DXC and Dell: Harnessing GenAI to reshape the enterprise

Welcome to the second in a series of three interviews conducted with leading executives from both DXC Technology and Dell Technologies, which explores some of the next-generation technologies in focus in 2024.

In this conversation, DXC Chief Architect, Data & AI Charlie Wardell and Dell Global Alliances CTO for AI and Data Management Rob Shear discuss the potential of GenAI to transform businesses, and how DXC and Dell — along with NVIDIA — can help you on the journey.

DXC and Dell: Partnering for innovation (Part 1)

Making multicloud part of your enterprise AI/GenAI strategy (Part 3)

Q: What impact will the AI/GenAI revolution have on customers?

Charlie Wardell: If you listen to NVIDIA founder and CEO Jensen Huang about the Blackwell chip that was just announced, everything is about to change. It's going to radically change the way we construct systems, data centers, and even the way we program. Companies will need a strategy to capitalize on the future of GenAI technology.

Today, many companies are adopting tip-of-the-spear AI solutions. The risk of this approach is that they miss an overall strategic play that could benefit their organization and introduce the risk of AI sprawl; AI solutions, typically models, are derived from data and, as such, need to be governed as data. GenAI driven by corporate data needs to be validated, protected and run through the same rigor as corporate data is.

The cloud has been absolutely phenomenal with regard to agility, but I think the AI/GenAI revolution will drive a cloud repatriation, where organizations start moving critical data back into the locality of on-prem GPUs where they can be closely managed and optimized for cost. I don't anticipate a mass exodus from cloud, but I do think that organizations will need to choose the best approach based on the size, security and velocity of the data. For some, it will make more sense to keep their data on prem, very close to the applications that create it, while others may want the flexibility and high burst that works well in the cloud. DXC’s Enterprise Intelligence Systems (EIS) strategy can support any approach (cloud, on-prem, edge or hybrid).

Rob Shear: The impact of GenAI to clients is profound. Every large and medium-sized company is looking at starting projects, both internal and external. The opportunity for productivity gains is probably the largest we've seen since the late 1990s/early 2000s.

To Charlie’s point about cloud repatriation, Dell's view is that you should bring AI to your data, not the reverse. If the bulk of your enterprise data is on prem and in various systems of record around your company, trying to relocate that to the cloud is very expensive. Running GenAI in the cloud — especially in the long-term — is very expensive. Standing up a POC in the cloud might seem quick and easy, but as you move to production the costs will escalate.
 

Q: What are the individual strengths of DXC and Dell in AI and GenAI, and how can these be combined to transform customers’ businesses?

Rob Shear: Dell brings a great portfolio of solutions, including all our best-of-breed hardware and software. GenAI is use case-driven, so there is a large consulting side to GenAI engagements. That's where DXC comes in. DXC takes our hardware and platforms, and ultimately brings business value to end users.

In fact, the power of Dell with DXC and all our partners is that we bring a full portfolio of servers, storage and solutions, as well as our Dell-validated designs for training, inferencing and retrieval-augmented generation. Our Dell Data Lake House makes it easier and quicker to get access to your data.

In this new world, data is king. Companies that can harness their data to bring it together and feed it into their AI deployments are going to have a huge advantage. It’s very easy to do siloed GenAI deployments and get something out. But having single business units doing their own thing in an organization is not going to lead to success. That’s where DXC’s Enterprise Intelligence Systems (EIS) approach comes in. DXC approaches these projects with a longer-term, whole-enterprise view. That is a much better strategy.
 

Q: Tell us more about DXC's EIS approach.

Charlie Wardell: Our EIS approach is based on research that IDC did around an enterprise intelligence strategy based on the premises of ingesting, assimilating and presenting data. EIS is really more focused on an enterprise gaze, creating a data-literate and data-driven organization. The underlying component of GenAI solutions is data, and that data has to be managed. You still must do the same data blocking and tackling to create, fine tune or augment those models. These models need to be integrated, so there is also a strategic engineering component to this, where we want the output of these models to be integrated into other systems, such as SAP and Workday.

Where DXC really excels and adds value through our EIS approach is that we know how to move data from point A to point B. We know how to manage, synthesize, organize and govern it, and we can create that interface layer so that data can be integrated into other systems as well. Our EIS strategy takes it a step further and aims to incorporate data literacy, data governance and everything else that's involved with data, with AI at its core. So we are promoting a standardized approach to AI.
 

Q: What about AI governance? Is it important, and what structures should companies put in place?

Charlie Wardell: I do believe that there will be regulatory and compliance requirements for AI outside of AI platform governance. But as for corporate AI models, these need to be catalogued, tagged, documented and protected just like any other digital object in an organization. Doing so drives data literacy within an organization. Understanding the lineage of the data involved in creating, fine-tuning or augmenting models would ensure compliance to governance, privacy laws and taking into consideration issues like the right to be forgotten, HIPAA-compliance in healthcare, or AI best practices that an organization’s Office of AI would publish. 

Rob Shear: I prefer the term best practices, and these are emerging. I think regulation is going to take a long time, because it's difficult for legislators to keep up with the speed of GenAI. I don't know if we're even ready for regulation yet because we need the industry to continue to mature. So, best practices around data governance and the quality of data does matter. Many companies store a lot of data that is outdated and no longer relevant. We have to ensure that it is not brought into the models. So when we're working with customers and building out an enterprise strategy, we identify relevant and current data for the input. We put some governance around it, ensuring that the output is accurate. That’s especially important for externally-facing GenAI projects, where there is no margin for error. 

