When Brad Clay became chief digital officer of GlobalFoundries in early 2021, he knew his role would be less about technology implementation and more about process change.

In 2018, the $8 billion global semiconductor manufacturer announced a pivot in its business strategy: The company would no longer develop and produce 7-nanometer and smaller chip technologies; instead GlobalFoundries would focus on producing specialized chips for high-growth markets such as automotive, 5G, and the internet of things.

“When we shifted from commodity contract wafer manufacturing to delivering more value to the device manufacturers, we faced a different set of business problems,” says Clay. “We had to align our business processes with the new strategy.”

Wholesale shift in process management

GlobalFoundries grew up from a collection of companies, each doing things differently, so processes had become siloed and fragmented, with no single person or organization accountable for an entire process. This, along with the change in business strategy, required a wholesale shift to global process management.

Clay, who is also CIO, worked with the IT team to spearhead development of a global business process owner model. “We wanted to understand and define how our processes could be interconnected and work end to end,” he says. “We had to break down the silos that naturally occur between finance, planning, and supply chain, for example. We did not start with an organizational construct; we started with a process construct.”

Once the global process model was defined, the technology team would create two common platforms, one for global processes and the other for data, but they could not put the cart before the horse.

“When you start a transformation, everyone wants a quick win,” says Clay. “We resisted that approach. We spent a year developing, defining, and visioning our new process model before we bought any software. That approach has paid huge dividends.”

Introducing a new role: Global process business owner

Clay and his senior executive peers identified eight global processes: idea to product, hire to retire, order to cash, demand to deliver, source to pay, market to contract, make to order, and record to report. They then identified people for a new role, global process business owner (GPO), which would be the linchpin to the new model.

In companies that have a relatively clear understanding of their global processes, that GPO might be a Six Sigma Black Belt with a continuous improvement lens, but this was not the case with GlobalFoundries. “Because the concept of global business processes was new to us, we needed VP-level leaders to step into the new role. We had to drive the transformation from the top down.”

Taking on a GPO role is not for the faint of heart. Clay needed the newly appointed leaders to understand the magnitude of the change they would drive. “We communicated that this was not a continuous improvement effort where the owner would make the process 5% better,” he says. “Our message was that transformation starts at 50%, and that our leaders had to have the vision and courage to sign up for that level of improvement. Anyone will sign up for 5%; it’s in the margins. But 50% is where people get nervous. That’s a very visible level of accountability.”

Once process owners were identified, they went through training so they could have a common understanding of processes, lexicon, and ways of interacting. “We even did 360-degree assessments on the leaders, because we needed the process owners to be a tight-knit group,” says Clay. “They would be accountable for driving common process through the company in a way that had never been done before.”

Under each GPO are process advisory groups that span the various departments involved in a single global process and have a stake in its improvement. Because a GPO cannot have detailed knowledge about every single piece of a process, these advisory groups are critical to making the global process owner model work.

“The advisory groups ensure that the GPO understands the user stories, and they make sure that everyone knows what is going on with the processes,” says Clay, who also reorganized IT so each GPO has a dedicated technology owner.

With the GPO model in place, Clay and his IT team could now address the challenge of implementing new software to automate the global processes. “We had primarily been using point solutions for specific requests held together by manual effort,” he says. “We had to cross ‘the gap of stranded investment’ and focus on platforms. We replaced pretty much everything — ERP, CRM, PLM, quality management — new software soup to nuts.”

GPO lessons learned

Now that Clay can see the faster decision-making and increased productivity that has resulted from the GPO model and platform architecture, he has some lessons to share.

The first: Transformation is more than software implementation. GlobalFoundries’ GPOs are aware that transformation has two elements: digital enablement and business change, which ensure that your operating model is aligned with the business strategy.

“That is why the GPO has to be a senior person,” Clay says. “The GPO aligns the processes to the corporate strategy and then makes sure that what IT is building into the platform aligns. I believe that digital transformations fail because its leaders miss that duality.”

The second lesson is that when you implement the software, minimizing customizations helps you avoid “fighting gravity.” Clay sometimes gets up in front of his colleagues and drops a rubber ball to make the point that when you choose to veer from a vanilla ERP, for example, you are trying to keep the ball in the air.

“Commercial software was built by people with expertise in business processes,” he says. “When you decide to customize the software, you are deciding about whether you are removing friction or fighting gravity. Our goal is to fight gravity only where we absolutely have to.”

Finally, Clay points to the importance of senior-level support and business engagement.  Early on in the transformation, hundreds of people across the entire global company came together to walk through each process, with each GPO standing up to address plans for their own area. “It was the global process owner explaining how processes are going to change,” he says. “It wasn’t an IT person explaining how SAP works.” That, and the fact that CEO Dr. Thomas Caulfield described the program as “business transformation enabled by IT,” were critical to the program’s success.  

“At GlobalFoundries, we manufacture semiconductor chips in four different facilities across three continents,” says Clay. “But the GPO model was a true transformation, which we had never done. And that’s the challenge of transformation. It’s always unique.”

Business Process Management, Digital Transformation

Over the next five years, the healthcare industry is expected to go through dramatic changes as service providers expand value-based care models and equipment manufacturers strive to keep pace in a digital-first world. One factor driving global transformation is the push to bring healthcare services, as well as responsibilities and control, closer to consumers.

By 2027, 70% of healthcare organizations will rely on digital-first strategies to empower patients to take a more active role in health responsibilities and experiences, notes IDC in its FutureScape Worldwide Healthcare Industry 2023 Predictions report. Spending on digital patient engagement and experience technologies that target such areas as mental health, telehealth, clinical trials, digital therapeutics, and ‘care anywhere’ solutions is expected to rise, according to the report.

Bedford, Mass.-based Novanta is a trusted technology partner to medical and advanced industrial OEMs, with deep proprietary expertise in photonics, vision, and precision motion technologies. Novanta engineers mission-critical components and subsystems that deliver extreme precision and performance, enabling customers to improve productivity, achieve breakthrough performance, and enhance people’s lives.

Leading the technology strategy at Novanta is Dr. Sarah Betadam, an experienced digital executive who, as presently the company’s CIO and CISO, is charged with delivering transformational solutions that scale operations and improve business agility, but she also helps drive business strategies and global objectives. Betadam is well-suited for the task, given the range of her experience prior to joining Novanta more than four years ago, and holds a doctorate in engineering and engineering management, as well as degrees in computer science from noted universities.

A self-confessed data analytics and research junkie, Betadam wrote a thesis presented to George Washington University a few years ago that outlines a contemporary model for IT program management that challenges many existing models, which she calls ‘overly subjective’ and less viable in today’s world.

