SPECIAL EDITION: Scaling, ROI, and what’s next for workplace AI

Businesses in the UK and throughout the world are ramping up their investments in AI as the benefits of the technology become increasingly apparent. Unlocking the value of AI isn't without its challenges though and just like other waves of technology innovation there are obstacles to overcome along with the potential for growth. How are businesses approaching their AI investments? What ROI are they starting to see and how can organizations move from experimentation to AI at scale. I'm Jane McCallion, and you're listening to a special edition of the IT Pro podcast brought to you in association with SAP. Joining me today to discuss all this and more is James Bates, Head of Industry and Customer Advisory at SAP UK and Ireland. James, welcome to the show. Fantastic to be here, Jane. So I'm sure many of our listeners will be familiar with SAP, but could you give me a bit of an overview of what SAP is and what you're up to as a company today? Absolutely. So SAP, Jane, is one of the world's largest enterprise software and AI companies. Our mission is to help the world run better and improve people's lives. And we do that by underpinning our customers' mission critical processes. So these are things like HR, finance, supply chain and customer experience. So I'm trying to give maybe some examples. If I reflect back on our Sapphire conference earlier in the year, we were lucky to share the stage with the likes of Levi Strauss, ExxonMobil, H&M. It's really, really exciting time to be working for SAP and in the industry. In terms of what we're up to, well, we're really at this sort of intersection. It's really the exciting intersection of software, AI and data sort of coming together to really drive transformation. So, yeah. Very impressive list of guests from Sapphire. Given SAP's focus on business AI, what are the biggest areas of growth and investment that you see among UK businesses? I imagine that agentic AI has to be quite high up the list. Yeah, certainly. If you actually look at some of the research that we've been doing with the likes of Oxford Economic Group, 70% of organizations are already piloting agent-based use cases. And that makes sense. So agentic AI is really exciting if you think about sort of where we've been with AI and where we're going. um ai so largely to this point has been fantastic in terms of productivity use cases fantastic at sort of um answering questions i'm sure um most of the audience are using it in their day-to-day lives like like like we are but if you think about sort of the prospect of agentic ai and its ability to transform end-to-end processes where actually you'll have uh agents really driving actions albeit still orchestrated i would i would stress by people and people very much at the sort of the center of that that that that decision making kind of process um you're right there is definitely a lot of excitement about it and and in terms of growth and if i reflect on some of the conversations that i'm having with customers A lot of it is around sort of that potential of agentic AI. However, I wouldn't sort of ignore some of the value of what we're able to deliver today. And if you look at sort of the overall investment picture, again, calling out some of that Oxford Economics Group research, the average investment is increasing. So going from around 16.2 million in 2025 to 20.6 million in 2026. And actually, if we think about where customers are investing and working and deploying capabilities today, it's also in building some of the foundations around AI. So things like data, processes, skills, governance frameworks. And I often think, actually, that the organizations and customers that will sort of win in the long term are those that are focused on building those initial foundations rather than perhaps experimenting sort of straight with agents. And we are also increasingly hearing about the challenges that businesses face when it comes to implementing AI. One thing that is a constant trend or perhaps a complaint, really, is the lack of ROI. Organizations are pouring money in, but they're not really getting much bang for their buck. Is this a technical issue, a planning issue, or perhaps a little bit of both? A little bit of both. um it's it's interesting jane i it feels like we're a little bit of a sort of um a pivot point an intersection if you will where if you look back 12 months ago there was a lot of data um pointing sort of to this roi challenge but it does feel like we've moved past that a lot now and we're sort of now into sort of like the real um execution phase of ai so actually we are starting to see that ROI data improve. Again, sort of introducing some of the data into the conversation, because I always think that's really interesting to do. The ROI that we were seeing last year was around 17%. That's now 22% in 2026. And if I reflect on what we're doing and our organization and the conversations, again, that I'm having with customers, those productivity use cases, So help me write a job description or help me set a series of measurable goals for my team. Those sorts of things are actually there now and is really driving, I think, some of that value conversation or value ROI for customers. I wonder if perhaps it's one of those cases where people were expecting ROI slightly too early on. that you know this is a new investment you're gonna have to wait a little bit before you start getting your investment back effectively yes i yeah i think i'd agree with that