we would like to do. So the deployment of Proactive AI in our space has meant we could do that for their heavy lifting.
And again, I go back to the segment of customers, the small segment of customers where GenAI isn't going to work, AI isn't going to work. You need to pick up the phone, you need to go and manage their expectation. We are now getting the capacity to do that whilst offering tailored solutions and interactions with the rest of our customer base, if that makes sense.
Absolutely. Giles, Chris said that you can't employ loads more humans to look after the humans. You employ to those that need that human touch because there's this interesting corporate paradox. Implementing AI across customer touchpoints can drastically lower overhead, but if it alters customer perception of human involvement in their customer care, it could affect your brand. So how do you balance the need to automate your operations to reduce costs, obviously, with the strategic necessity of maintaining a premium human touch?
So this is a crux of where NICE and my company sits and essentially the automation of customer engagement while making it also human seems like a paradox, but actually it's got relatively straightforward answers. The first is that you automate all the low value touchpoints, all those low value touchpoints, the reordering, the resubmitting, all that stuff you have to call in about, you automate all those touchpoints. What that gives you then is vastly more time for the human agents, as Chris said, for the human agents you have to actually deal with the customers. And to my mind, one of the metrics we seek to eradicate is what they call average handling time, which is a call centre metric which basically measures how little time your call centre agents should spend and less is seen as more. We want to make it such that more time spent serving your customers is better.
That provides great customer experience. So the first thing is to automate the low value stuff. The second thing is to deal with issues proactively. Here's the thing. By the time a customer calls up, logs in, writes to you, whatever they do, emails you, the chance for an optimum customer journey has already been missed. It's gone because the customers had to do the work. You have to engage proactively with people in order to serve them best.
And Tom, could I just add to that? Okay. So I think the core metric in our business now, again, isn't necessarily pound shit in Pence. It's cancellation rates, it's escalation rates, it's contact rates. For me, they are the real key core KPIs that we're using them to measure the customer experience, because that's going to set us apart from the competition.
Well, Zach, looking at the market more broadly, from your perspective, what will differentiate the companies that successfully leverage AI for customer experience in the best way over the next five years from those that get it wrong and potentially alienate their customer base?
The companies who get it right are going to solve the answer to the question, if you could automate everything in your life, where would you stop? The automation boundary question is a fun one. Companies that get it wrong are going to automate too much or in some cases too little. They're going to have costs that don't allow them to compete or they're going to strip the relationship with their actual customer out of the experience and it will come at a terrible costable trust and brand.
Here's something I'd like for all of you to comment on. In 2021, a popular fast food restaurant launched a high profile pilot that looked like the absolute future of automated operations, deploying AI-powered voice ordering systems across more than a hundred drive-through locations. The strategic goal was efficiency, accelerate service speed/overhead costs and optimise peak hour traffic. Technically, the machine learning models performed remarkably well, capturing an impressive 85% accuracy rate over three years of testing. In a software lab, an 85% score is a definitive success, but in the unpredictable environment of a real world business, the remaining 15% margin of error created real world challenges that no lab test had surfaced. Without human intuition to act as a buffer, the AI lacked the context to handle chaotic real world variables. Viral videos flooded social media exposing a comedy of errors. The system routinely cross-contaminated audio streams from parallel lanes, added hundreds of unwanted chicken nuggets to single transactions and suggested bizarre culinary combinations like adding bacon to ice cream cones.
Rather than unlocking capacity, their system frustrated customers, created massive bottlenecks for kitchen staff and fell short of the projected efficiency gains. The experiment also introduced severe corporate risk culminating in a federal class action lawsuit alleging that the technology extracted and tracked unique customer voice prints without proper biometric data consent. By July 2024, corporate leadership officially pulled the plug on the programme. Now, what lessons can be learned here? What does this say about the importance of customer experience and perception? And even though we will never be able to get a hundred percent accuracy all the time, how should organisations handle some of those UX issues and bring customers along in the journey rather than alienating them? Let's start with Zach.
Some of this, by the way, is just experiment, trial and error. I think it's actually kind of fun that this happened in public. It happens in private all the time and then companies often figure it out and don't release it. So my hot take is we're going to see a lot more of this and some of it the company will resolve quite reasonably and some of it the company will fail to resolve and it will be a black eye that will live for some period of time. And to the extent to which it actually teaches the public what the technology is capable or not capable of is important, acknowledging that the technology is improving at a rate that most people don't appreciate. I think again, the bigger point is the degree to which humans have an exceptional tolerance for human failure and we have none for machine failure.
I write about this a lot. I talk to companies about this a lot. I call this one of the gaps in the societal threshold. What do we want machines to do is way more important than what machines... Can actually do. And it's the reason that 1.3 million people die on the global roadways every year, but if a Waymo swerves into the wrong lane, there's an international outcry to shut the whole project down. So this particular fast food company probably messes up way more orders with humans than it ever did with the machine, but we fixate and fascinate on this particular issue. I think a lot of companies are going to actually have to tow this line and some of them are going to get it very wrong and some are going to get it very right to the collective benefit of learning.
Charles. So there's a couple of things. Firstly, so this was launched in 24, so it was built on 23 tech.
Launched in 2021.
2021.
Yes.
2021. The technology between 2021 and 2026 is fundamentally different. Even in 23, we were talking about the foundation models had maybe 40% competence compared to PhD level. Hallucinations were rife. I can't go back as far as 2021. I mean, I hate to think what they were trying to do. So the technology now is radically different. We can't be compared to 21 in 24 to 26. So my advice would be try it again, but start smaller, start safe, iterate, get it perfect and then roll out. And crucially, make sure you're compliant with all data requirements. Don't mess up on that.