“Everybody wants to have a business impact and business outcome,” she explains. “But business value is what is important for a particular business, which stems from the business capabilities rolled up from the business strategy. There’s revenue generation, the compliance aspect, operational efficiency, and legal. If you total all of this up it gives you a spectrum,” Betadam says, noting that this model was used at a previous organization and resulted in a more focused and successfully executed strategy.

More details on Betadam’s program management model, as well as her thoughts and insights on the challenges that lie ahead in terms of leading a transformational digital business, are presented in this CIO Executive Council/IDC Future Forward podcast interview. The following are edited excerpts from the podcast. Click on the links at the end of the excerpts to listen to Parts 1 & 2 of the full conversation.

Tim Scannell: IT organizations are going through a lot of structural changes, with reporting lines changing and the focus becoming more intense on business and customer facing objectives. How has your IT organization changed over the past year or two, and how has this made Novanta more flexible and adaptable?

Sarah Betadam: We’re expanding our competency levels — not just within IT, but across the business — to be successful. There’s also another aspect in terms of the agile transformation within our team and flexibility in terms of product roadmap ownership. This means really emphasizing the business partnership and making sure we’re aligning with the business strategy. We’re very big into Lean processes; aligning to those is how you get the momentum going. It depends on the resilience of the CIO to make sure there’s a match between the capabilities you’re trying to add as components to the Novanta organization and the current flow of programs and projects that each business unit is running.

Technology initiatives that have a direct impact on the bottom line are more of an emphasis today. Does this create any pressures on you and your IT organization to complete projects more quickly and with less risk involved?

Betadam: This is why agile is important for our entire organization. It’s educating the business that it’s okay that something is not going to be perfect the first time around. I also use Lean language, which is more familiar, letting them know there will be continuous development, process improvement, iterations, and that everything is not going to be perfect. It’s a journey and people are getting more comfortable with that approach than in the past two years.

A lot of companies today are investing more in data analytics and business intelligence. Some are even establishing digital strategy teams consisting of people from both IT and the business communities. The problem, however, is finding experienced IT and business talent. How is that impacting your data analytics strategies?

Betadam: We started our journey pre-pandemic with one or two people and now have a BI Council. We have brought in experts from the business to take ownership of the different data that we are trying to centralize, but finding people and talent is very difficult. One solution that has worked for us is to expand your search to places where you wouldn’t think that there would be talent. We’ve started doing analytics training, where everybody is invited to join, and more people from all departments are interested in learning more about data reporting. We are evangelizing this across the organization, so you don’t have to be IT person. You could be anywhere within Novanta, and we will train you to make use of the data that could help you and your team.

A lot of IT organizations are being asked to minimize risk and focus on projects that deliver a definable and maximum ROI. Do you see this as impacting your innovation initiatives in any way?

Betadam: I think there is definitely a slowdown in innovation activities, especially during times of economic uncertainty. Innovation comes with a certain tolerance level of accepting failure. At the same time, you are trying to keep the lights on and making sure customers are satisfied when there is a lot happening outside of your own control. Depending on the leader, however, there might be some incubation to keep things going. We’ve been looking at different efficiency solutions within the organization and at tools like AI. But this is done by a smaller group because we really needed to shift our focus to meet the customer amid all these different vectors that surround us. So, it slows down, but I don’t think as a leader you should shut it down because then you’d be behind.

Digital Transformation, IT Leadership

As organizations shape the contours of a secure edge-to-cloud strategy, it’s important to align with partners that prioritize both cybersecurity and risk management, with clear boundaries of shared responsibility.

The security-shared-responsibility model is essential when choosing as-a-service offerings, which make a third-party partner responsible for some element of the enterprise operational model. Outsourcing IT operations has become a smart business strategy. But outsourcing operational risk is untenable, given the criticality of data-first modernization to overall enterprise success.

“Intellectual property is key to a company’s success,” notes Simon Leech, operational security lead for HPE GreenLake Cloud Services. “Therefore, it’s up to CIOs to do due diligence about what sort of security controls are in place and to ensure data is well protected in an [as-a-service] operating model. The security-shared-responsibility model provides a clear definition of the roles and responsibilities for security.”

Having a well-articulated and seamlessly integrated security-shared-responsibility model is table stakes. Organizations are spending far more time grappling with the costs and consequences of highly complex cyberattacks, to the tune of a 72% spike in costs over the last five years, according to the Accenture/Ponemon Institute’s “Ninth Annual Cost of Cybercrime” study. Specifically, the study attributed an average $4 million loss to business disruption, with another $5.9 million associated with information losses. In total, the global cost of cybercrime is skyrocketing, expected to grow 15% annually to hit the $10.5 trillion mark by 2025, noted the “2020 Cybersecurity Ventures” report. 

HPE GreenLake: Security by Design

Against this backdrop of heightened cybercrime activity, organizations are more vulnerable as the proliferation of platforms, internet-of-things (IoT) devices, and cloud applications has created an expanded attack surface and widened security gaps. A new security-by-design approach infuses security practices and capabilities directly into new systems as they are built — versus addressing security requirements later as an afterthought. 

An organization’s approach to security must also scale at the speed of digital transformation. This means that security must be automated and integrated directly into continuous-integration/continuous-delivery (CI/CD) pipelines, ensuring that safeguards are applied consistently across workloads, no matter where data resides. This also makes it easier for developers to create secure code. As organizations grapple with additional complexity challenges, they need access to third-party security experts to close any internal security gaps.

The HPE GreenLake security-shared-responsibility model differs from that of the typical cloud provider, because the as-a-service platform delivers a public cloud experience everywhere, including in a company’s private data center and/or in a shared colocation facility. The company or colocation provider maintains responsibility for securing the connectivity and physical data center, and HPE’s responsibilities vary, depending on the chosen HPE GreenLake consumption model. For example:

In a bare metal model, HPE is responsible for securing the HPE GreenLake infrastructure and cloud experience, but the customer takes ownership of everything on top of that infrastructure, including the operating system (OS), hypervisor, container orchestration, applications, and more.

With containers and virtual machines, the responsibility shifts and HPE GreenLake handles security for the lower layers that includes the hypervisors, software-defined networking, and container orchestration. Here again, the customer is responsible for securing the guest OS, applications, and data.

For workloads, such as SAP Hana delivered as a service or electronic health records as a service, HPE GreenLake takes security responsibility for everything up through the application layer whereas the customer maintains ownership of data security.

“In all three scenarios, security of customer data is always the responsibility of the customer,” Leech says. “It’s ultimately their responsibility to decide what data they put in the cloud, what data they keep out of the cloud, and how they keep that data protected.”

Best Practices for Security Success

Drill down into the details. Leech cautions that the No. 1 rule for security success is understanding the boundaries of responsibility and not making any premature assumptions. Organizations should confer with their cloud service provider to clearly understand and delineate who has responsibility for what. Most cloud providers, including HPE, offer collateral that drills down into the details of their security-shared-responsibility model, and customers should take full advantage.