um if i think back to some of that foundational sort of uh build out when it comes to AI that the ROI for some of those things i think is always going to be challenging but that investment is building for the future we've already touched on sort of the potential of agentic AI and if if customers aren't investing in uh you know processes data governance context then actually i think you know that will slow them down in in the future whereas investing in that now i think then really sort of accelerates that ROI going forward you just mentioned data governance you know this is another topic that people are starting to talk about quite a lot and the importance of ensuring that greater use of data doesn't mean putting that data at risk how should organizations approach this so the one the one thing the great great question uh i'm really actually really glad you've raised it i mean um some of the events that have been well publicized recently have certainly driven a lot of focus on this and rightly so um you know right from the start of the AI conversation SAP has put governance ethics compliance auditability um right at the heart of everything that we're doing and that's really important particularly if you think about um the sorts of processes and data that SAP's customers are dealing with um so we often sort of talk about some of this concept of um you know nearly right isn't good enough in our world um it actually has to be a hundred percent right which is interesting if you start thinking about the nature of probabilistic systems versus deterministic and of course large language models are probabilistic and we're having to sort of uh try and understand how we solve some of those challenges uh when it comes to sort of the the nature of um data within uh within an SAP kind of landscape the other thing i'd sort of add is for me yes data governance is important but actually it's a far bigger topic than that um i think you know the whole end-to-end process governance understanding auditability permissions identity these are the sorts of things that we're working with customers to try and sort of solve yes my sort of one recommendation around data governance and governance more generally is for customers to put that as a first principle right from the start so if you're designing and building out AI systems um you know that should be part of a consideration right from the start rather than an afterthought um i think it has to be sort of really foundational let's talk a little bit about skills obviously generative and agentic AI are still a relatively new technology so the number of experts or even just people with a decent amount of experience on the job market will be limited how are businesses in the UK addressing this so i think i think this is again with sort of the governance topic also another really important one to raise i think that the tendency could be to kind of think of well do we need to go out and hire x number of AI experts and and actually that's that's not how we're thinking about it you know people that understand the models and how they're architected will be important but actually i think this is uh the skills question is a question for the entire organization and it's almost sort of the organizations i see sort of really getting ahead and if i reflect on sort of what we're we're working on we're looking at this in terms of all roles so what skills are needed um to support AI enablement regardless of sort of what your role is so if i reflect on on my own organization everybody has um an AI goal um usage of AI so for me it's not about well we need to create different roles it's how AI can empower and drive productivity for all roles so one of the things that that we've done at SAP is we built a community of 9,000 AI ambassadors um that sort of really driving this sort of We like to call it work the AI way out to the entire organization for how everybody can use AI to the most effect, regardless of what role you are. So I think this industry term of almost democratizing AI, I agree with that. I think this is something for all of us. Looking into the future now, what is SAP's vision for where all of this leads? So, we're really excited about what we're working on. We are envisaging a future where organizations and organizational processes will run autonomously. Now, to be really clear, we're still very much seeing the human in the loop here. So with the human being the individual that sort of drives the decision making, but if you think about a financial close process or you think about some supply chain processes or other processes that we support within our customers, if you look at a lot of those things, tasks involved in those processes can be quite long winded and can be quite administrative. So we're seeing a really exciting vision that we call the autonomous enterprise, where the tasks are driven by agents and assistants, so orchestrated by the assistants, tasks driven by the agents, but with the human absolutely sort of driving that value, if you will. So in that sense, perhaps back quickly to the skills conversation, actually skills around business acumen, understanding business processes become even more important in a way rather than sort of the specifics around understanding the AI element, if that makes sense. Yeah, yeah, absolutely. Well, unfortunately, that's all we have time for this episode. But James, thank you very much for joining me. thank you bobby and i will be back on friday with our regular look at the world of IT but until then goodbye