Following a briefing in Belfast, Sync NI caught up with Ben Stanton, a senior manager in Deloitte’s Technology, Media and Telecoms Insights team, to discuss some of the hot topics in the world of technology and get his predictions on the future of generative AI, smartphones and streaming.
Read our Q&A below in full to find out more about Ben's views on AI and how we can monitor the improved productivity it brings, Agentic AI, screen fatigue and the attraction of analogue:
Q: Ben, one of the main topics in your presentation was Gen AI and whether people are using it. You said you haven’t seen it being as widely adopted at an organisational level as might be expected?
It's usually younger members of a workforce who are the most prolific users of generative AI.
Chat GPT came out in late 2022 so you have two and a half years’ worth of graduates coming into companies who have experience creating content, maybe disguising that they used AI to create content, and they bring those skills into the workforce.
They are able to use it more because if you are at a junior grade, the chances are the role is very task oriented. But it is a sliding scale - the more senior you are, the less you use Gen AI tools, because if you're a partner or director, your role will mainly be face to face with clients or thinking about big picture, a more strategic role. Gen AI can help with tasks assigned by that senior person but most are still not using it widely themselves.
Q: You spoke about the likely evolution of Agentic AI over the next 12 months and whether organisations can move from using AI for basic tasks to more in-depth functions. What level of change are you expecting, if any?
Building agentic AI systems is really hard, but that doesn't mean that adoption will be low. Actually, what I think is already happening is a lot of the big software platforms, for example, HR software, are building AI agents that interface with that software, that can complete tasks on that software that are given really narrowly defined parameters in which they can operate. So it might be that their only ability is to raise an HR ticket to log someone's holiday request or a paternity leave request.
But instead of me having to go to a portal and navigate 10 buttons and a form to fill out, it could be a quick message to an agent that exists in Microsoft Teams that will go away and do that for me.
I think, for a lot of major organizations that already buy big software platforms, it'll be an upsell on a tool that you already use, rather than creating some kind of digital assistant workforce from scratch. Most companies, even with Gen AI, are buyers, not builders, and with agentic AI, it's even harder to build, so it'll be a case of consuming a software product that hits the market, and there are already loads of examples that can do increasingly clever stuff.
Q: Many organisations will see that adoption of AI makes sense in many ways, and it should improve productivity. But there is a challenge around measuring that, isn’t there?
With generative AI, most of the proof of concepts that get created do not get deployed at scale because it will cross the desk of the CFO, and the CFO will say, ‘Why should I sign off a big budget for this big, transformational AI tool that you wanted to build or buy when I can't tell what impact it's going to make to our workforce’? It is really hard to do. There are economic advisory teams that are desperately trying to solve this problem of quantifying output per worker, and then how that will change. But roles are so nuanced, particularly for knowledge workers, where it's already difficult to quantify someone's output, let alone before and after tools have been implemented. So that makes the process of getting sign off for this stuff hard, and that's what is bottlenecking a lot of these projects from turning into scale AI deployment.
The other thing I'd say is, I see lots of companies that are becoming fixated with trying to measure their AI success. I think it is such a foolish question, because I don't think any company should try and measure their AI success. My advice is always, what are your business problems, and does AI help those problems get solved faster or not?
For example, I was speaking to a big consumer loans company the other day, and they were talking about all the stuff that we're trying to do with AI, but really, fundamentally, their business problem is, how can we get credit to people who deserve it, but normally get rejected in a credit check? How can we find ways of scrutinising their banking data to work out actually these people are credit worthy, and we can give loans to them. And they'd forgotten about that, and they were just thinking, what's our progress on our AI roadmap? Forget that. Just think about, what's your business problem, and does AI solve it? That's an archetypal mistake of 2025.
Q: On the consumer side, when it comes to AI a lot has been written about trust issues. You made the point that having great technology by itself isn’t going to change how consumers view a particular service. Can you expand on that?