“The risk is really one of blissful ignorance,” he says. “The assumption can be made that security is there, but unless you actually go into the contract and look at the details, you might be making a wrong assumption.”

Include the enterprise risk management team. Invite the enterprise risk management team into the discussion early on, so it has a clear understanding of the potential risks. With that knowledge, it can help determine what is acceptable, based on a variety of factors, including the industry, specific regulatory climate, and customer demands. 

Follow security-by-design principles. Use the security-shared-responsibility model as an opportunity to address security early on and identify potential gaps. In addition to automation and ensuring that security is code-driven, embrace zero trust and identity and privilege as foundational principles. “By understanding what those gaps are early enough, you can build compensating controls into your environment and make sure it is protected in a way you’d expect it to be,” Leech explains.

Know that visibility is essential. Security monitoring should be a part of the routine to gain a full understanding of what’s happening in the environment. Organizations can opt to do security monitoring on their own or enlist additional services as part of an HPE GreenLake contract. “It goes back to that idea of blissful ignorance,” Leech says. “If I’m not doing any security monitoring, then I never have any security incidents, because I don’t know about them.”

The HPE GreenLake edge-to-cloud platform was designed with zero-trust principles and scalable security as cornerstones of its architecture and development — leveraging common security building blocks, from silicon to cloud, that continuously protect your infrastructure, workloads, and data so you can adapt to increasingly complex threats. For more information, visit https://www.hpe.com/us/en/solutions/security.html


Breaking communication siloes between contact center and non-contact center employees is paramount to improving customer satisfaction, employee engagement, and operating costs. 

The average contact center agent spends 15% of their working day chasing down information needed to serve customers. These hours can add up to a financial loss of $1.5 million annually for a 200-agent contact center, according to Aberdeen. How much is your organization losing?

Who’s handling your company’s customer experience (CX)?  

Customer inquiries today far exceed basic requests that can be completed faster and cheaper via self-service. People now reach out with complicated issues (and prefer not having to repeat themselves about them) that involve multiple parties and/or lines of business. Why should agents be the gatekeeper of complex customer journeys so many others have a stake in? 

Business leaders need to understand and embrace a shared CX ownership model in which every employee is connected and fully vested. 

Actions in three areas make all the difference, and they are within every contact center’s reach:

Better connect departments for faster knowledge sharingEmpower employees to be more productive and creativeQuickly bring in experts to find crucial information

Avaya expects Workstream Collaboration (WSC) will be a key investment this year for achieving these collective CX objectives, as well as for achieving innovation without disruption.  

Create a conversational workspace 

Workstream Collaboration is a cloud-based application that converges siloed unified communications (UC) and contact center technology into one tool with one single application. 

UC capabilities like presence, messaging, screen share, voice, and video conferencing are available directly within contact center architecture so agents can effortlessly and organically communicate with anyone across the organization – whoever they are, wherever they happen to be – and vice versa, with full security and compliance. Integrated with agent desktop applications, there’s no stopping your contact center workers. 

Two scenarios demonstrating this include a parent who doesn’t know their child’s healthcare costs aren’t being covered or a customer who doesn’t understand why he or she has an outstanding balance on his healthcare account. 

You could see who’s available to help using presence (functionality) and then chat about things such as policy changes and customer qualifications in real-time. You could find the person who has the answers to get the information you need, pull them into the customer conversation with you, or have them help you creatively problem solve. 

All of this happens while the customer is on the call, meaning no need for multiple transfers or even hold time depending on how quickly information can be obtained.

Just as importantly, you can innovate your customer experience while retaining your core service. Workstream Collaboration is a fantastic step toward innovation at a pace and path that fits your business needs. 

You’ll improve information-sharing – helping agents get a complete view of situations to make informed decisions quickly – without abandoning current investments. 

By the end of this year, it’s expected that 75% of organizations will be using Workstream Collaboration company-wide to improve CX, productivity, and company culture. 

CX is everyone’s business 

Customers don’t see communication siloes. When they have a bad experience, they blame the brand. This often falls on the contact center agent who is generally considered the voice of the company. This needs to change. Every employee impacts CX in a certain way, not just the agent who says that first “hello.”  

This year, we will see a significant focus on enabling contact center agents to quickly get in touch with employees from different departments while handling customer requests. The ability to consult subject matter experts or even pull them into interactions with customers will become table stakes when it comes to delivering a great experience.

Start connecting different departments for faster knowledge sharing, empowering your employees to be more productive and creative, and making it easier to engage experts. 

Learn more about Avaya’s award-winning Workstream Collaboration solutions.   

Digital Transformation

Are you overthinking your cloud model? If so, you’re likely in need of a well-defined cloud strategy. 

Companies with a clear cloud strategy position themselves to achieve more from cloud computing than those without. A well-defined cloud strategy provides a playbook inclusive of principles, baselines, services, financial models, and prioritization guidelines that enable companies to make informed decisions that support their goals.

In addition, a cloud strategy that is concise, actionable, and reviewed continuously allows organizations to assess current and future states in alignment with security, architecture, governance, compliance, human resources, quote-to-cash processes, and business objectives.

Unfortunately, many organizations mistakenly take their cloud adoption or migration plan as a cloud strategy. Whereas a cloud adoption or migration plan is focused on the “how,” a true cloud strategy is focused on answering the “what” and “why.” A cloud strategy provides a clear decision framework that directly supports business goals and outcomes.

In this article, I’ll share what a cloud strategy looks like, how it helps companies make better decisions, and how to get started.

What a cloud strategy is (and isn’t)

Unfortunately, many misperceptions surround the cloud, as outlined in Gartner’s Top 10 Mistakes in Building a Cloud Strategy. One such misperception is that a cloud strategy is a cloud adoption or migration plan. It’s not. A cloud strategy is not a data center strategy, a cloud-first strategy, or an IT-only strategy. Furthermore, it’s not the execution-level implementation plan many organizations think of. Take a minute to review the top 10 misperceptions that lead to mistakes in building a cloud strategy, as defined per Gartner:


In The Cloud Strategy Cookbook, Gartner defines a cloud strategy as “…a concise viewpoint on the role of cloud computing in your organization.” A true cloud strategy describes the role of the cloud as a business accelerant. It exists to align stakeholders and establish guardrails for decision-making. It includes leaders across the business, including C-level executives. Most of all, it directly supports whatever your business is trying to achieve.

A formal cloud strategy streamlines decision-making, drives better ROI, and reduces risks in strategy execution. A clearly defined cloud strategy helps businesses maximize their cloud investments and align with other synergetic domains of the business. Ultimately, a cloud strategy reduces frustration, disappointment, overspending, and low-value creation.