Q: Are there industries that are better suited for GenAI projects, and what are some of the use cases for these?

Charlie Wardell: The real value of GenAI is when you can train it with the depth and breadth of the organizational data and achieve insights that you've never had before. Whether it be used for hospitals improving operational efficiencies, or corporations understanding their supply chain risks, this is where AI gets extremely valuable as a strategy.

There are insurance industry organizations that, with the help of domain subject experts, are identifying claims that could be better handled by AI, while routing questionable ones to be validated by experts. So there are ways of improving operational efficiencies within organizations.

We recently set up managed services for an international insurance firm that is based on the Dell AI Factory and NVIDIA's Enterprise AI suite. I’m also working with a global pharmaceutical company that is very interested in a more comprehensive enterprise AI strategy for their data. In the banking and capital markets sector, I am demonstrating how GenAI can be used to automate research, identify risk and optimize investments. 

Rob Shear: I think every major industry is looking for use cases for GenAI, including some of the more regulated industries. For example, financial organizations are looking at how GenAI can help them generate some of the vast regulatory documentation that usually takes months to produce. It is these industry-specific use cases that show a very quick ROI for companies.

And if you look at, say, patient outcomes in a hospital setting, there is a tremendous amount of data generated that is not currently leveraged. It might show up on a screen at a nursing station. But it's not captured and looked at longitudinally across all the patients that come through that hospital. There is a lot of work being done now around patient outcomes. But adding GenAI within a patient focus, privacy, regulations and governance are much more complicated. If you don't have an Office of AI and rock-solid governance, you're never going to get that through any major hospital.
 

Q: What about outcomes reporting? How do you draw a direct line from AI/GenAI projects to value delivered in customer organizations?

Charlie Wardell: The IDC report about an enterprise intelligence strategy was pretty clear, that organizations that embrace an enterprise intelligence strategy have three to four times better outcomes. So certain types of metrics will have to be measured, the same as any enterprise initiative: What’s the projected ROI of the project? How is it going to help us? How is this going to reduce cost or make us more nimble?

Rob Shear: I see outcomes being measured at the business unit level, depending on the way the company is structured. So if you're working a project around developer productivity, you’re measuring lines of code or quickness of rollouts. In manufacturing, for instance, you’re measuring increased worker productivity, reduced error rates in the facility and improved overall equipment effectiveness (OEE).
 

Q: What questions are our customers asking us about AI and Gen AI?

Rob Shear: The biggest one is where to begin. They have a thousand possible use cases they can and want to do, so they’re stuck in analysis paralysis. I think that’s where a consulting engagement with DXC can help. DXC comes in with the maturity and expertise to help the company develop a plan and pick some projects to get started with out of the gate. Then they can ultimately measure the success of the venture.

Charlie Wardell: Agreed. Many customers struggle with knowing where to begin. Everybody knows they need to be in this space; they just don't quite know how to get started. GenAI is moving so fast — language models are improving every day and are absolutely mind blowing — and as the landscape changes, you become a little bit more insecure with what you thought you knew yesterday. GenAI is exciting and revolutionary, and you don't want to inhibit that excitement in an IT organization. We just need to be mindful of the risk that we experienced with data fragmentation and the resulting untrusted nature of the data. If GenAI initiatives are not managed with a strategy in mind, we will see AI fragmentation or AI sprawl and, like data sprawl, there will be a reckoning in trying to bring it all back to center.

It's better to have a strategy. What I’m hearing from the customers I have been speaking to is that they want us to help them with the vision, help them avoid mistakes and help future-proof their decisions.
 

Q: What would you say to companies wanting to get started?

Charlie Wardell: The strategy I'm advocating is simple: Start with a centralized AI strategy. Even though you may not have the specific use case, we all know the need for centralizing AI, standardizing and governing it. This is where the DXC-Dell partnership has been amazing. The Dell AI Factory is a great place to start. It has all the components you need to build out your AI strategy and let your data scientists (or whoever is responsible in your organization) start in an environment that is going to be part of an enterprise strategy going forward. This minimizes AI fragmentation or AI sprawl, and enables you to start building the layers of data management on top of that AI strategy, and then further layers of governance, integration and decisioning. Most organizations want a GenAI strategy now. DXC has a strategy that will protect your investment, scale, and provide for heterogeneous hybrid workloads. So start with that. DXC has a managed service for AI and we can guide customers into their GenAI path and work with you to optimize your AI and enterprise intelligence strategy.
 

Q: Looking forward, what will the DXC-Dell partnership be focused on this year around AI/GenAI?

Rob Shear: We've had a 20+ year relationship with DXC going back to the CSC days. DXC is one of our top partners across all of Dell, and their bringing GenAI consulting and use cases is so critical to success in this space. Dell makes great platforms and we have great partnerships with ISVs and great solutions, including with NVIDIA. In order to be successful and do this at scale, we need DXC as a partner, to bring value beyond the platforms. I'm super excited going into FY25 / FY26, to expand this relationship, grow this business and bring business value to our customers. As regards GenAI, we need to draw all parts of DXC — consulting, global infrastructure services and global business services — together with Dell and NVIDIA and bring this “power of three” to customers to show ultimate value.

Charlie Wardell: The “power of three” messaging really sets us apart from everyone else in the industry. DXC, Dell and NVIDIA have a cohesive, well thought-out strategy across services, hardware, compute capability and  the integration that goes with it — whether it’s edge or cloud or on prem. I think the opportunities are amazing.