One of the areas that generative AI could transform is the contact centre – the point of interface with a customer. But what people often forget is the contact centre has been relentlessly automated for the past 15 years, and that a lot of consumers have already had experiences of chat bots, or automated voice bots.
But I think a lot of customers look at a chat bot as a cost saving exercise by the company they're trying to talk to, so very often the most common question a customer still has is what's the quickest route to get through to human operator. So even if you deploy a Gen AI version of the same chat bot that can do far more for customers, the customer may not change their behavior because they've got the scars of older versions.
I spoke to one of our clients, a big automotive manufacturer, and they are really interested in how they can bring AI into the cockpit of cars to give people the ability to converse with their car, change settings on their car, just through voice. The challenge is, you've had voice assistants in the cockpit of luxury cars for the last 10 years, and customers have had terrible experiences over that time. So why would you bother changing your behavior now? Particularly because the average person on the street has never used a Gen AI tool before. It's only around one in three that have, so my call to action is always, if you're going to deploy a really frontier technology like this, you need to make sure you're not ostracising some demographics of your workforce or customer because they don't always know what these tools are or how they're different.
So there's a big fluency challenge that organizations face internally but also with whatever they want to deploy at the intersection with customers.
Q: You also said that for consumers, with products like cars and phones, there’s actually a trend towards more basic functionality?
Yes, we are seeing manufacturers putting buttons back into products. Smart phones have evolved so much that all phone manufacturers have left to add is buttons and make very incremental changes. They've gone so far with the device or product that this is all they're left with.
I do think that we're probably going to get to a point of screen fatigue. When you look at the amount of time that different generations spend on screens, it is still increasing the younger you get. But I think we're getting to biological limits. There's only so many hours per day that you can use to consume screens and actually a kind of a tactile experience, like sticking a button on a phone or even sticking a set of buttons back into a car cockpit, sometimes is a better user experience than having a screen there, because you can use it without looking or it gives you tactile feedback.
We have lurched headlong into this kind of digital world where screens are everywhere and everything, and now there's a slight kind of pullback counter culture towards more analog experiences. I wouldn't say it's a proper reversal. It's just a kind of, it's a trend and then counter trend. The counter culture and the desire to have analog experiences is strengthening this year, for sure.
Q: You used the example of young people using headphones with a plug in jack rather than wireless headphones – is that driven by this trend?
Partly, yes. You have to remember for young people it can be a cost thing. But look at the number of professional footballers when they walk from the team bus to the changing room that have wired headphones, because it gives you a cool kind of retro esthetic. And these people can afford the best Bluetooth earbuds on the market.
Q: You made a prediction about entertainment, specifically streaming content, that the main way these services are now achieving growth is through bundling with other services – for example when you buy a new phone. Is that market plateauing because we’re overwhelmed with the amount of options out there?
I think we're at a point of screen fatigue, but also subscription fatigue across customers. If you look at the average number of streaming video services that the UK household has, it's pretty much plateaued around two and a half services per house, and it's not increasing for a number of reasons. Some people feel like they've just got enough streaming platforms. There are cohorts of people that do something called subscription cycling, where you will maybe take one video platform for two months and binge everything on there, then cancel it and switch to a different one for another two months, binge everything on there. So this is a challenging cohort of people for a streaming platform. If you spend $200 million on a big budget, eight part TV show, you need someone to really be subscribing for a number of months to recoup that.
They are asking can I find new route to customers that wouldn't buy web service in the first place, but also, how can I find ways of locking customers in for slightly longer contracts. If they buy an 18-month internet subscription and there is a video platform bolted on it's not impossible, but it's more of a hardship to go about decoupling that and canceling. If I buy a streaming service on its own, usually you can cancel in the same month, unless you bought an annual plan. So bundling with other services makes sense.
Read the latest edition, Spring 2025, of Sync NI here completely free of charge.
Subscribe to the Sync NI newsletter for all the latest technology news, jobs and upcoming events in Northern Ireland.
Visit Sync NI online for the latest technology news in Northern Ireland.