Getting started with your cloud strategy

If you don’t have a cloud strategy, you’re not alone. Most organizations don’t start with a strategy. Instead, they may begin with a business driver, such as closing a data center or merging with or acquiring another company. They may start with experimentation or heavier adoption in one silo of the organization. In most cases, it’s not until later that they circle back and create a cloud strategy that is not intermingled with adoption or execution planning.

A cloud strategy playbook does not need to be long or written in stone. In fact, it should be relatively short (10-20 pages) and considered a living document —updated regularly to reflect the shifting needs of your business, changes in the market, or anything that affects a significant organizational change.

For example, if your company merges with or acquires another company, that will impact your principles and R-Lane prioritizations. Or, if a new market opportunity comes along, you may need to revisit your cloud strategy to ensure it will help you drive competitive advantage.

The steps for creating a cloud strategy are simple:

Secure executive sponsorship.Gather the right people and set up a cloud strategy office, business office, or steering committee to help ensure your cloud strategy aligns with other domains in the business, such as security, architecture, data center, compliance, finance, IT service management, HR, and legal.Finally, host an iterative assessment series with workshop sessions to define or refine your cloud strategy, including principles and alignment to other domains (architecture, security, procurement, HR, etc.). Establish a cadence to review it regularly.

10 questions your cloud strategy should answer

As part of your cloud strategy definition, you should consider assessment and hands-on workshop sessions to define guiding principles. To do so, be sure that you answer the following 10 foundationally strategic questions:

What is your delivery and operational model? Align your workload and business objectives to a public, hybrid, private, multi-cloud, distributed, or smart cloud (workload-by-workload decision).What is your service model prioritization guideline (for SaaS, PaaS, IaaS, or XaaS)? Use guiding principles to define your service models based on defined guidelines.What is your consumption/development model? An example might be buying SaaS before building.What is your cloud deployment model? Will it be hybrid, multi-cloud, or cloud-native, distributed cloud?What is your R-Lane prioritization model? For example, define if lift-and-shift is the preferred option or last resort as a principle. Should a long-term and short-term R-Lane vision be defined for each workload? For instance, one workload can be a rehost in the short term but a re-architect/refactor in the long term.What is your FinOps model? Charge back or show back? Proportional or even split?What is your workload analysis model? Is it a workload-by-workload analysis or a big-bang analysis? How sophisticated are the inventory capabilities for workload-by-workload exercises (done outside of the strategy workstream itself)?What is your data center (or off-cloud) strategy?What is your cloud exit strategy?What is your level of strategic alignment with other parts of the business?

Define your cloud strategy with help from GDT

A solid cloud strategy helps ensure your organization makes suitable investments to support your business and streamlines decision-making around your cloud model.

With decades of experience, GDT can help you define your cloud strategy and create a cloud strategy playbook based on Gartner’s principles and best practices. To learn more, contact the cloud experts at GDT today.

Cloud Computing

Differentiating your brand in the telecommunications market is hard—just ask Vodafone’s CDIOScott Petty.

Despite the British multinational telco’s continued investments in fibre and 5G, and growing consumption of broadband and cellular services, intense price competition, rising energy prices, market regulation and economic headwinds have made for an industry where single-digit revenue growth is difficult to find.

But with a new insourced operating model and a ‘mixed matrix’ team—which gives autonomy centrally and to local markets—Petty believes Vodafone is on a journey to become a next-gen telco.

Vodafone’s digital transformation strategy

Over the last three years, Vodafone has used digital transformation as a centrepiece for how it unlocks efficiencies and new growth opportunities.

Underpinned by its 2025 tech strategy—which commits to building in-house scaled platforms, launching products and services to market faster, building a talent pipeline, and establishing Vodafone as a gigabit broadband leader—the firm’s digital transformation strategy has three pillars.

Operational efficiency: The company is leveraging automation and observability tools to improve efficiency, as well as data analytics, machine learning (ML) and AI through Google Cloud Platform to make better decisions and find incremental savings. An example of the latter is its use of ML and AI to overlay data—such as customer movements from social media as well as profitability—to find out where to deploy new base stations. Dubbed ‘smart capex’, Petty says Vodafone has found a 15% efficiency and £500m in cost savings through this since 2019.Digital channels: Petty says digital channels enable Vodafone to “sell more, service better and reduce costs”. The company app, chatbots and websites have all driven opportunities to better serve customers, particularly through retail-specific events, such as the launch of Apple’s iPhone 14, where web traffic can grow 500% within an hour.New platform businesses: Created in parallel to its legacy connectivity business, new platform businesses aim to drive revenue growth. Vodafone now has an IoT platform connecting 150 million devices, including connected cars, smart meters and smart home monitoring systems. In Africa, it claims to run the largest payment platform, while it’s adding new financial servicing in Europe—such as handset insurance—to wrap around standard tariffs. Petty says that 10% of total company revenue is now coming from these platform businesses.

New operating model, a mixed matrix team

The key to achieving transformation at scale, and delivering better products and services for customers, has been the creation of a European digital and IT organisation that combines the expertise of local markets with the breadth and scale of centralisation.

Last April, Vodafone Technology created a pan-European digital & IT organisation with nearly 17,000 employees led by Petty, bringing together IT functions from 12 European markets to work as one team.

Formerly CTO for Vodafone UK, Petty says Vodafone previously operated independent sets within countries, so each market was responsible for their region, with shared services for data centres and platforms, such as ERP. But this up and down model made code reuse difficult. Historically, scaling across markets was also complicated, with many overlapping projects, high expenditure and limitations in how many projects Vodafone could run simultaneously.

Vodafone didn’t move to centralisation—a model where you don’t have “proximity to the local market”—but rather to a ‘mixed matrix’ model, where each operational local CIO had a broader technology role for the organisation in addition to their local jurisdiction.

“Our UK CIO runs everything in the UK, but also digital engineering for all of Europe,” says Petty. “Our Spanish CIO runs digital channels and the Spanish technology. That model is replicated down into the organisation, so everyone has an operational function close to the markets they operate in, and a domain function, which is trying to execute our strategy. That creates interlocks between each other.” He also highlights that, for instance, the Italian CIO responsible across Europe for data and analytics needs analytics to run well in the UK, and to have a close working relationship with his UK counterpart to ensure digital engineering operates effectively in Italy.

Insourcing and reuse in engineering

A key component of this strategy was Vodafone’s approach to insourcing. Petty says the telecom provider was previously 30% insourced, and 70% outsourced, with a heavy reliance on system integrators and third parties. He says that model isn’t effective and doesn’t scale, with system integrators not incentivised to build replicable platforms.

“Our belief is you have to have a significant portion of your own engineering,” he told CIO.com. “There will always be a role for systems integrators, but integrating into our platforms and standards, not a free-for-all.”