Chris.
Yeah. So look, I think a lot of organisations are shooting for the accuracy rates, the automation levels, whereas actually they're forgetting about what matters most to the customer. So start with the customer experience. Is it clear? Is it easy? Does it help the customer to get to the end outcome better, faster, smoother, more, with less effort for them? I think Charles mentioned a point around testing. I know the work we've done with NICE, we've had a lot of testing, making sure the guardrails for want of a bit of a phrase at the moment are in place almost trying to test and break the AI capability to its limits before it's launched to thousands and millions of customers. And we've done just that. So I think it's about testing, using real customers to test that with if possible, but making sure you're clear on what the customer experience outcomes you are trying to achieve are.
By the way, none of us have pointed out the fact that customers may simply not want a machine to take their order. I mean, the reason that people go to Sonic is people deliver their food. And in fact, Starbucks acknowledged that it had overindex on the experience for the mobile order and underindexed on the store to the massive detriment of the brand. I just think we should spend a lot more time actually asking, do people want this thing to be automated? You can automate things to the hilts. It's not clear that it actually serves the customer experience or the business.
Or given the customer the choice in terms of what do they want to go down this route or would they like to speak to someone to order their fast food, for example?
Yeah.
Well, before I let you go, a quick fire round for all three of you. What is the single biggest roadblock, cultural, technological, or financial that could cause an organization's AI strategy to falter over the next five years? And on the transverse, what's the one biggest factor that will determine if an AI initiative thrives, Zach?
The lack of a north star in both cases. I think great leadership right now requires, as it probably always has, courage and compassion, but we have a dearth of courage that stems all over and you can see it everywhere. I think great leaders will emerge as describing a castle on a hill, a place that their company can go achieve, which probably looks like expanded access to their goods and services for more people will work towards a north star. And bad AI deployments will work towards EBITDA, bottom line, accuracy measures or other miscellaneous metrics.
Charles, a quick response from you.
Sure. So biggest roadblock I would say is departmental point IT solutions. They do not build on the platform principle. They don't build on the data of the organisation. You need to have IT, the business, transformation, the executive, everyone, as Zach said, pointing towards, directing towards an aim they're trying to achieve and much stronger together and leveraging the platform in the company.
Chris.
Our biggest roadblock I think is a lack of discipline around the value that you are trying to create. If you can't clearly link the AI user cases with measurable business outcome, pounds, shillings, pence, customer experience, fewer cancellations, escalations, then I think it's going to stall regardless of the capability that you have.
Okay. In five years time, will the average consumer feel that customer service has become completely frictionless due to AI or will they miss human contact? Zach.
In five years time, businesses that have solved this will have figured out the degree to which their customers expect frictionless. And for many businesses, especially those that are experiential, I think they will actually have invested more in human touchpoint. Charles.
In five years time, customers aren't going to miss what's disappeared because what we perceive as connection now is often just friction and that will give them much more time, everyone much more time for the human engagement they actually want to have. Because again, you automate the straightforward things, you keep the precious touches.
Chris.
I think it will become significantly more seamless provided organisations use AI to help rather than deflect. I don't think a human is going to miss per se human contact if the experience is seamless and frictionless. And more importantly, back to the point, Zach, I think you raised, they have access to a human if they want to.
Okay. Final question. What is your one hope for how the balance between human talent and AI automation will reshape organisational culture by 2031? Zach?
I hope that we do not lose etiquette and tradition as we break the traditional career ladders. I hope that we raise the next generation to appreciate the shoulders of the giants they're standing on and also to learn about the ways in which business has been done in order to actually pass down meaningful stories. But I really hope that we give the next generation an opportunity to do way more than the last one was able to earlier in their career. I think young people are actually given far too little credit and opportunity and I think this technology supercharges the next generation to improve a lot of things that we did not. Charles.
So by combining the powers of AI and humans, my hope is that organisations, governments, institutions, rather than focusing on how they operate, how they manage, actually ending up putting the customer, the patient, the member, putting them at the heart of their business and that AI enables them to do that rather than focus inwardly.
Chris.
So AI should handle complexity and scale in the background, which allows our people to focus on judgement , empathy, and solving those real complex problems that you just need that human touch to do.
While it's unfortunately time to bring this episode to a close, many thanks to our fantastic guests. Giles, Bryan, thank you.
Thank you very much, Tom.
Chris Herbert, a real pleasure.
Thank you, Tom.
And Zach Cass, thank you for sharing your insights.
Zack:
Thanks, Tom.
Look, as we close this chapter of the next five, my takeaway from the drive-through of AI hype and value is one of wonderment. I wonder if measuring the returns of AI is less about the challenges of the technology and more about the human strengths and frailties of a living breathing organisation. It's easy to look at the massive investments and see only hype a speculative bubble waiting to burst, but as Zach, Giles and Chris have shown us today, real business value is quietly being unlocked by those who look past the interface. They bathe in the trough and look ahead to the sunny slopes of enlightenment with VIM and vigour. The true return on investment isn't found in a software licence or a standalone chatbot. It is realised at the intersection of automated efficiency, a fundamentally repurposed human workforce and an elevated, deeply empathetic customer experience.
The ultimate winners of the next five years won't be the companies that deploy the most AI. It will be the ones that use AI to make their human capital more impactful and their customer relationships more authentic. It is absolutely a time to experiment with people and purpose, but perhaps draw the line at putting bacon on your ice cream. I'm Tom Parker and this has been The Next Five Podcast. Thanks for listening.