Schneider Electric CIO Elizabeth Hackenson on AI, IT's energy consumption, and CFO demands...
"I don't think there's enough energy in the world for what we're going to need for AI."
Just how many technological revolutions can one CIO take?
Elizabeth Hackenson had a front row seat for the internet and mobile explosions while working at UUNet, MCI and Lucent in the late 1990s and early 2000s. She then moved into the energy world, spending eight years as CIO at AES Corp, before joining Schneider Electric Corp as CIO in 2017.
It’s fair to say that “power” (however dirty) was a given and “sustainability” was at best a nice to have in the 1990s. Now it’s a critical issue for CIOs, as they contemplate how to take advantage of the latest technology revolution – GenAI – while meeting its voracious power demands AND more stringent government regulations (and investor pressure) around sustainability.
We spoke to Hackenson as Schneider Electic launched an update to its EcoStruxure data centre infrastructure management (DCIM) platform to help customers meet upcoming regulatory requirements, such as the European Energy Efficiency Directive (EED). As of April the platform will deliver updated reporting, which uses machine learning to track both historic and realtime energy performance, and one-click reporting to meet EED and other reporting requirements.
Schneider Electric CIO: Trialling Copilots...
Hackenson said the company has a number of transformation programs in play across the company: "Everything from how we interact with our customers, to our supply chain, finance, HR..."
Unsurprisingly, a lot of this involves the use of AI. The company has its own internal "conversational assistant" on the Microsoft Azure OpenAI Service, “which employees are getting comfortable with.” She also namechecked partners that were coming out with copilot technologies including SAP, AWS and Microsoft.
With the latter, “We’ve got a sales copilot going on with them right now. We're going to pilot Office 365 Copilot here any day now.
But with AI technology coming thick and fast, Hackenson said this year was about “Experimenting as much as we can to see where we get traction with employees. And then probably next year is when we scale.”
This all comes in the wake of a wide-ranging effort across the company over the last few years to pare down its hardware estate and make it more efficient.
“We had a situation where we had at least 140, what we would call high-value server rooms,” says Hackenson. “So these are on-premise, and they support our factories or distribution centres and large office locations. And we uncovered that we could do better.”
Rethinking energy consumption analytics
Unsurprisingly, whilst rethinking this infrastructure footprint, Hackenson’s organization tapped Schneider’s own EcoStruxure expertise to understand the power implications of its estate.
“We had no real intel or baseline on our energy consumption. We have hundreds of thousands of IT assets in the company that we manage. They're all over the place all over the world. And they're on different utility bills”.
This all made it “really hard to decipher” the true costs of its IT assets, Hackenson said. “So now for the first time, we have the ability to see in real time energy consumption, we have the ability to look backwards; start to compare trends.”
This experience was common to a lot of CIOs, Hackenson suggested. “They want to be part of sustainability efforts in their company, but it's not easy to get there.”
But working out energy consumption and trends was just the first step, she explained. “What we're trying to do is reduce our CO2 emissions. And that's by reducing the equipment that we have in our ecosystem.”
This effort saw it move thousands of servers to the cloud, while moving others to co-lo providers. In both cases, it targeted providers on their sustainability efforts.
But some infrastructure, particularly that associated with manufacturing plants, couldn’t simply be migrated or relocated. In those edge cases, it kicked off a ruthless and methodical modernization program to replace aging and therefore more inefficient server, storage and networking kit.
This meant working with its suppliers, the likes of Cisco, Palo Alto, et al, “We want them to have the most energy efficient products as well.”
This was matched with a change in procurement processes to keep kit evergreen – or at least greener – with a shift to an opex rather than capex model.
“There's always pressure to take an asset to its longest life, which can be even way more than 10 years. And we just think in general it's not a good idea from a sustainability perspective.”
However, that raises a sustainability issue around the equipment that has been retired. So Schneider takes pains to educate itself about the embedded carbon of its kit, as well as its power consumption, and the recycling standards and certifications of its suppliers to ensure its being processed in the most sustainable way.
And it takes a similar approach to its off prem infrastructure providers, says Hackenson. “We're also getting reports from our big cloud providers, [about] how they're driving energy efficiency as well.”
Those same providers are often customers of Schneider Electric. Hackenson said, “We really try and make this a partnership…everyone’s learning at the same time here.”
And there are external factors to consider as well, particularly in Europe with new regulations on the way.
As for other CIOs, she said, “You have to start somewhere." This could be just the energy consumption as stated by a supplier "and then use something like EcoStruxure IT Expert to then start to manage it in real time.”
This effort to collaborate on sustainability has been given added impetus by the GenAI explosion. “I believe that what we're going to see with AI is so exponential, that we're all going to have to figure this out together, because I don't think there's enough energy in the world for what we're going to need for AI.
This also highlights the changing role of the CIO, she says, compared to her earlier career working in telecoms for organizations like Concert and UUnet as the internet was taking off.
Sustainability wasn’t thought about too much back then, she said, if only because the big challenge was that a lot of the technology didn’t always do what it was supposed to do.
“It felt like we spent a lot of time in the data centres, software applications, other technology like integration tools. You were so focused on, does it really work?”
By contrast, she says, technology has advanced so much that the CIO now plays a very different role with the business and stakeholders, “which is really trying to help figure out, what is the best technology or digital solution for a problem that our stakeholders are trying to solve?”
So while the technology itself might work, she says, the question is whether it meets privacy standards, or security standards, or, of course, sustainability requirements.
There’s another difference between the onset of AI and the shift from private connections to the internet, she says. The latter was comparatively gradual. This is definitely not the case with AI, she suggests.
“If we think about generative AI, so many companies are coming up with copilots, and they're all adding a pretty significant price to join in, right? And all CFOs are now asking for what's the business case? What's the ROI?”
The problem is, she continues, is that, despite some obvious use cases, most organizations are still working out what to do with the technology even as they rush to adopt it. “This is the year to experiment with these copilots, and then try to figure out which ones are best for your company to scale, let’s say, next year.”
That might not suit every CIO or CFO. But that’s the thing about revolutions. They can get messy. Or as Hackenson says, “It’s a super exciting time.” Again.