Petty also says that investing in in-house engineering has an upfront benefit as its cheaper, but that it also gives the organisation the technical and engineering talent to build “horizontal scale”.

Vodafone, which has 13,000 software engineers with ambitions to reach 16,000 by 2025, sees a greater willingness for teams to share and reuse through this same model.

“They are more open to sharing and reusing,” he says. “And as they replicate that model in their teams, you just create the connections. Ultimately, culture is about humans and how they interact.”

Petty says staff across the European markets will volunteer their time for projects, such as building APIs for new cloud platforms, sensing the opportunity to do interesting work and develop their core skills. It’s this collaborative approach that’s also improving efficiency. He contrasts the new approach with how the Vodafone UK team once operated. Spending approximately £80 million each year on digital investments, and with 1,000 engineers developing capabilities in Spotify-style squads, the company experienced more defects, cybersecurity issues and slower business speed as the number of projects increased.

The new operating model and team restructure saw technology squads split into technical functions, such as cloud engineering and economics, with these squads responsible for service maintenance, cybersecurity quality and achieving performance across the various European markets.

“Taking this approach dramatically improved our velocity,” says Petty. “Our biggest supporter of hyper-scaling is our CFO because he sees that the more money he pours into engineering, it comes back as increased velocity and capacity. The business can go faster.”

Petty gives the example of Vodafone’s rewards app. Previously, the mere difference between a latte in the UK and café con leche in Spain would have meant the two regions building separate solutions. Under the new reuse model, the rewards platform was developed 100% in-house in Turkey, then launched in Portugal inside three months. The next market, Germany, took just one sprint to bring it to market.

He says the cost difference is significant too, explaining that while the first market spends 110% of their capacity (10% more to make the platform reusable), the payback comes as each subsequent market adopter spends significantly less.

“You start to get a flywheel effect on creating that reuse across your platforms.”

CDIO becomes CTO

Despite progress, there are hurdles ahead for Petty. He admits the skills shortage remains a critical issue, particularly in data science, and there’s the migration of on-premise platforms to Google Cloud Platform (GCP), through which Vodafone hopes to unlock more data and release new capability. Investment must continue on fibre and 5G through recession, too, while Vodafone has started exploring quantum-safe cybersecurity with IBM.

There are deeper changes in the boardroom as well. In early December, Vodafone announced its CEO of nearly 20 years, Nick Read, would be stepping down with finance director Margherita Della Valle taking over as an interim replacement. And with global CTO Johan Wibergh retiring in December 2022, a new role beckons for Petty, too, as he starts as the new CTO in early 2023. Together with Alberto Ripepi, the new Chief Network Officer, he will join the executive committee and expected to co-lead the Vodafone Technology group.

Petty says the hires were down in part to the success of the insourcing model, and a realisation that technology executives can challenge the business. “When we changed the operating model, we saw benefits for technology to challenge the business a little bit more,” he says. “You can work as a team to try and drive the right outcome.”

CTO, Data Engineering, IT Leadership, IT Operations, IT Skills, Operating Systems

What is transformational leadership?

The transformational leadership approach encourages, inspires, and motivates employees to innovate and create the change necessary to shape the future success of the company. This is accomplished by setting an example at the executive level through authenticity, a strong sense of corporate culture, employee ownership, and independence in the workplace. Transformational leaders are change agents in the business, who can identify innovative and shifting trends in technology, and then help the organization embrace that change.

Transformational leaders inspire and motivate their workforce without micromanaging — they trust trained employees to take authority over decisions in their assigned jobs. It’s a management style that’s designed to give employees more room to be creative, look to the future, and find new solutions to old problems. Employees on the leadership track will also be prepared to become transformational leaders themselves through mentorship and training.

[ Find out how to apply transformational leadership at your organization and what it means to be a transformational leader today. | For the latest insights on leadership, sign up for our newsletters. ]

Transformational leadership theory

The concept of transformational leadership started with James V. Downton in 1973 and was expanded by James Burns in 1978. In 1985, researcher Bernard M. Bass further expanded the concept to include ways for measuring the success of transformational leadership. This model encourages leaders to demonstrate authentic, strong leadership with the idea that employees will be inspired to follow suit.

While Bass’ transformational leadership theory dates to the ’70s, it’s still an effective leadership model practiced today — this style of authentic leadership never changes, just the environments it’s used in. It’s applicable across every industry, but it’s especially vital to the fast-paced tech industry where innovation and agility can make or break a company.

For a deeper look at the transformational leadership model, see “How to apply transformational leadership at your company.”

Transformational leadership model

Four main elements define the transformational leadership model and style. These factors were developed by Bass in 1985 to help define what transformational leadership looks like and how to be successful as this type of leader:

Idealized influence: The most important thing you can do as a transformational leader is to lead by example. Employees will look to you as a role model for behavior in all areas of the workplace. If you lead with authenticity, employees will pick up on that behavior and feel inspired to maintain that high standard for performance. It’s not about manipulating employees into working hard, it’s about leading by example and positively influencing others through a commitment to trust, transparency, and respect.Intellectual stimulation: To help create change, it’s important to challenge long-standing beliefs in the company and push the status quo by encouraging innovation, creativity, critical thinking, and problem-solving. Transformative leaders should help employees feel comfortable exploring new ideas and opportunities that can inject innovation into the organization. You want to establish an environment that welcomes growth and gets everyone excited about digital transformation and other important initiatives in the organization.Inspirational motivation: As a transformational leader, you will need to encourage your team to feel attached and committed to the vision of the organization. You want to ensure employees feel as committed to these goals as you do as a leader by giving employees a strong sense purpose, rather than attempting to motivate them through fear.Individual consideration: Employees need to feel a sense of independence and ownership in the overall business goals. As a transformational leader, it’s important to understand every employee is a unique person within the company and will have specific needs, mentorship styles, and their own contributions to the company. These leaders will tailor their coaching and mentorship styles to the employee and help them reach goals both inside and outside of the organization.

Transformational leadership characteristics

Businesses want transformational leaders who display an “executive presence,” according to David E. Ulicne, senior director of executive education at Carnegie Mellon University’s Heinz College of Information Systems and Public Policy.

“The most important attribute for the success of the CIO would be their ability to articulate their vision and gain support of key stakeholders.” It’s a skill that takes practice, and you’ll need to “build your storytelling skills” by presenting at conferences and symposiums, and even recording yourself delivering speeches and asking your PR or media team for “constructive feedback.”

For a deeper look at leadership traits, see “7 leadership traits major enterprises look for in a CIO.”

