As part of our Interconnection Between People, Process and Technology book produced in partnership with Tanium, we spoke to Jim Green, Technical Director, AI & Data Team Lead, AtkinsRéalis.

With a background running big data programmes in Defence Intelligence, Jim joined engineering consultancy company AtkinsRéalis as a Technical Director. He mainly works with public sector and critical national infrastructure clients where AI is the hot topic of conversation. Jim believes AI's true value currently lies in automating dull, repetitive tasks that humans find tedious. And in an industry where security is so key, it is helping to bolster cybersecurity defence.

“When we look at what AI is optimising, it really is back-end efficiencies. It tackles those mundane tasks that humans often find repetitive and dull. Automating these processes is where AI shines. While discussions about generative or general AI and their future potential are exciting, the real impact for me is in eliminating tedious tasks.”

Jim feels passionately about how AI automation can push employees up the value chain.

“What I'm excited to achieve is helping get our government organisations to a place where they've got that high threshold already and we can start sprinkling the AI on top so that everyone is performing where they need to perform.”

Jim believes that the AI hype cycle peaked last year and has since slowed down, with the focus shifting from flashy features to practical uses and real-world results. From his experience, he believes AI has evolved beyond pilot projects and experimentation to implementation, and this brings with it a whole set of new challenges with integration, scalability and trust. Jim believes that rather than seeing AI as a technology integration programme, it should be approached as a change management initiative with AI being just one piece of the jigsaw. But scaling this to enterprise level is, he admits, very challenging.

“With scalability you need to have the hardware enabled for that and a lot of that is about migrating to the cloud. Significant swathes of the public sector have now moved to cloud, but there's still a challenge with certain kinds of business units doing that migration and building confidence to migrate to cloud from a security perspective.”

Trust issues are particularly key in areas of Defence and Critical National Security. Trust in the algorithm is vital but there also needs to be transparency about what ‘good’ and ‘bad’ looks like. That requires upskilling your workforce not only in AI skills but also fundamental digital literacy which is still lacking in many organisations.

Jim believes that the biggest pain points in the organisation can be traced back to legacy processes rather than technology.

“The technology is generally pretty simple. I say that you can buy a new server rack or you can migrate to cloud, and there are some great cloud providers available and they’re secured to the appropriate levels. That's all achievable. The people and the processes are the bit where it gets really interesting to my mind, particularly when you start to scale.”

Jim also highlights the dangers of outdated workflows and an underlying reticence to change that’s holding back the inculcation of AI at scale. Risk aversion is also high among the defence and critical national infrastructure community.

“All change is scary in some shape or form because it's moving into an unknown system. It's about education and that can be done in a really light, informal manner.”

Jim recommends spending time with people and breaking down AI into bite-sized, understandable chunks as this helps them understand where they fit in the AI value chain.

“It's about prediction. It's about classification and once you understand the kind of building blocks of what AI can actually achieve, it's a lot less scary. Humans are fantastic at multitasking and juggling all sorts of complex problems while AI at the moment is not really good at that.”

When it comes to decisions on purchasing new technology, Jim cautions, “Is ‘new’ always better?”, when the pain of migrating from one software package to another at enterprise level can be so significant. Indeed he advises that leaders ask themselves, is the benefit worth it or are you just adding complexity?”.

“Often, organisations haven't integrated the tool properly, so they try and move software and think that will solve the problem, but that’s not the problem. The problems come back time and time again to the people and the processes, and the software is just the relatively static, simple piece.”