AI may be becoming part of the furniture, but challenges remain. From data management hurdles to ongoing concerns about employee trust, leaders are still grappling with how to manage change. Here, five experts share how organisations can move past the sticking points and implement AI more effectively.
Panel:
Johan Huss, vice president and head of digital machining at Sandvik Coromant
Greg Smith, executive director and general manager, EMEA SSG at Lenovo
Matthew Unangst, senior director at AMD
Clare Walsh, director of education at The Institute of Analytics
Meri Williams, chief technology officer at Pleo
What challenges are we still seeing around managing data?
There are a handful of challenges that the industry and ecosystem are working through. I think data is certainly one of the biggest ones. When an organisation is thinking about a problem to address with AI, it’s vital to understand which data sets are needed and how those inputs will shape the solution.
It’s then fundamental to understand how to train the models and applications. We have to teach people and teams how to use these tools because they are new. It’s a different way of thinking, a different way of structuring problems and a different way of using data.
I agree completely. I’d also add that dealing with very private and personal data in my industry helped with this. It had a chilling effect on us in the beginning because, early on, it wasn’t possible to use ChatGPT and keep your own data out of the models. That’s been solved now, but that privacy and security angle continues to be challenging and is something that we need to keep focusing on as we roll this out.
In my industry, there’s a lot of legacy. How can we leverage that? Where do we store it? How do we manage access? Before we can start realising value with it, how do we even dig it up and start combining it? So for us, data management and structuring are a big challenge. The progress of AI accelerates that need. We have to tackle that before we can move on to more advanced use cases. There’s no shortage of data, it’s how it’s structured that is a big challenge right now.
It’s amazing, isn’t it that? If you go back 30 years, we were having exactly the same conversation about data, just in a different context. How do you ingest and manage data? How do you secure, store and access it? A lot of the AI challenges we see now might have been solved if the industry had taken data management more seriously.
There are four things you have to get right for a successful AI implementation: security, people, technology and process
How can leaders manage the people side of AI adoption?
Our CIO here at Lenovo says there are four things you have to get right for a successful AI implementation: security, people, technology and process. And he says the easiest one to fail on is people. If you don’t get people to come along on your journey and buy into it, your project will fail – no matter how good the business case is.
It tends to be the CIO leading this effort, and they need to really become part HR, part facilitator, part conductor and part technologist. It is a phenomenal role that the CIO now has to adopt.
Organisations tend to forget that this is about managing change. When you deploy GenAI into the workforce to boost productivity, satisfaction and all those wonderful things, it only works if change management is done properly. If it isn’t handled or positioned well, the whole initiative is likely to fail.
We’ve been having this conversation for so many years. With every new iteration of the technology, we’re seeing some very forward-thinking companies starting to recognise the multiple opportunities to upskill staff. For example, if you have some high-performing individuals, but there’s no promotion available to them this year, you’ve got to do something to keep them motivated and performing at their peak.
A lot of companies that we’re speaking to that are doing very well now are beginning to offer training in digital and data skills to keep those people engaged while they wait for an opportunity. When development is integrated into other business aims, we’re starting to see a more collaborative approach – not just ‘I’m going to tell you what I’m doing’, but ‘let’s work together and see how we can all achieve mutually beneficial goals’.
How can leaders ensure employees don’t feel threatened by AI?
In a way, it’s a bit easier in a heavily regulated industry like mine, as we need to keep a human in the loop. The message is clearer when we’re not fully automating or handing things over to AI because we’re simply not allowed to. In that scenario, it’s very obviously about augmenting humans, rather than replacing them. But I completely understand that’s not the case for everybody, and it’s causing anxiety for a lot of people.
It’s important to provide training. We’re considering doing a whole programme of work around GenAI because we don’t want to be in a situation where engineers who want to upskill have to do that on their own time. Not only is that an unreasonable expectation, but it also exacerbates inequality. If some people have more responsibilities outside of work, you don’t want them to be falling behind because you haven’t provided time to learn during business hours.
AI may be becoming part of the furniture, but challenges remain. From data management hurdles to ongoing concerns about employee trust, leaders are still grappling with how to manage change. Here, five experts share how organisations can move past the sticking points and implement AI more effectively.
Panel:
Johan Huss, vice president and head of digital machining at Sandvik Coromant
Greg Smith, executive director and general manager, EMEA SSG at Lenovo
Matthew Unangst, senior director at AMD
Clare Walsh, director of education at The Institute of Analytics
Meri Williams, chief technology officer at Pleo