According to Bass’ model, transformational leaders set themselves apart from other types of leaders by doing the following:

Encouraging the motivation and positive development of followersExemplifying moral standards within the organization and encouraging the same of othersFostering an ethical work environment with clear values, priorities, and standardsBuilding company culture by encouraging employees to move from an attitude of self-interest to a mindset where they are working for the common goodHolding an emphasis on authenticity, cooperation, and open communicationProviding coaching and mentoring but allowing employees to make decisions and take ownership of tasks

For a look at how to draw attention to your transformational leadership qualities in your resume, see “IT resume makeover: Highlighting transformational leadership.”

Transformational leadership examples

Harvard Business Review analyzed companies on the S&P and Fortune Global 500 list to uncover the best examples of transformational leadership. These businesses were judged on “new products, services and business models; repositioning its core business; and financial performance.”

Jeff Bezos, Amazon: Harvard Business Review attribute’s Bezos’ “insider, outsider” status as part of what makes him a great transformational leader. As someone who jumped from the finance world, he brought a fresh perspective to e-commerce through years of experience in a different industry.Reed Hastings, Netflix: Hastings tied for first alongside Bezos, and for similar reasons. Hailing from the software industry, he wasn’t rooted in pre-established process and procedure in the television industry.Jeff Boyd and Glenn Fogel, Priceline: Boyd and Fogel reinvented travel reservations by charging lower commission fees on reservations, but focused on smaller niche markets (inns, B&Bs, and apartments), eventually spawning Booking.com.Steve Jobs and Tim Cook, Apple: HBR points to Apple as an example of “dual transformation”: Jobs innovated on original Microsoft products while also building a software ecosystem. Cook has extended on Jobs’ vision, maintaining a focus on innovation, software, and brand loyalty.Mark Bertolini, Aetna: Bertolini is known for his realistic management approach in the healthcare industry. He says his goal is to build strategies around a realistic vision of the future.Kent Thiry, DaVita: Thiry managed to take a bankrupt company and turn it into a thriving business through firm core values that included “service excellence, teamwork, accountability and fun,” according to Harvard Business Review.Satya Nadella, Microsoft: Nadella started at Microsoft in 1992 and worked his way up the corporate ladder, eventually running the business’ cloud computing efforts, which landed him the executive position.Emmanuel Faber, Danone: Faber started out as an architect for Danone and earned the CEO job after he helped develop the company’s vision to turn the company into a sustainable health and nutrition company.Heinrich Hiesinger, ThyssenKrupp: Hiesinger become CEO of ThyssenKrupp in 2011 and helped alleviate pressure from Asian competitors in the steel market by embracing newer forms of manufacturing, including 3D printing — “new growth areas” that now make up 47% of the business’ sales.

Transformational leadership in IT

Although the concept of transformational leadership can apply to every industry — including healthcare, education, and government agencies — it’s increasingly important in IT as companies embrace digital transformation. Adapting to rapidly changing technology requires innovation and strong leadership to stay ahead of the curve and to remain competitive.

As leaders in IT, CIOs are responsible for setting the example as transformative leaders — especially considering they’re largely responsible for the digital transformation of the organization. Gartner reports that 40% of CIOs are leaders of digital transformation in their organization, while 34% say they’re responsible for innovation. Inspiring and motivating employees is an important aspect of digital transformation success, as everyone must buy into and embrace growth and change.

Digital transformation has become an even bigger priority since the COVID-19 pandemic changed the way companies operate across every industry. According to the Digital Leadership Report 2021 from Harvey Nash Group, the pandemic sparked a “time of unprecedented change and transformation of digital at the majority of organizations.” While this digital shift began years ago, the pandemic only accelerated it through an increased demand for remote work and significant changes in supply chains and customer behavior. 

In 2021, 50% of digital leaders said they expect major or radical changes to products and services, 47% planned to unlock new value through digital, 43% were tasked with supporting innovation, and 48% had expectations to transform and digitize the enterprise, according to data from Harvey Nash Group.

While there is certainly a growing need to keep an eye on the future — whether it’s security, new technology, or shifting platforms — not every part of IT will benefit from transformational leadership. Some processes, procedures and development projects require more structure, consistency and reliability; this is called transactional leadership.

To read how CIOs are making good on transformational leadership, see “Transformational CIOs juggle innovation and operations.”

Transactional vs. transformational leadership

Transactional leadership is the exact opposite of transformational leadership — it relies on motivating employees through rewards and punishments. It requires supervision, oversight, organization, and performance-monitoring. This leadership model doesn’t try to innovate. Instead, it’s rooted in keeping things consistent and predictable over time. Errors and faults are closely investigated, and the overall goal is to create efficient, routine procedures.

This style is best suited to departments or organizations that require routine and structure — areas where businesses want to reduce chaos or inefficiency. But it doesn’t allow for innovation or future planning the same way transformational leadership will.

Transformational leadership, on the other hand, supports agile environments, especially where failure carries less risk. You want the development and maintenance of a current product to remain consistent and error free, but you don’t want that to hinder the progress and growth of future updates and improvements.

Transactional leadership takes care of creating a consistent development process, while transformational leadership leaves people free to come up with new ideas and look at the future of products, services and ideas.

Training and certification

Although transformational leadership skills are considered soft skills, there are still plenty of resources, certificates, and training programs aimed at developing transformational leaders. Here’s a short list of online resources to earn training and certifications in transformational leadership. However, it’s likely you can find an in-person training course at a local college or university in your home state.

Georgetown University School of Continuing Studies: Institute for Transformational LeadershipUniversity of Notre Dame: Transformational Leaders ProgramCoursera: Transformational leadership: Developing othersUniversity of Washington: Transformational Leadership AcademyThe Great Courses: Transformational Leadership: How leaders change teams, companies and organizationsThe University of Texas at Dallas: Transformational Leadership Certificate programUdemy: Transformational Leadership and Leadership Development TrainingBinghamton University School of Management: Transformational Leaders Program

More on transformational leadership

How to apply transformational leadership at your companyWhat does it mean to be a transformational IT leader today?15 leadership certificate programs to boost your career10 IT leadership development programs to help you level upThe 9 new rules of IT leadershipIT Leadership

Following increasing numbers of cyberattacks on Western energy companies, Hydro-Québec decided to conduct a world-first “electrical containment” exercise this summer.

For four hours, the state-owned company managed to completely isolate itself from the internet without sustaining any service failure. A reassuring exercise for its customers in Québec but also for those in New Brunswick, Ontario, New England, and New York State – where it fulfills up to 15% of power needs.

Where and when did this experiment take place?

“We have to keep that a secret, but I can tell you it made me very nervous,” says Jean-François Morin, VP – Information and Communications Technologies, who oversaw the exercise from start to finish. “What kept me awake was forgetting a machine somewhere or cutting off some customers by mistake.”

This secret exercise is more than just a cybersecurity milestone: it’s part of Hydro-Québec’s plan to push its digital shift – a key component of its 2022-2026 strategic plan.

“Our meters generate one billion bits of data per day, and this number will go up very soon. We’re developing huge models to manage all this information and make it talk. But it won’t work unless our cybersecurity is foolproof.”

Ethan Cohen, Gartner VP and analyst for Utility Transformation and Innovation, doesn’t conceal his admiration for the government-owned corporation.

“There is an element of show in such an exercise, but it demonstrates a level of competence that most utilities would like to have. It shows that Hydro-Québec is not merely sustainable, but resilient,” Cohen says.

“A lot of utilities formulate grand strategies, but the issue is executing them. What matters is that the CIO actually does it.”

Transformation, phase two

For Hydro-Québec, its current digital transformation is the second phase of an energy transition plan that began in the 1970s, when the utility was able to use its large production of green hydropower to satisfy almost all residential demand, including heating. But today, the company needs to work on the government’s new requirement – electrifying all transportation by 2040.

Jean-François Morin, VP – Information and Communications Technologies, Hydro Québec

Hydro Québec

“To reach this goal, we’ll have to build an IT ecosystem that allows us to better predict and control power demand, but also to use the production capacities of residential, commercial and industrial customers,” Morin says. “There’s going to be AI everywhere – and our operators are going to become computer specialists.”

Hydro-Québec produces, moves, and distributes about 180 terawatt-hours per year – more than almost any other utility in North America. The digital transition will be on a similar scale.

“Of all viable solutions,” Morin says, “the most profitable will be the use of data. Analytics will help optimize maintenance and consumption but also automate production and decision-making, including the analysis of new infrastructure projects.”

The Université du Québec and Université de Sherbrooke computer science and finance graduate gives the example of maintenance, which costs billions every year – and currently follows “blind” procedures.

“We’re now changing parts that we don’t really have to change. By putting sensors everywhere, we’ll be able to control what’s going on, to collect histories, and intervene where it’s really needed.”

Most of the utility’s 20,000 employees’ jobs will be affected in some way by the digital shift. With the installation of smart meters, the job of meter reader has already disappeared, although Hydro Québec hired more computer technicians. Among other things, this new technology made it possible to better detect energy theft by analyzing clients’ consumption in real time and comparing it to that of neighbours.

“That was just the beginning,” Morin says. “One of my roles is to identify the jobs of the future, what we’ll need in terms of AI and IT specialists.”

“Hydro-Québec has a tradition of innovation and R&D,” Cohen says. “It’s a very entrepreneurial organization at a level you don’t see elsewhere. And they’re willing to shake up the way they do things to achieve real breakthroughs.”

Eliminating waste

The digital shift will be key for solving Hydro-Québec’s biggest problem: waste. In a way, the company is the victim of its own success: because it offers North America’s cheapest and greenest energy in massive quantities, it has created a class of ultra-dependent, hyper-hungry consumers who are devouring energy that could be put to better use – to electrify transportation and industrial processes, for example.

“There are people warming their driveway to melt the snow and heating their outdoor Jacuzzis all week long in the winter,” says Morin, who will play a key role in planning how to get customers to use power more efficiently, especially during peak hours – which are very costly.

Hydro’s residential customers currently receive no warning about the actual cost of their unbridled consumption. “We need to develop ways to better inform them of their use. Somewhat like Tesla, which is very good at telling its clients how much they saved in their journey. My dream would be for all customers to get notifications at 4 p.m. about what they’re paying for and how much they’d save by turning off their pool or water heater for a few hours.”

Such goals are not limited to data management; infrared drones, for example, could produce a stunning view of the consumption profile of the most energy-hungry customers. Pricing would also be a powerful awareness-raising tool if changes in habits are matched by real savings, says the company’s VP, who has worked his way up through the ranks since starting as an IT project manager in 1999.

The transition won’t apply only to IT processes: Morin’s office is now involved in all the company’s fundamental policy decisions. Hydro-Québec will have to meet major power needs in the next 50 years, and its management has promised to consider all possible avenues to delay the building of new hydroelectric megaprojects.

“We’re juggling a lot of new ideas, like scaling existing dams, putting in more efficient turbines, but also solar roofs, which look good, or small residential windmills that would allow customers to produce their own electricity and even power the grid at certain times.”

“Hydro-Québec may be further along in the energy transition than other utilities, but they still have to address the current need for self-production,” Cohen says. “There are many new opportunities that have to be analyzed.”

A radical, careful transition

Morin believes Hydro-Québec could move much faster in its digital shift, but he’s holding back on purpose because of cybersecurity and privacy issues. This cautious approach, according to Cohen, is not a bad thing: “There would be big advantages to moving fast but for utilities, the regulatory environment is an inescapable reality.”

A company that sells electrons is particularly susceptible to “malicious” electrons. The more sensors or home automation services it provides, the more exposed it becomes to hackers.

“We need to think ahead,” Morin says. “We could monitor consumption in every home. We could hire bigwigs in Paris and New York to work remotely. But our participation in the North American energy market requires us to comply with strict reliability rules.”

Now that he has successfully achieved the ultimate electric containment, Morin believes cybersecurity will be one of the first AI applications in grid management and control.

He also must deal with Québec’s new data privacy regulations, the most advanced in North America. Like any company that manages information about Québec’s residents, Hydro has to guarantee that these data are protected. Neglecting to do so can be very costly – up to 4% of a company’s worldwide revenue.

Legal and Regulatory Affairs must therefore validate Morin’s decisions. “Do I have the right to use such and such data for such and such application?” he asks – explaining that he had to deliberately slow down the expansion of the Hilo smart-home subsidiary because of these issues.

“We need very robust data governance to make sure we comply with the law but also to determine what data is actually useful. Even with AI, the old law of computing applies: ‘garbage in, garbage out’.”

Translation by Daniel Pérusse

Digital Transformation, IT Leadership, Utilities, Utilities Industry

In March 2022, Steve Demetriou, chair and CEO of Jacobs, announced a new multiyear strategy that would focus on three accelerators for the $15B professional services business: climate response, consulting and advisory, and data solutions, each supported by significant acquisitions.

Madhuri Andrews, who became chief digital and information officer of Jacobs in 2019, knew the strategy would require a new IT operating model, both to enable IT to act as a truly global organization and to leverage the IT team as business leaders in the company’s acquisitions-fueled transformational journey.

“We had previously acquired a business whose water and environmental focus is critical to addressing climate change. We also divested our oil and gas business, which was a big part of our history, and we made some targeted acquisitions that enhanced our advisory and data capabilities,” says Andrews. “In shifting our portfolio to help our clients with their own transformations, we needed to shift our IT operating model to support the new portfolio.”

Three modes of IT

To achieve this, Andrews began with an overarching vision of IT working in three modes. The first mode she defined as helping Jacobs’ customers transform their own businesses. “Our clients know that they need to transform if they are going to emerge stronger than before from the digital disruption that is happening in the market,” she says. “We want to help our clients disrupt their own industries.”

For example, during the pandemic, Jacobs’ life sciences team worked with a large pharmaceutical client to speed up vaccine production from 12 months to 12 weeks. “We helped them redesign their manufacturing processes using digital twins, iterative design methodologies, and advanced analytics,” says Andrews. “We leveraged our global integrated delivery footprint, with employees around the world working together on accelerated timelines, to partner with our client and make this happen.”

Andrews’ second mode for Jacobs IT is to support the company’s global operations, which involves a high level of collaboration to leverage and develop talent fluidly across a global footprint.

To enable this, Andrews and her team have created a virtual workspace capability that can be scaled to allow more than 20,000 engineers and data scientists around the world to collaborate in real-time on complex engineering projects.

“With the virtual workspace foundation, we’ve been able to move faster on automated and generative design,” Andrews says. “We are leveraging agile, cloud, advanced analytics, and have a great platform for experimenting with emerging technologies such as blockchain and Web 3.0. It is exciting for our people to engage differently across the global organization and learn new technologies.”

The third mode is about the IT organization itself, which has had to transform to make the other modes happen. “In the past, many of our businesses, like aerospace and defense, had their own IT organizations and platforms,” says Andrews. “To help leverage Jacobs’ global workforce and capabilities at scale for our customers, we needed to operate as one global IT organization – utilizing a mix of common platforms, processes, and tools that are informed by our operational and customer needs. This has also created opportunities for our people to expand their knowledge of our lines of business and bring innovative ideas on using technology to address our operational and customer needs.”

Business success teams

As a key element of the new operating model, Andrews has introduced “business success teams,” each dedicated to a different line of business to work alongside leaders of foundational IT capabilities such as common enterprise platforms, cybersecurity, data, and advanced analytics.

“When I do organizational design, I think of a spectrum,” says Andrews. “We know that all roles require soft skills in order to collaborate and innovate effectively. However, some roles are heavily relationship-driven, whereas others are very technical. The business success teams understand enough about technology to know how to use it to solve business problems or create opportunity, but they don’t always need to have in-depth expertise on Java, Jira, or cloud computing.”

To make the business success teams work, Andrews resisted a one-size-fits-all model.
“Business success teams range from 10 people to 100-plus,” she says. “The small teams work with businesses that have stable operations where we are engaged in continuous improvement activities, and the large teams work with businesses that are specialized or driving extensive changes in how technology is transforming the work they do.”

Clarity of purpose

For Andrews, driving major operational change across a large complex global organization required clarity of purpose and focused communication. “As we were creating our new IT operating model, we worked with a diverse cross-section of our organization to develop a purpose statement that would resonate with our employees and clearly connect to our corporate strategy. In addition, we developed six themes such as creation of frictionless employee and customer experiences, driving a seamless data fabric to harness the power of data business insights and flexible partnerships to increase the velocity of our business, all of which ensure that the IT operations support our business strategy,” she says.

Andrews and her team then tied all current and future IT initiatives into at least one of those themes. “This way, our people can pull the thread through the needle and see how their work actually impacts the business,” she adds.

Andrews sees the changes she is making in the IT organization as reflective of the role of a CIO. “In some IT organizations, our people advance because they are really good at execution, but they don’t necessarily understand how to adapt or drive change,” she says. “As CDIO, my role is both as a consultative partner to the business and as a mentor to the IT organization. If I can connect people to the purpose, they will think more creatively. I can go into the weeds on any technology, but my more important role is to ensure that the work we are doing in IT is connected to our overall Jacobs strategy.”

Business IT Alignment, IT Leadership, IT Strategy

Many companies that begin their AI projects in the cloud often reach a point when cost and time variables become issues. That’s typically due to the exponential growth in dataset size and complexity of AI models.

“In an early phase, you might submit a job to the cloud where a training run would execute and the AI model would converge quickly,” says Tony Paikeday, senior director of AI systems at NVIDIA. “But as models and datasets grow, there’s a stifling effect associated with the escalating compute cost and time. Developers find that a training job now takes many hours or even days, and in the case of some language models, it could take many weeks. What used to be fast, iterative model prototyping, grinds to a halt and creative exploration starts to get stifled.”

This inflection point related to the increasing amount of time needed for AI model training — as well as increasing costs around data gravity and compute cycles — spurs many companies to adopt a hybridized approach and move their AI projects from the cloud back to an on-premises infrastructure or one that’s colocated with their data lake.

But there’s an additional trap that many companies might encounter. Paikeday says it occurs if they choose to build such infrastructure themselves or repurpose existing IT infrastructure instead of going to a purpose-built architecture designed specifically for AI.

“The IT team might say, ‘We have lots of servers, let’s just configure them with GPUs and throw these jobs at them’,” he says. “But then they realize it’s not the same as a system that is designed specifically to train AI models at scale, across a cluster that’s optimized to deliver results in minutes instead of weeks.”

With AI development, companies need fast ROI, by ensuring data scientists are working on the right things. “You’re paying a lot of money for data-science talent,” Paikeday says. “The more time they spend not doing data science — like waiting on a training run, troubleshooting software, or talking to network, storage or server vendors to solve an issue — that’s lost money and a lot of sweat equity that has nothing to do with creating models that deliver business value.”

That’s a significant benefit of a purpose-built appliance for AI models that can be installed on premises or in a colocation facility. For example, NVIDIA’s DGX A100 is meant to be unpacked, plugged in and powered-up enabling data scientists to be productive within hours, instead of weeks. The DGX system offers companies five key benefits to scale AI development:

A hardware design that is optimized for AI, along with parallelism throughout the architecture to efficiently distribute computational work across all the GPUs and DGX systems connected together. It’s not just a system; it’s an infrastructure that scales to any size problem.A field-proven, fully integrated AI software stack including drivers, libraries and AI frameworks that are optimized to work seamlessly together.A turnkey, integrated data center solution that companies can buy from their favorite value-added reseller that brings together compute, storage, networking, software and consultants to get things up and running quickly.The DGX system is a platform, not just a box, from a company that specializes in AI, and has already created state-of-the-art models, including natural language processing, recommender systems, autonomous systems, and more — all of which are continually being improved by the NVIDIA team and made available to every DGX customer.“DGXperts” bring AI-fluency and know-how, giving guidance on the best way to build a model, solve a challenge, or just assist a customer that is working on an AI project.

When it’s time to move an AI project from exploration to a production application, the right choice can speed and scale the ROI of your AI investment.

Discover how NVIDIA DGX A100, powered by NVIDIA A100 Tensor Core GPUs and AMD EPYC CPUs, meets the unique demands of AI.

Artificial Intelligence, IT Leadership